3D Art

The

& Design Inside

r Photoshop
r 3ds Max r Maya
r Cinema 4D

Boo Volume 3

5IF
VMUJNBUF
HVJEF
UP
HFUUJOH
UIF
NPTU
PVU
PG
ZPVS
%
QSPKFDUT
r ZBrush

$IBSBDUFST
t
7FIJDMFT
t
"SDI
7JT
t
1IPUPSFBMJTN
t
"OJNBUJPO
 Welcome to
The
3D Art
& Design
The world of 3D art and design is vast and continuing to grow. It’s an exciting place for an
artist to develop skills and techniques in order to ultimately express their artistic creativity. The
third volume of The 3D Art & Design Book not only assists in developing these areas, but also
provides inspiration – the other essential ingredient. Whether you follow along with the tutorials
step by step, or simply use the ideas and techniques to create your own, original artwork, the
choice is yours and, either way, you are guaranteed to create something to be proud of. Split into
six key areas of art and design, the book explores character, architectural visualisation,
photorealism, environment, vehicle and animation. There is certainly something for everyone on
the pages within, and perhaps you will discover a passion for something you hadn’t considered
before. If that wasn’t enough, we’ve also included a free CD packed full of 3D models, video
tutorials and source files in order to help you complete many of the tutorials found in the book.
 The
3D Art
& Design
Imagine Publishing Ltd
Richmond House
33 Richmond Hill
Bournemouth
Dorset BH2 6EZ
 +44 (0) 1202 586200
Website: www.imagine-publishing.co.uk
Twitter: @Books_Imagine
Facebook: www.facebook.com/ImagineBookazines

Head of Publishing
Aaron Asadi

Head of Design
Ross Andrews

Production Editors
Hannah Kelly and Sarah Harrison

Senior Art Editor

Greg Whitaker

Design
Lauren Debono-Elliot

Photographer
James Sheppard

Cover image
Matthew Burke

Printed by
William Gibbons, 26 Planetary Road, Willenhall, West Midlands, WV13 3XT

Distributed in the UK & Eire by

Imagine Publishing Ltd, www.imagineshop.co.uk. Tel 01202 586200

Distributed in Australia by
Gordon & Gotch, Equinox Centre, 18 Rodborough Road, Frenchs Forest,
NSW 2086. Tel + 61 2 9972 8800

Distributed in the Rest of the World by

Marketforce, Blue Fin Building, 110 Southwark Street, London, SE1 0SU.

Disclaimer
The publisher cannot accept responsibility for any unsolicited material lost or damaged in the
post. All text and layout is the copyright of Imagine Publishing Ltd. Nothing in this bookazine may
be reproduced in whole or part without the written permission of the publisher. All copyrights are
recognised and used specifically for the purpose of criticism and review. Although the bookazine has
endeavoured to ensure all information is correct at time of print, prices and availability may change.
This bookazine is fully independent and not affiliated in any way with the companies mentioned herein.

The 3D Art & Design Book Volume 3 © 2014 Imagine Publishing Ltd

ISBN 978 1909 758 933

Part of the

bookazine series
 08
The evolution of CG software
Character Architectural Photorealism
18 Create a portrait visualisation 70 Make a realistic human portrait
24 Render your character's skin 48 Build a wooden interior 77 Gallery: Then Flying The Youth
26 Sculpt intense facial expressions 52 Build C4D arch-vis assets 78 Design a stunning dress

30 Design and model sci-fi characters 56 Make atmospheric renders 84 Create hair and fur

37 Gallery: Equilibrium 63 Gallery: Butterfly 91 Gallery: Photorealistic CGI

38 Mesh-shatter your characters 64 Add post-production touches 92 Bring still-life objects to life

45 Gallery: Fatal 67 Gallery: Contemporary Living Room

6 3D Art & Design

 178
161 154

171
Environment Vehicle Animation
100 Build a city in 60 minutes 134 Model a complete car 166 Animate a dialogue shot
106 Use modular textures 141 Gallery: No More Gas 171 Gallery: CTN Animation Expo Kirk & Lucy

112 Create a videogame environment 142 Build a vehicle game asset 172 Shatter moving characters
118 Build matte paintings 149 Gallery: AEG27Cern 05 178 Master blend cycles
124 Create underwater renders 150 Create Bunkspeed renders
185 Gallery: General Shellshock
131 Gallery: Seedy City 154 Model a fighter jet
186 Animate vehicles
161 Gallery: Waldo
162 Create wireframe renders

3D Art & Design 7

 Evolution of CG software
THE

EVOLUTION OF

CG SOFTWARE

From early wireframes to Avatar, we talk to the

pioneering founders of the most prominent CG
software, finding out where the tools came from and
how they got to where they are today
8 3D Art & Design
 Will Anielewicz
Head of computer animation at
Omnibus. Early work at Alias
working on PowerAnimator

Jim Blinn
Inventor of Bump-mapping, the
Blinn-Phong shading model.
Researches realistic rendering

Kim Davidson
Co-founder, chief executive
officer and president of Side
Effects Software Inc.

Greg Hermanovic
Founder of Derivative and
co-founder of Side Effects
Software Inc.

Tom Hudson
Main developer of 3D Studio, as
well as its precursors and
prototypes (Yost Group)

Tilo Kühn
Senior developer and
programmer of CINEMA 4D at
MAXON studio

Chris Landreth
Hired to define, test and abuse
software such as Maya. Has
Oscar-winning animation work

Philip Losch
Senior software engineer at
MAXON. Co-writer of the
precursor of CINEMA 4D

Terrence Masson
Author of Cg101: A computer
Graphics Industry Reference.
Technical director at ILM

Rob Powers
President of the LightWave 3D
Group, a division of NewTek Inc,
with numerous film credits

Fred Ruff
Lead designer for 3ds Max.
Worked with the original creator
of 3D Studio (Bent Image Lab)

Gary Yost
Leader of the Yost Group, which
created and developed an early
version of 3ds Max
Evolution of CG sofware

3D Art & Design 9

 Evolution of CG software

W
e may not think about it computer-generated sequence for Star Trek 01
often, but it’s difficult to II: The Wrath of Khan. Barely a month later a
go even one waking day huge 20 minutes of full 3D CGI created by
without some form of CG Information International Inc, Robert Abel
somehow making its way past our eyes. and Associates, MAGI and Digital Effects
Whether in adverts, music videos, YouTube was revealed within Disney’s TRON release.
streams, screensavers, or any other number It was only a matter of time before
of computer-generated imagery, the art form commercial software companies with
has fully entrenched itself in our day-to-day products made specifically for computer-
lives. Strange then, to think that just 50 years generated imagery started to trickle into the
ago there was effectively no such thing. public sphere, with many co-developing their
By the late Sixties, the few computers in software together with the in-house
the world capable of producing any form of software at successful studios such as ILM.
line-drawn animation cost the equivalent of As the hardware also began to lower in price,
nearly $700,000 today, were the size of a the early Eighties marked the introduction of
bunk bed and used only paper tape. Even the first startups, as well as successes and
02
computer science itself was so unheard of failures of businesses such as Autodesk,
that when computer graphics pioneer and Alias Research, Wavefront and Omnibus
inventor of Bump-mapping Jim Blinn
attended the University of Michigan in 1967, HOUDINI & PRISMS
it wasn’t even offered as a subject. “My big break came in 1985,” begins Kim
However, during the early Seventies, hubs Davidson. “I was hired at Omnibus
like the University of Utah began to Computer Graphics in Toronto as a
showcase 3D imagery produced by pioneers programmer, helping it develop its internal
like Blinn. There were early examples of 3D animation software.” The largest
shaded 3D CGI, produced with frame- computer graphics company in the world,
buffers for display on TV monitors, and even Omnibus was already using mainframe
the first 3D, computer-animated graduate computers to create flying logos and station
shorts like ‘A Computer Animated Hand’, by IDs as well as boasting state-of-the-art
Fred Parke and future Pixar founder Ed computer animation studios in Toronto, New
Catmull. Nevertheless, CG remained York and Los Angeles. It was here that
something only a programmer could Davidson first met Greg Hermanovic who,
realistically develop for some time to come. after a background as a programmer
Also, bizarre by today’s standards, special working on everything from real-time flight
effects in the movie industry were by no simulators to a training simulator for the
means prevalent. In fact, when George Space Shuttle’s robot arm, was hired as the
Lucas began to produce Star Wars in 1973, new head of the Toronto R&D team.
he quickly discovered that the Special Effects Together, they began developing the early
Department at 20th Century Fox was version of Houdini’s predecessor: Production
actually no longer operational. of Realistic Image Scene Mathematical
It was in order to create the effects Simulation, or PRISMS. However, their time
required to match Lucas’ vision that at Omnibus was short-lived. In 1987, after
Industrial Light and Magic was born in 1975, having swallowed both Digital Productions
and the result of its work is hailed today as and Robert Abel & Associates, the company
one of the first examples of the modern-day that had seemed so powerful suddenly went
blockbuster. Star Wars even included 3D bankrupt under the weight of its many debts.
effects. A year after a brief 3D-digitised “I don’t think of myself as entrepreneurial,
representation of a face and hand was first or a risk-taker, but I loved it so much that
seen in Futureworld, 1977’s Star Wars Episode when Omnibus went bankrupt in 1987, I
IV contained the first truly extensive use of didn’t hesitate to jump into starting a
3D computer-generated imagery and company so I could continue working on
animation ever seen in a feature film, during computer graphics in Toronto,” explains
a very basic, untextured wireframe sequence Davidson. “There was nothing close to what
of the Death Star trench. we had created at Omnibus and that was
“Larry Cuba, back in the Sixties and the reason for starting Side Effects.”
Seventies, single-handedly completed the Luckily, the struggling Omnibus had
wireframe of the Death Star simulation in the licensed PRISMS to other companies in Kim Davidson
original Star Wars,” explains Terrence order to bring in extra revenue. This meant was hired by Omnibus
Computer Graphics in
Masson, author of CG101: A Computer that when Davidson and Hermanovic 1985 and went on to
Graphics Industry Reference. He tells us that by acquired the rights to PRISMS, they had their found Side Effects
Software in 1987
1982, ILM had produced the first entirely first licensed customers right from the start.

CG software: a few
key events
1968

1968

1969

First public demonstration Evans & Sutherland founded, UNIX operating system developed ‘A Computer Animated Hand’, a short by
1972

introducing the computer marking the first computer by Kenneth Thompson and Ed Catmull and Fred Parke, is developed
mouse by Doug Englebart of SRI graphics company Dennis Ritchie at Bell Labs at the University of Utah
10 3D Art & Design
“By the end of the Sixties it got advanced enough that people began
realizing it had potential outside the research lab” Terrence Masson
 Evolution of CG sofware
03

01 “The procedural
version of PRISMS started
to see the light of day
around 1989,” explains
Hermanovic. “It became
popular because it
automated and visualised
what CG artists were
already doing blindly”
02 PRISMS also included
Mojo, a morphing package,
as well as motion-control
interface Moca and ICE
03 The Omnibus team in
1987. The struggling
company had licensed
PRISMS to companies in
Japan and the UK a few
months earlier, so when
Davidson and Hermanovic
acquired the rights they Emerging as early as 1989, PRISMS
already had their first was extensively tested by the
licensed customers founders of Side Effects, Kim
04
Davidson and Greg Hermanovic
04 SIGGRAPH 1993

In their first years running Side Effects, the of a sudden, Side Effects found itself in
co-founders continued to not only keep competition with three public companies all
The need to innovate coding the software but also to use and test vying for a market share. With Houdini, the
“It’s amazing how many advancements have happened in
computer graphics over the last 25 plus years. Mobile phones
it themselves as a production company, co-founders aimed to shift from PRISMS’ C
have graphics capabilities that far exceed the early graphics which enabled PRISMS to improve all the to C++ and take advantage of the
workstations and at a fraction of the cost and size,” says Kim more quickly. “The procedural version of advancements in hardware graphics.
Davidson, co-founder and president of Side Effects. “However, I’m PRISMS started to see the light of day However, they also made the conscious
frustrated that Side Effects and the industry aren’t further than we
around 1989,” explains Hermanovic. “It decision to switch tactics, rising above the
are. There were animated CG humans 25 years ago… The
industry is just doing more a lot easier, but I don’t think we are became popular because it automated and competition by focusing on a segment of the
doing it a whole lot better. The majority of animation is still done visualised what computer graphics artists market where they felt they could make the
by an animator sitting with a mouse moving a limb and creating a were doing blindly and non-interactively most difference: film VFX.
keyframe and it’s still nearly impossible and expensive to create before.” With their client base growing “It was a good move for us,” Davidson
photoreal CG humans. John Pennie, president of Omnibus, said in
1986 that in two years we would be creating CG humans that were
rapidly, the pair soon made the decision to concludes, something that is more than
indistinguishable from real actors… I knew it wouldn’t be two pull away from production to focus solely on clear, given that Houdini is now used by
years, but I thought it would be a reality 27 years later. So for all developing the software, and by the Nineties leading film studios and Side Effects has won
the accomplishments we have obtained, there’s still lot to do. PRISMS had been used to produce the visual three Academy Awards in recognition of its
That’s exciting to me and that continues to be our motivation.”
effects for feature films as impressive as phenomenal work. “The inspiration for
Apollo 13, Twister and Titanic. Yet despite this PRISMS was to create efficiencies for
success, Side Effects had long been thinking ourselves, as clients are always changing
NEW STUDENTS ARE BETTER IF about where to take the software next. their mind. It’s satisfying to see how well the
“Although Houdini 1.0 shipped in overall architecture and procedural
THEY UNDERSTAND THE December of 1996, we had started the paradigm has stood the test of time. While
EVOLUTION OF THE SOFTWARE design of Houdini in 1991,” reveals Davidson. we have rewritten some big chunks of
AND WHAT’S BEEN MADE WITH IT The Nineties saw the start of exponential Houdini and added tons of features, the
– THAT’S ABSOLUTELY CRITICAL growth in the 3D industry thanks to further basic architecture and data flow have
TERRENCE MASSON, independent VFX supervisor advancements in affordable graphics and, all remained the same.”
University of Utah

AMC-SIGGRAPH, the special interest Martin Newell develops the now Bump-mapping is
1978
1974

1975

group of the Association for Computing famous Utah Teapot, now one of the introduced by Jim Blinn
© 1972 The

Machinery, is first established most iconic images in the world of CG and Martin Newell
3D Art & Design 11
“The other big milestone probably was the explosion of
commercialisation in the Seventies “ Terrence Masson
 Evolution of CG software

04

05

Today, Will Anielewicz works at Caustic

Graphics Inc, writing shaders. “One of my main
roles at ILM was to match the in-house prman
shaders using mental ray. That history makes
me the perfect person for this task,” he explains 06

MAYA & ALIAS 1 during Toronto’s harsh winter months, a lack developing software – which was later to
Before Greg Hermanovic took over as head of funds lead to a new change in focus: Alias become Maya – in an actual animated film to
of R&D at Omnibus, Will Anielewicz had the 1. The team aimed to produce a system that test the software,” Landreth continues.
job. After studying a Masters program – would enable users with minimal training to In 1995 Silicon Graphics bought Alias
during which his computer science advisor achieve complex computer animation. Research and Wavefront Technologies,
declared his project was not viable due to However, in 1985, Alias signed a landmark merging the two companies. Landreth spent
there being no future in computer art – deal with car manufacturer GM, receiving a the following year developing multiple new
Anielewicz set out to prove him wrong. “In sizeable investment out of an interest in using Maya features, putting them to the test on
1979 I was hired by Omnibus Video to head the spline-based software for design instead. his short film, ‘The End’, which in turn earned
up its Computer Animation division,” he tells “I created the first prototype graphic him an Academy Award nomination. In 1998,
us. “Although I knew nothing about interface for this system in a few days,” the new Alias|Wavefront merger finally
animation, they hired me anyway since they Anielewicz says. “Our first demo was mostly released the first version of Maya, which
got no replies for their ad: ‘Required: smoke and mirrors but we won the contract.” quickly became a new industry standard. The
04 Interior Pool by Experienced computer animator.’ I was the Chris Landreth was hired and started team continued to develop it until 2006,
Ryan Montrucchio
and Will Anielewicz, only employee for almost the first year.” working at Alias in 1994. “It was a good time when Autodesk acquired the company. “I
Geometry: Tasso It was only three years later, when to start because the company really wanted
Ringas and George
Nikopoulos; Omnibus was becoming a powerhouse, that to show its potential market that the software
Software: Caustic
Visualizer for Maya
he first heard of a tiny startup founded by it was developing could be used to do I HAVE ALWAYS
Stephen Bingham, Nigel McGrath, Susan production.” Alias’ profits had soared
05 An example of
McKenna and David Springer, named Alias throughout the late-Eighties and early
THOUGHT OF MYSELF
plotter art by Will
Anielewicz, created Research. “They welcomed me with open Nineties, as its second- and third-generation
AS AN ARTIST. THE
from 1976-79
arms. I now had three years of experience in Alias and PowerAnimator software began to COMPUTER IS BOTH
06 Image from
‘Subconcious
computer animation, which in those days be used for Academy Award-winning VFX MY BRUSH AND
Password,’ Chris
Landreth’s latest
was a relative lifetime,” Anielewicz explains. on films such as The Abyss, Terminator 2, COLOUR PALETTE
short film, which he
Working in a run-down warehouse where Jurassic Park, Forrest Gump and The Mask. WILL ANIELEWICZ, software engineer,
created using Maya the computers doubled as room-heaters “What my job turned into was to use Alias’ Caustic Graphics Inc
1985

Autodesk is 20 minutes of full 3D Jim Blinn receives the first Alias founded in Toronto by Stephen NewTek
1982

1982

1983

1983

founded in CA CGI released within ACM SIGGRAPH CG Bingham, Nigel McGrath, Susan founded by
by John Walker Disney’s TRON Achievement Award McKenna and David Springer Tim Jenison
12 3D Art & Design
1980s
 Evolution of CG sofware
really did not expect Maya to still be in use,” “We were motivated by a lot of things, not WE ALWAYS SAID THAT WE
reveals Anielewicz. “It’s been almost 30 least of which was our royalty-based licence
years. I use it every day.” with Autodesk,” says Yost. “We were in our
WERE TRAINING OUR USERS TO
30s, so if what we did became popular we’d BECOME MUTANTS WHO WANTED
3DS MAX & 3D DOSR1 see enough financial reward to make all the TO PLAY GOD, CREATING WORLDS
“I’d been going to SIGGRAPH since 1986, intense work worthwhile. We were on what WITHIN THE COMPUTER
drooling over the SGI Iris workstations and we called a ‘deferred life plan’ for so long.” GARY YOST, leader of the Yost Group
software from companies such as In addition to the extra funds a royalty
Wavefront, Alias and Vertigo,” begins Gary agreement brought, the team were also
Yost. “However, those solutions cost motivated by their users. At the start of each
upwards of $40-50,000 at the time and 18-month development cycle, Yost would
were totally unavailable to mere mortals such spend three months collecting feedback to
as my friends and I.” help develop the feature set for the next
Yost had already founded Antic Software version of the software, before embarking on
in 1984 and, together with fellow developer development over the next 14 months. “That
Tom Hudson, had been working on the enabled me to rapidly develop ideas for new
creation of a set of 3D tools, before becoming features. Tom, Don, Rolf and Dan took every
limited by the graphics their budget could suggested function as a personal challenge. I
afford. Then, in 1987, they were asked if they don’t remember any of them ever telling me
would be interested in developing 3D that something couldn’t be done. Our motto
software for the PC under license to the was ‘it’s only software, so why not?’”
ever-growing Autodesk. Each member of the group worked from
By 1988 the Yost Group was formed with home in different cities, sending data files
Gary Yost at the helm. The small team began back and forth over slow modems. The main
working on a modelling and rendering goal for the developers was to essentially
application internally, code-named THUD democratise 3D software, making it possible
after Tom Hudson, who was the only for creative people who were working for
07
programmer on the project. Impressively, the others to be able to begin working for
first version of 3D Studio DOS was released themselves. “I think a huge part of 3D
only two years later. “The response to DOSr1 Studio’s legacy that continued into 3ds Max
was huge,” remembers Yost. “There was just is the concept of making easy-to-use
a lot of pent-up demand for something like software that is priced at an affordable level,”
that. Once it took off, we just got on the train explains Hudson. After unveiling the new
and didn’t look back.” software at SIGGRAPH in 1995 to a shocked
There were two huge shortcomings with audience, the Yost Group delivered 3ds Max
3D Studio DOS. There was no Undo function in 1996. By 1997 the team decided it was
and there was a lack of animatable controls, time to start living out their previously
meaning that once the user created deferred life plan.
something, the mesh would be fixed aside “Between 1997 and 1999 we transitioned
from scaling, rotation or morphing. The core the code to a great in-house development 08
group made up of Yost, Hudson, Dan Silva, team at Autodesk and then went on to other
Rolf Berteig, and Don Brittain were soon things,” Yost continues, “I tear up a bit when I
joined by Mark Meier and Gus Grubba. They recall how exciting it was to wake up every
spent the next seven years working to rewrite morning after only four hours sleep and get
the entire code base of 3DS DOS, leaving to work. Never did so few developers make
what would become 3ds Max. so many people so happy.”

Giving up on 3ds Max

“In 1994, when we started writing code in earnest, our source code control/versioning system
was running over analog modems from a server in Don Brittain’s house,” explains Gary Yost,
leader of the Yost Group. “Within the first two months of development it was obvious that dialup 09
was too unreliable for us to share code in this way.” Incredibly, by September the developers had
almost given up work on 3ds Max. “I had a conversation with Autodesk’s then-CEO Carol Bartz
07 ”This is my character, Klanky the Robot, with the cornerstone character
in early October, saying that I thought we’d have to give up the project unless we could figure out
from my first 3D Studio DOS animation, ‘Corner Stone’,” says Hudson. “The
something else to do,” says Yost. “Just that week, Gus Grubba found that the new Integrated animation was created with a pre-release version of 3DS DOS and I had
Services Digital Network (ISDN) system was being deployed nationwide and that it would thought the scene files were lost, but I found them in 2010”
support what we needed. Between October and December we were able to get the brand-new 08 “I felt like a roadie working with rock stars! I remember Gary Yost and his
ISDN service and that solved the problem, enabling us to get on with development in earnest in Hawaiian shirts,” says Fred Ruff, who worked as a lead designer for 3ds Max
January. What’s amazing is that only seven months later, at SIGGRAPH LA in July, we unveiled and created this image using an early version of the software
Max along with the Character Studio/Biped plug-in by Michael Girard and Susan Amkraut. Max
09 Right after the launch of Max at SIGGRAPH in 1996. From left to right: Jack
was kept secret until then and nobody expected us to ditch all of our 3D Studio DOS code.”
Powell, Don Brittain, Rolf Berteig, Mark Meier, Martin Doudoroff, Gary Yost,
Dan Silva, Kyle Peacock, Tom Hudson. Missing is Gus Grubba
1988
1986

1986

MAXON formed by current Steve Jobs purchases the Computer Omnibus caves to Side Effects Yost Group
1987

1987

CEOs Harald Egel, Harald Graphics Division from George Lucas and crippling debt and Software formed with Gary
Schneider and Uwe Baertels establishes Pixar files for bankruptcy established Yost at the helm
3D Art & Design 13
 10 11

Image by Philip Losch using early

CINEMA 4D. Losch still works as one
of the most important developers for
the software today I THOUGHT OF BUMP-MAPPING WHILE LOOKING
AT MY SHOES. I THOUGHT WHAT MAKES THEM LOOK
BUMPY IS NOT THE HEIGHT OF THE DISPLACEMENT
BUT THAT THE ANGLE OF THE SURFACE CHANGES
FROM ONE PLACE TO THE NEXT
JIM BLINN, computer graphics pioneer

LIGHTWAVE showcased all of this amazing new

When Tim Jenison founded NewTek in 1985, computer-generated imagery made with the
he too aimed to develop sophisticated yet LightWave 3D software. The genius behind
affordable tools that everyone could use. In those cutting-edge images was Ron
1990 the company first began to enjoy Thornton, who subsequently released video
widespread fame with its release of the tutorials that were excellent learning tools.
Video Toaster for the Commodore Amiga They helped me build my LightWave 3D skill
personal computer, which also incorporated set, which made it very easy to get work in
the first version of LightWave 3D. Replacing Hollywood at that time, because computer-
video-processing software that had generated graphics were essentially in their
previously cost around $100,000 for less infancy. It was a golden opportunity for a
than a tenth of the price, today Jenison is still student fresh out of film school.”
considered the visionary force behind the
desktop video revolution that followed. FASTRAY & CINEMA 4D
Despite the success, NewTek ran into “In 1988, when I was 13 years old, my
problems after the failure of the Commodore parents bought an Amiga computer and I
Computer Company and the Amiga was instantly fascinated by its graphics
computer, subsequently affecting the Video capabilities,” begins Philip Losch. “Shortly
12
10 Spinning top from The Mechanical Toaster. “This caused NewTek to port after this, my brother and I decided to write
Universe (1985) by Jim Blinn, forming LightWave to Windows and Mac computers our own raytracing software.” Aiming to
part of a 52-part telecourse to teach
college level physics. During these as a standalone product not associated with convince MAXON Computer of their newly
productions he developed other the Video Toaster,” explains president of the written software’s potential, the brothers
standard CG techniques
LightWave 3D Group, Rob Powers. entered it into a monthly contest run by
11 Images by Ron Thornton, an
innovative LightWave artist who NewTek survived the challenge and the MAXON’s Kickstart magazine and promptly
created the iconic CG in Babylon 5 standalone version of LightWave has since won the competition. Little did they know,
using the software
enjoyed worldwide success. “The founder they had just developed the very beginnings
12 “I never understood why in most
Tim Jenison is one of those unique genius of what would become CINEMA 4D.
applications you had to define a light
source, a camera and material before personalities that easily inspires others,” The Losch brothers were hired as
rendering… We implemented a effuses Powers. “While I was in school, developers and, in 1991, FastRay was born.
different behaviour in CINEMA 4D,”
explains Losch Babylon 5 premiered on television and it “To be honest, I would say FastRay was just
1990

1990

1994

NewTek Video Toaster, for 3D Studio released by the FastRay developed by DreamWorks SKG founded
1991

the Commodore Amiga Yost Group only two years brothers Christian and by Steven Spielberg, Jeffrey
computer, is launched after the group was founded Philip Losch Katzenberg and David Geffen
14 3D Art & Design
1990s
 Interviewing Jim Blinn
Computer graphics legend Jim Blinn (see www.
jimblinn.com/publications for his work) discusses
what led him to take an interest in 3D art, as well as
his research into the Blinn-Phong shading model.
When did you first gain interest in CGI and 3D imagery?
I first took the introductory programming course in spring 1967
while an undergraduate at the University of Michigan… Somehow
during this time I saw a picture of a graphics terminal and knew
right away that was what I wanted to play with.
While there was no official major called Computer Science at
the UofM there were several programming courses in the Science
department and in the Engineering School. I tried to sign up for all
the ones I could. I finally declared a double major in Physics and in
13 Communications Science – a sort of precursor to what later
became the Computer Science department.
The computer I used at UofM was expensive and there were
only a few of them in the world. I did a few line-drawing animation
projects with students in the Art School but never really dreamed
13 The Utah Teapot is one of a handful of that the computers would get so cheap so fast that artists would
iconic models from the early have their own dedicated machine.
development of 3D computer graphics,
having been developed by Martin Newell Can you tell us more about your time at the University of Utah?
and modified by his co-worker, Jim Blinn, What led you to first begin research on realistic rendering?
in 1975 I was first made aware of the UofU when my office mate at
14 Image created by Philip Losch using Michigan showed me the PhD thesis of Utah student Gordon
early CINEMA 4D. When he was just 13 Romney (in 1970) that had some early crudely shaded images. I
years old, he and brother Christian Losch was fascinated by the images but the computers I had a UofM
14 wrote their own raytracing software – the were not capable of displaying shaded images. I continued to play
earliest predecessor to CINEMA 4D with line drawing for several years, but in 1974 I decided to apply
to UofU graduate school to join the crew making images there. I
had kept up with the developments at UofU during the intervening
a big learning exercise!” says Losch. years ago are still on board,” continues years and was particularly intrigued by the work of Ed Catmull.
“Everything we did was self-taught and at Losch, who is himself still one of the most When I got to UofU I was surprised to find that they didn’t have
the time it was very hard to come by any real important developers at MAXON. computers that could display shaded images either. All their
information. We often went to the library “Everyone on our team works from home in images were created by doing a time exposure of the computer
display slowly scanning out the results of their computations. I got
and ordered books or scientific papers that a relaxed atmosphere and loves 3D – it’s not there just as they were getting the first frame buffer from Evans &
then would arrive weeks later.” He explains just a normal job.” Sutherland Computer Corporation, which could display a shaded
the developers faced significant challenges image on a TV monitor.
at the start – not least because though they BEHIND THE PIXELS I started implementing Catmull’s rendering algorithm and began
tinkering with it. I didn’t have a long-term goal, but just began
had entered a very new industry, by the early As exciting as it can be to theorise on what’s tweaking various lighting and shading algorithms to make the
Nineties many competitors had already next for CG, it’s important to remember the images look better. I guess that’s symptomatic of my career – I just
existed for five or even ten years. past – to think on the people that brought us did what seemed interesting at the time and wound up being
However, as the team began to get to where we are today. “The people that involved in some of the most interesting projects.
increased access to literature and several have made this industry – they’re just the What first led you to create realistic specular light models, bump
examples of the competing software itself, smartest, the most creative and the most mapping and environment/reflection-mapping?
Environment mapping began as Martin Newell’s idea. We were
Losch realised there were several crucial friendly, it’s so humbling,” praises Masson. teaching the Computer Graphics course at Utah at the time. He
ways their software could stand out. “20 It’s interesting to consider that, despite his was a professor there and I was a teaching fellow, and we spent a
years ago I never understood why in most years of research on the history of the lot of time taking about rendering and geometry in preparing for
applications you had to define a light source, computer graphics industry for his CG:101 the lectures. I think we taught each other more than we ever
taught the students. He came in one day and told me he had been
a camera and material before rendering. If books, when Mike Wahrman, a friend who thinking in the shower and said, ‘What if were to take the
you left out any of those steps, your render used to work at Wavefront, asked Masson calculation for the lighting and, instead of doing the regular dot
would show up pitch black,” he explains. “So whether he knew who first wrote the product with the light source that would give you the colour of the
we implemented a different behaviour in standard OBJ file format, he had no idea. surface, what if you took the light and reflected it backwards?’
I first thought of Bump-mapping while looking at my shoes.
CINEMA 4D.” MAXON’s new software Together, they had asked everybody who They were made of leather but they had this embossed pattern in
became focused on avoiding the same had worked with Masson in the Eighties and, the surface, and I was looking at the highlights on them and
difficult UI and many illogical steps required despite how widely used OBJ is today, thought what is making these things look bumpy is not the height
to do even basic tasks in other software, nobody could remember. of the displacement but that the angle of the surface changes
from one place to the next. I realised I could nudge the surface
making 3D more accessible, inexpensive and “Just about everybody who started this normal from one place to the next, plug that into the lighting
as fun to use as possible. “Pretty much every industry in the Sixties and Seventies is still equation and that would be a good approximation of a more
time you need to consult the manual, it’s a alive and active, but it’s getting to the age realistic effect.
failure on our side, because we didn’t design that we’re starting to lose some of our Did you expect 3D imagery and software to be able to reach the
an interface or component obvious or founding members. They’re going to pass heights they are at now in so relatively short a time?
intuitive enough,” Losch reveals. away and even before they do, memories go I’ve never used any commercial software, only seen it from a
distance. I come from an era where everybody did his or her own
As the years passed and other as well; like the OBJ file – somebody made programming. Having off-the-shelf software that’s reliable
competitors began to fail, he tells us it was a an OBJ file but nobody knows who it is,” enough for non-programmers to use is an impressive feat. CG has
matter of persistence coupled with a Masson continues. “It would be a tragedy to already gone way beyond anything I could have dreamed of. There
committed team and healthy working have some of these folks pass away with are two aspects to visual creative art: figuring out what to draw
and figuring out how to draw it. The second of these is what a lot
environment that led to CINEMA 4D being a these histories in their head… I’m interested of people think about and teach, but the ‘what to draw’ part is
success today. “Most developers that in the people behind the pixels. I want the more personal and subjective.
started to work for us five, ten or even 15 stories of how all this happened.”
1998

1998
1996

1996

Pixar releases Toy Story, the Houdini 1.0 ships 3ds Max delivered after Alias|Wavefront first introduces Blender is first
1995

world’s first computer- after five years being initially unveiled at its new flagship 3D product, released to the
animated feature film in development SIGGRAPH in 1995 Maya, to the 3D community public by NaN
15
 Character
Bring your character
concepts to life
45 26

18 Create a portrait
Follow this ZBrush workflow to create a
3D interpretation of a character concept

24 Render your character’s skin

Learn how to use ZBrush in order to
render your sculpts

26 Sculpt intense facial expressions

Sculpt a character, building emotive facial
expressions using ZBrush

30 Design & model sci-fi characters

Learn to create a highly complex and
functional armoured alien character

37 Gallery 24
Equilibrium by
Daniel Arnold-Mist

38 Mesh-shatter your characters

Create an abstract character constructed
from various pieces of geometry

45 Gallery
Fatal by
Adam Martinakis

18 38

30

16 3D Art & Design

 You feel a great sense of
achievement when you create a
fully developed character from
its initial concept
37

3D Art & Design 17

 Character

Easy-to-follow guides
take you from concept
to the final render
Create a
Artist info

Adam Fisher
portrait
Personal portfolio site
www.afisher.com.au
The Prisoner 2013
Country Australia
Software used Follow this ZBrush workflow to create a 3D
ZBrush 4R5
Expertise 3D character art for
interpretation of a stylised character concept
the videogame industry
Adam Fisher works as a 3D artist for the videogame and simulation industries

T
aking the concept of a menacing
convict, I’ll be taking you through
my workflow for creating a
Establish initial form
character bust. We’ll go over the whole
process, from sculpting the main forms to
Block in the basic shapes of the character
refining the model with pore and wrinkle
details. We’ll be using DynaMesh to create
the clothing and the chain, FiberMesh to
add facial hair and we’ll also go over some
01  Build the base mesh With the DynaMesh tools in
ZBrush it doesn’t really matter if you’re starting with
a sphere or a pre-existing base mesh. The main thing is to
Polypainting techniques to bring the keep the subdivision low at this point. I like to keep my brush
character to life. size large and avoid zooming in on the character early on.
ZBrush is the tool of choice for this This way I’m only concentrating on the overall silhouette of
process because it gives us the freedom to the character. To start we’ll concentrate on getting the main
create without the need to jump between shapes and proportions of the head, neck and shoulders.
multiple software packages or worry about Hitting Y will toggle between the default white and black of
01
technical limitations. Once our character the SubTool, enabling us to check the silhouette easier.
has been sculpted and Polypainted, we’ll
pose him and render multiple passes to be
composited in Photoshop. 02  Work out the major forms I like to exclusively
use the Move brush early on to manipulate the
major shapes into place. Looking at the concept we’ll try to
establish where the eyes, nose and mouth sit in relation to
one another. At this point we can begin to subdivide and,
Control subdivision using the ClayTubes brush, block in the major forms and
When starting a sculpt it’s important to structure. Using the DamStandard brush we can also cut in
keep the subdivision level low, especially some guidelines for skin folds and concave shapes, such as
when blocking in proportions and main the nasolabial fold, philtrum, eyelids, brow and lips.
shapes. You can then slowly subdivide

03 
(Cmd/Ctrl+D) to add polygon density as Add eyes and teeth For the eyes, click on
needed. Doing this helps prevent your
sculpts from looking bloated and gives you Append in the SubTool palette and choose
more control over your mesh. If you are Sphere3D. Using Deformation>Size, scale down the eye to
02
working on a higher subdivision level and fit the character’s head. Then, using Transpose Move, we
need to make large changes to the mesh, can position the eye in place. To create the other eye, simply
drop down to a lower subdivision level
(Shift+D) to enable smoother transitions use ZPlugin>SubTool Master>Mirror and choose Merge Into
and easier control. One SubTool. Once the eyes are in place, make adjustments
to the eyelids to fit the curve of the eye. For the teeth we can
append a Cube3D and resize it using the Transpose tools.
Under Geometry>DynaMesh, set the Resolution to a low
01 Having the SubTool black is easier to quickly
visualise and check the silhouette amount (such as 128) and create a U shape for the gums
and teeth. Turn off DynaMesh, increase the subdivision,
02 Concentrate on establishing the forms before
starting on the details then block in the teeth and gums using the Clay brush.
03 Once the teeth are sculpted, adjust their
Slowly refine the teeth with the DamStandard and Standard
03
position within the character’s mouth brush while increasing the subdivision level.

18 3D Art & Design

 Create a portrait

ěũũĥ1(2.-#1Ĵ.+8/(-3Ĵ
(,#+/2#ē,/ČĦ

Concept
We’re going to be making
an interpretation of Simon
Dominic’s concept. The
character has some strong
features that we’ll focus
on, like his jaw, brows and
substantial scars.

3D Art & Design 19

 Character

Make refinements
Create clothing and add details

04  Sculpt the scars In the concept this guy has some pretty gruesome scars. The
one over his left eye and upper lip are indented, so we’ll start by using the
DamStandard brush to create the initial line and flow of the scar. Next we’ll go back over the
04

line and include some irregularities, then by using small strokes and the ClayTubes brush we
can build up the scar tissue around the edges. The ClayTubes brush is great for these kinds
of details and we’ll use it again for the small raised scars on his head, by blocking in the
rough flow and shape of the scar. To add some more irregularities we’ll set the Intensity a
little lower. Using short strokes that alternate between Zadd and Zsub (holding Opt/Alt) we
can build up the grisly scar texture.

05  It’s all in the eyes I like to apply some Polypaint to the eyes before I finish the
sculpting phase, to bring some life into the character as I’m working on it. The
MatCap I’ll be using is the zbro_EyeReflection and can be downloaded from http://
luckilytip.blogspot.com.au. Select a base off-white tone and, with MRGB turned on, go to 04 Use DamStandard and 05
ClayTubes to create the scar
Color>Fill Object. Next, choose the Standard brush, turn off Zadd and only have RGB turned
on. We’ll select a brush size that fits the size of the iris and paint the darkest outer-edge of 05 Polypainting the eye with the
Standard brush
the eye. Next we’ll choose a lighter saturated tone and paint in the iris, giving a slight taper
to the lower half. This enhances a concave effect around the iris. Add some lighter flecks to 06 Use masking to pull out the
overlapping shapes
the iris, then include a black pupil as well as some subtle veins and colour variation to the
07 Position the SubTools then
sclera. To finish this element, we’ll paint in some fake occlusion to give the eyes some use DynaMesh to merge and
grounding in their sockets. subtract the shapes

06

06  Create the shirt To clothe our character we’ll

begin by appending a sphere and resizing it to
roughly fit. With the DynaMesh resolution low, we can use
07  Build the chain collar To shape the chain collar we’re going to use DynaMesh to
subtract and merge different meshes. Subtract one cylinder from another to create
the neck opening, then merge a cube and subtract another cube to produce the area where
the Move brush to block in the basic shape of the shirt. We the chain links in. To subtract one mesh from another in ZBrush, go to the SubTool palette,
can then begin to subdivide the mesh and build up details. move to DynaMesh (A) above the SubTool (B) you are going to use to subtract. The lower
For the collar area of the shirt, we can mask the areas where SubTool (B) needs to have the Difference SubTool icon selected, which is the second icon.
we want to create the appearance of an overlap and use the Select the top SubTool (A) and go to SubTool>Merge>MergeDown to combine both
Move brush to pull these areas up and over. Use the SubTools. To create the subtraction, hold down Cmd/Ctrl, then click and drag on the canvas
DamStandard and Standard brushes to refine the details. to re-DynaMesh.

07

20 3D Art & Design

 Create a portrait

Morph Targets and layers

The use of Morph Targets and layers can be extremely
beneficial to non-destructively add detail. For example, you
can store a Morph Target then start to sculpt scars and if
you aren’t happy with a particular area, use the Morph brush
to paint back to the stored version of the sculpt. Layers can
be helpful for controlling tiny details like pores. By placing
the pore details on their own layer, you can increase or
decrease the amount that they’re visible, giving you greater
control over them.

08  Add the chain links For the chains we’ll need to
append a Ring3D. Go to Deformation>Inflat to add
some thickness. By masking half of the ring we can use the
Transpose Move tool to drag half of the ring upwards to
extend the middle area. Next we’ll use DynaMesh to
reconstruct the topology. With the first link created we can
now duplicate and position a few more links using the
Transpose tools. This enables us to add a sense of
movement to the character.

09  Detail the accessories We can now do a

cleanup pass over our accessories and apply some
more detail. For instance, we’ll want to add some wear and
tear to the chain links. To do this we can use the Mallet Fast
brush which can be found in Lightbox>Brush>Mallet>Mallet
Fast.ZBP. Just use small strokes to add some dents and
scratches to the collar and chains. We can add some seams
to the shirt using masking with the Move and DamStandard
brushes, then also add some more folds to the shirt using
the Standard brush.

08

10

10 
09
Focus on the face With the sculpting phase
almost complete, it’s now time to add some finer
features. DamStandard is the main brush I use for applying
wrinkles and the smaller details. For the pores and stubble
texture, I use the Standard brush with a pore Alpha and the
Stroke set to Spray with a low Intensity. Try to keep in mind
places like the nose, which tend to have larger and more
visible pores. You can also use a mask for areas like the lips,
where you don’t want any pore detail.

08 Create the chain links and 10 Use a pore Alpha with the
position them using the Stroke set to Spray to achieve
Transform tools the pore details
09 Add a mask, use the Move
brush to define the seam,
then Standard and
DamStandard to refine

3D Art & Design 21

 Character

FiberMesh and Polypainting

Apply the texture and build the hair

11 12

11  Apply and adjust the hair FiberMesh works by applying hair strands to any masked
area of a SubTool. To create the head stubble, we’ll first need to create a 054
black-and-white image in Photoshop to be used as a mask. We can then mask the area of
the head we want to add hair to and subtract the 054 shape using the created Alpha. Click
the Preview button under the FiberMesh palette. Under Modifiers, change Length to 10 and
Coverage to roughly 70. Change the Width Profile to create thinner strands and the Base
and Tip colours should be a dark brown. Use the BPR Render (Shift+R) to test what this
looks like. Once you’re happy with the results, go to FiberMesh>Accept.

12  Shape sideburns and eyebrows Mask the sideburns area and select
Lightbox>Fibers under the FiberMesh menu. This will provide some presets to choose
from and edit. For the sideburns we can use the Fibers196 preset. Set Length to 14, Coverage
to 10 and adjust the Width Profile. Turn the texture off and set the Base and Tip Color to
brown again. Click Accept and then mask the area for the eyebrows. This time we’ll use the
Fiber160 preset and change Length to 110, Coverage to 2.25 and Gravity to -0.5. We’ll also
need to adjust the Width Profile, and change the Twist and Revolve to approximately 135.
Select dark brown for the colours and render to test the hair. When happy with the results
click Accept. Now we can use the Groom brushes to style the eyebrows a bit more and Masking tips
really get them looking rugged.
There are some masking tools in ZBrush that can be used to
help speed up the workflow and create some interesting
13 effects. Masking>Mask PeaksAndValleys can be used to
quickly create versatile patterns for texturing. In this case it
was used for Polypainting the chain. Masking>Mask By
Smoothness is also very useful for quickly painting edges
and dents, especially when working on hard surfaces. Play
around with the sliders to see what you can come up with.

13  Begin Polypainting For the skin I like to use the

zbro_Viewport_Skin2 MatCap, which can be
downloaded from http://luckilytip.blogspot.com.au. To
begin painting the skin, fill the SubTool with a neutral skin
tone as the base colour. Using a Standard brush, with just
RGB on and a blue selected, block in the beard area and
around the eyes. Add some red around the nose, lips and
cheeks, then a yellow tone on the forehead and slightly on
the cheeks. Using these more-saturated colours will give us
some nice variation and a good foundation to work from.
We can then begin to build up on the darker and lighter skin
tones and use a pore Alpha with the Stroke set to Spray to
11 Mask the head and apply the 12 Edit FiberMesh presets to 13 Layer on the colours and
FiberMesh options to fully create the sideburns and tones for the most-effective
add some more texture to the skin. Now we can slowly build
build up the hair rugged eyebrows skin tone up and refine the textures and values.

22 3D Art & Design

 Create a portrait

Finalise the sculpt

Add the last touches to bring out real personality

14  Final posing To finish we’ll need to add asymmetry

and pose the character. This can be done with
ZPlugin>Transpose Master>TPoseMesh. Then, using
14

masking and the Transpose tools, we can position the

character into an interesting position. In this case I’ve tried
to achieve a pose that would fit the concept artist’s
description of this character. When this is done, go to
Transpose Master>TPose-SubT to transfer the pose back to
the SubTools. We can now tweak the mean expression.

15

15  Set up the render When the sculpt is finished and

ready to be rendered, we’ll need to find a nice
camera angle and store its position in Document>ZAppLink
Properties>Cust1. For the lighting setup I like to use one
main light from the front and two rim lights from behind.
We’ll render these out separately to be composited together
in Photoshop. To create a rim light, simply click on the dot in
the Light Placement Preview window in the Light palette.
16  Make render passes Once the lighting is set up, we can begin to render out the
different passes. The starting render passes I use are the main BPR Render, Shadow,
the Rim Lighting passes, Ambient Occlusion and Depth. The next passes I render are the
This will send the light source behind the character. When Cavity, using the SketchShade2 MatCap; a Specular pass, using the Basic material and the
rendering the rim lights, turn off Polypaint for each SubTool Color set to black; then finally a Reflection pass for the chains, using a combination of the
and use the Basic material. I also set my colour to black. Reflected Map and Chrome materials.

16 17

17  Unify the bust Once we’ve rendered out our passes
it’s time to put everything together. I like to
experiment with the different layer blending modes in
14 Use Transpose Master to
pose the character Experimenting with MatCaps Photoshop to see which effects I can find. In this case my
Ambient Occlusion, Cavity and Shadow passes are all set to
15 Set up the main and rim lights
Don’t be afraid to experiment with layering different Multiply with the layer opacity lowered. For the Rim
16 The ZBrush render passes
MatCaps in your render passes. You can also edit existing Lighting, Reflection and Specular passes I’ve ended up using
17 Compositing all of the render MatCaps to suit the material you’re trying to render. Trying a combination of Color Dodge and Screen. Once you’re
passes together out different MatCaps and blend modes can lead to some
interesting effects and an overall more engaging render. happy with how the image is looking, you can make a final
Levels and colour correction to really make the image pop.

3D Art & Design 23

 Character

Easy-to-follow guides ěũ(-+ũ2*(-ũ,3#1(+ũ2ũ

take you from concept ũũũ!1#3#"ũ(-ũ343.1(+ũ
ěũ(-+ũũ!.,/
to the final render
ěũ43.1(+ũ2!1##-2'.32
Artist info

Thomas Lishman
Personal portfolio site
666ē3+(2',-ē!.ē4*
Country UK
Software used 142'
Expertiseũ'.,2ũ(2ũ-ũ
#7/#1(#-!#"ũ"(%(3+ũ13(23ũ
2/#!(+(2(-%ũ(-ũ2!4+/3(-%ēũ#ũ
!411#-3+8ũ6.1*2ũ2ũũ$1##+-!#ũ
2!4+/3.1ũ$.1ũ24!'ũ!+(#-32ũ2ũ

ũ.82ũ-"ũ !1+-#ũ.82

Render your
character’s skin
Learn how to use ZBrush to render your sculpts

R
endering in ZBrush has improved vastly since its
initial release, with users now able to create results
that can even match the results of any high-end
01 Set up your skin
shader Take
SkinShade 4 and click Copy
software that’s been designed specifically with rendering SH in the material loadout,
in mind. then open up a new
In this tutorial, I’m going to show you how to set up Double-Shader and paste SH
materials, lights and render settings that can provide into the S1 slot. Now for the
brilliant results with minimal time, using only ZBrush S1 slot; reduce the Ambient
and Photoshop. to 10, Diffuse to 60, Spec to 5
When creating a progress render for a client, I like to keep and also tighten the specular
a steady workflow. As such I use a preset file that I can curve. Add a value of just 0.2
import my sculpts into to instantly render export passes then Noise to the curve. Now take
composite within Photoshop. This keeps my entire workflow the ToyPlastic material and
within my main two programs: ZBrush and Photoshop. copy the S1 slot into the S2
Although it may take a few hours to set up initially, once slot of the Double-Shader
completed it provides a quick way to present your work, and that we are working with.
then make alterations afterwards. Reduce the specular of the S2
Unlike some methods, I prefer to render each light out slot to 5, and then increase
separately then comp those together so I have more control the Colorize Specular to 100.
over the mood and feel of the scene. As rendering in Change the spec in the
ZBrush is relatively much faster than external renderers, material to a pale, sky blue.
you can afford to do this. With practice you can create a For the wax settings, simply
range of lighting setups, and even save those out separately do: 30 Strength, 0 Spec,
01
and load those them in whenever they are required for use. 20 Radius.

24 3D Art & Design

 Render skin

02 Apply your Shaders Copy your Double-Shader

and paste it into a new slot, so you have Double-
Shader1. Increase the specular on both shader slots to 15. This
new shader will be for your high spec areas such as lips and
sweat. Once you have textured your sculpt, it’s time to set up
your materials. On each subtool for skin, get any brush and
make sure only M is enabled. Now go to Color>Fill Object with
the material you want selected, for example Toy Shader for
eyes and your Double-Shader for skin. You can then select
Double-Shader1 and paint on lip areas and wherever else your
model will be wet.

03 Render settings and canvas In the attached

screenshot you will find the settings I use for my
renders. I usually keep the AO resolution at half the Shadow
resolution to keep renders quick, but be sure to set your
resolution at roughly the same as your final output. Next you
need to set your canvas up. Go to Document and adjust the
Back colour to black, with Rate to 0. Below that, uncheck the
PRO button and adjust the sliders to your final render size, then
hit resize. Use Ctrl+N to clear the background, then drag your
02 03
model into view.

04 Lighting and passes I use a single light for each render, only enabling different ones
for rim and fill. My key light is set at 1.5 brightness with default settings. For my rim I
usually set a blue hue and to bring it to the back. I click once on the small preview sphere. I drag
and angle this so it creates a decent rim. Set this to about 3. I usually create a third, cool light for
fill, at about 1. I render each lighting pass out and export them via Document>Export Document.
Next export the AO and depth passes from Render>BPR Renderpasses.

04

ZBrush perspective
ZBrush doesn’t use true to life perspective; instead it’s
quite distorted. Use low levels to zero for sculpting. When
rendering, amp up your Angle of View to get the best
possible result. Below the floor button is a Local
Transformations button, which when enabled will rotate
the camera around a certain point on the model. Try
disabling this when rendering as it will snap to the world 05
grid centre instead, enabling you to find better angles to
view and render your scene.

05 Compose in Photoshop Open your passes in Photoshop, and copy and paste each
one into the first pass. Hide all but fill. Now unhide the rim and set it to screen. This will
overlay it over your fill layer. Adjust Opacity and Color however you desire. Next unhide your fill
layer and bring it down to a low opacity, and change the colour with Ctrl+U to a cool blue. I
usually experiment with levels, colour ranges and opacity to try and find something that looks
good. But if I don’t have the time, default settings are usually best.

06 Photoshop Composition 2 For your AO pass you can try a few things. I usually
increase the levels and set it to multiply to darken those cracks and shadow areas, but I
also change the colour with Ctrl+U then click Colourize. You can use the AO pass to create
ambient lighting. Set your unaltered layer to Hard Light and reduce the opacity to 5-15%. The
best thing is to play around and push each pass to add realism. I take the depth pass, Ctrl+Select
the RGB layer in channels and invert that selection to use on the main render. I then use Lens Blur
06
and also add some small noise to really achieve that depth of field.

3D Art & Design 25

 Character

Easy-to-follow guides
take you from concept
to the final render
Artist info

Chévon leo
Personal portfolio site
www.danroarty.com
Location Cape Town,
South Africa ěũ-ũ3'#ũ"(2!ũ8.4Ħ++ũăũ-"ũ++ũ
.$ũ'_5.-Ħ2ũ343.1(+ũ
Software used ZBrush screenshots, as well as
Expertise Chévon specialises 3'#ũăũ-+ũũăũ+#
in creating believable
facial expressions
Final render of the sculpted character

Sculpt intense
facial expressions
Sculpt a character, building highly emotive facial
expressions using the optimal tools in ZBrush

W
ithout expression, characters run the Look at yourself in the mirror and try to keep a and details in all individual areas of the face is key
risk of becoming boring. Giving them neutral expression. Now smile. Switch between the to producing a successful expression.
a good expression is vital, as it portrays two expressions and observe the changes that are Here we will be creating the character/bust from
personality and brings them to life. Throughout occurring. Notice that it’s not only the mouth and scratch, using ZSpheres to start on our base mesh.
these steps I will be demonstrating my own its immediate surrounding areas that are affected, We will then give the character an identity, pose
process for sculpting a character, focusing mainly but there are changes happening across the entire him and spend a lot of focus on the eye and mouth
on giving him a believable, yet characterised, facial face. For example, your nasolabial fold or laughing areas of the facial expression.
expression using the optimal tools in ZBrush. lines are more defined, your nostrils have widened, The final steps will be focused on refining the
Expressions play a huge role in everyday life. your eyes have narrowed and your ears have even model as well as adding supporting wrinkles, folds
When we look at somebody, we can tell what kind moved up slightly. and hair to further enhance the personality and
of mood they are in. We are even able to get a lot of Due to this observation, it’s safe to say that final look of the character. We will then export our
information about people without ever having to conveying expression in your artwork can be a character to Maya and render him using V-Ray.
speak to them. This is made possible by our natural fairly complex task. In order to simplify things, we This process shows the approach that I have
ability to recognise human expressions. need to break the face down into individual areas, found is best suited to my projects’ needs, however,
When sculpting expressions, believability is while still remembering to observe all the changes it is by no means the only way to achieve the same
imperative. One of the first steps to a believable in the face as a whole. Keep your mirror close by result. I believe that, as an artist, you have to find
expression is to look at the face as a single object. and grab a sheet of paper. While smiling, cover the the workflow that best suits your needs, so
So, grab a mirror, follow this little experiment and lower half of your face. You’re still able to identify experiment and I hope these tips point you towards
things should become a lot clearer. the expression, right? So, nailing the subtle changes some great results.

26 3D Art & Design

 Facial expressions

Give your character some personality

Sculpt a character with realistic facial expressions
Maya viewport screenshot ZBrush screenshot of
final sculpt

01

01 Start with ZSpheres ZSpheres can be used to create the basic structure of the
character. While this is not the only way to go about it, I prefer using them because they
provide a quick and flexible way of building a custom, unique base mesh. When working with
ZSpheres, I set the Draw Size to 1, because this makes it easier to manipulate them without
affecting any adjacent ZSpheres. I start by making a neck and head from the base sphere, and
02 also create a shape that represents the chest. I then continue by adding the torso and arms. Once
I am happy with the structure of my ZSpheres, I navigate to the Tool palette and click Make
Adaptive Skin.

02 Form and proportion Once the base mesh is complete, the next step is to capture
the general form and proportions of the character. Proportions play a vital role in the
believability of a character and their expressions. Whether you’re working on a photoreal or a
stylised character, your proportions need to be appealing to the eye. Having an amazing
expression and detail means nothing if your proportions are wrong, so spend some time on this
important step. Enable DynaMesh at a Resolution of 64, then tweak the proportions using the
Move brush, all the while adding and subtracting volume with the Clay Tubes brush. When
constructing forms, identify the bony landmarks on the body and use them as a guide.

03
03 Refine the forms Now that we have our general form set in stone, we are able to
increase the resolution and get to work on refining the character to give him a unique
identity. Use the Standard brush to define areas where bone and cartilage protrude, as well as the
Dam Standard brush for creating creases/folds and the Clay Tubes brush, with a low Intensity
(3-6), for volumes. During this stage, focus on the planes and structure of the face. Don’t go into
too much detail here, as we’ll get to this in a later stage, once our character has been posed. It’s a
good idea to retopologise your mesh now, as it makes posing easier and more efficient later on.

Anatomy
In order to achieve a believable humanoid character, complete with good expression and
personality, it’s beneficial to have some understanding of basic human anatomy. When
solely relying on reference, without any basic understanding, you could easily fall into the
trap of thinking your anatomy is correct, when it could perhaps be slightly faulty. You don’t
necessarily need to devote your life to understanding every bone and muscle in the body,
but understanding the forms and planes that these bones and muscles create can be key to
a believable sculpt. I encourage you to do some research on the subject as it will help you a
great deal on all further projects. There are many fantastic books available on the subject,
one of my favourites being Artistic Anatomy by Dr. Paul Richer.

3D Art & Design 27

 Character

Workflow efficiency
ZBrush is fully customisable and enables you to create a shortcut for almost all commands
offered. To apply custom shortcuts, hold Cmd/Ctrl+Opt/Alt and click on the function you
wish to create a shortcut for, followed by the key you want to allocate to the shortcut. I use
shortcuts for everything from brushes to creating polygroups. Try to create shortcuts for the
tools used most in your workflow. Another efficiency tip is to store a morph target before
sculpting and utilise the Morph brush as an eraser, rather than the Undo command, as it can
become sluggish when working on dense meshes.

04

04 Capture good reference Reference plays a vital

role in the sculpting process and with the invention of
the internet and Google, it’s just a mouse-click or two away.
However, because we have something very specific in mind in
terms of expression, a trusty old camera and a helpful model to
pose for you is more useful. Another way to obtain reference is
by observing yourself. Keep a mirror at your workstation and
constantly use it to analyse the creases, wrinkles and tension 05
areas that form expression. It may feel weird at first, but it’s one
of the quickest and most convenient ways to obtain reference.

05 Establish the pose Body language is important when portraying emotion. As such,
it’s beneficial to give the character a captivating pose. At this point in the creation
process you can turn off Symmetry. When posing a character, drop down to the lowest
subdivision available, then isolate and mask the polygroups of the parts that you need to pose.
Once you’ve properly masked the required body part, invert it and reveal the rest of the body.
You can then proceed to manipulate the part into the desired position using the Transpose tools.
During this process, be sure to observe the silhouette of the model from all angles by hitting the
V key on your keyboard. This switches to your secondary colour, which is black by default.

06
06 Block out the expression Now that the pose is properly established, use the Move
brush in conjunction with the Smooth brush to repair any errors that may have occurred
during the posing phase. Next, begin loosely blocking out the major influence areas of the
expression, such as the brows and general shape of the mouth.
07

07 Sculpt the eyes and nose As mentioned earlier, we will be isolating areas of the face
while working on the expression. This is to ensure we get as much emotion in the eye
region as possible, without being distracted by the rest of the face. If we are able to properly
capture the details of the expression in each isolated area, the overall expression should be easily
readable. I rely mainly on the Move brush to shape the expression. It’s best to tweak the brow
areas, raise the lower eyelids and focus on the areas directly around the eyeballs. Even the areas
under the eyes need to be given some form and definition. It’s also important to focus on the
area between the eyebrows, known as the glabella, to ensure that the skin is folding – or not, as
the case may be – as it would in real life. If you want to convey extreme detail, then focus on all
areas of the nose, including the supra-alar crease, alar-sidewall, columella and nostrils, ensuring
all these different parts of the nose are reacting to the changes in the face.

08 Build up the mouth The Move Topological brush is

useful for morphing the mouth into the desired shape,
which affects the mesh based on the topology, rather than the
08

overall area. This lets us manipulate the lower lip without

affecting the upper lip. Remember, you should be continuously
looking in the mirror as a source of reference. By using this
reference, you’ll see what the character is lacking, such as
some volume in the lips. To correct this, add volume using the
Inflate brush. When sculpting the mouth, focus on the bottom
half of the face as if it were a single object. Look at the tip of the
nose, the chin and the nasolabial folds. It’s vital to correctly
capture these areas, as they portray tension in the muscles. Pay
particular attention to the corners of the mouth and the shape
of the lips.

28 3D Art & Design

 Facial expressions

09 Polish things up In order to polish the model, increase the subdivision level and use
the Trim Dynamic brush, as opposed to the Smooth brush, which is far too destructive
on the forms we have spent so much time creating. It’s at this stage that we can also look at the
10

expression as a whole and focus on getting all of the elements to work well with one another. Any
changes we make at this stage should be incredibly subtle. To view the before and after meshes,
first store a morph target before sculpting. Then, by clicking the Switch button in the Morph
palette, you can view all the changes made and review their merit.

09

10 Add wrinkles and folds By utilising ZBrush’s layer

system to maintain control over the Intensity, you can
start adding supporting wrinkles to the face. Begin with the
larger folds, such as those below the chin, using the Dam
Standard brush with a large Draw size and a small Focal Shift to
give a more gradual, less harsh crease. To give the feel of
compression, gently go over the creases with the Inflate brush.
The clay 11
Add finer wrinkles and details wherever you deem necessary.
material While there are many Alphas available on the internet,
ZBrush’s default set of Alphas is more than sufficient. Use the
When creating the Standard brush with the Alpha set to 58, as well as a Radial Blur
material, I opted for a and Spray Stroke to apply finer details. Play with various
simplistic approach. I Alphas, as they can produce some interesting results.
created a VRayMtl,
setting the Reflection to
white, decreasing the
Glossiness and enabling
Fresnel Reflections.
11 Make some tweaks Before exporting the character to
your external 3D package of choice, make sure you
assess the model, make subtle changes where necessary and
Fresnel uses the viewing start work on the hair. Begin by creating an additional layer and
angle to determine the
amount of reflection that blocking out clumps of hair with the Clay Tubes brush. Refine
occurs on any given the hair using a low-intensity Rake brush, which mimics strands
material. This mimics of hair. Both these brushes are used in conjunction with the
real-life reflection Lazy Mouse to create clean strokes. Obviously hair does not
behaviour – if you would
like more information on react to expression, but a furrowed or straightened brow can
this, I encourage you to move the hairline. As for the eyebrows, I created a custom
search for BRDF online. Alpha and applied it with the Standard brush and the Drag
For my specific material, I Direction stroke.
wanted it to resemble
clay, so, I plugged a 12
fingerprint texture created
by Jeff Patton in as a
Bump map. I played
12 Set up the render We are now ready to bring our
character to life. In order to provide the scene with
some ambience, create two V-Ray Dome Lights. Light 1 affects
around with the Intensity the Diffuse attribute, while Light 2 affects the Reflection. This
until I found something
that I was satisfied with. ultimately provides us with additional control over the final
Because I hadn’t created render, because we are able to tweak the Reflection and
any UVs for my character, Diffuse amounts separately. Lighting is important when
I applied the texture as a creating expression, as it will ultimately add more definition to
projection based on my the character and the details you have created. Accompanying
render camera.
the V-Ray Dome Lights are three V-Ray Rectangular Lights.
Light 3 is the key light and Light 4 is the rim light. To calculate
GI, use an unbiased solution by setting both the Primary and
Secondary Bounce to Brute Force. Please note that this will
increase your render time significantly, but in my experience it
produces the most accurate results. By default, V-Ray also
produces too much GI, so lower the Bounce values to around
0.65. The final image is rendered out at 3,000 x 1,800 pixels as
a 32-bit OpenEXR.

3D Art & Design 29

 Character

Easy-to-follow guides
take you from concept
to the final render
Design & model
sci-fi characters
Artist info

Ben Erdt
Lukruk – The Thelarian Elite 2012
Here I’ll discuss how I created a highly complex and functional
Username: Benjamin Erdt armoured alien character, from concept through to final vision
Personal portfolio site Ben Erdt is a professional artist currently living in Amsterdam and working at Guerrilla Games
www.ben-erdt.de
Country Netherlands

L
Software used Maya, 3ds Max, ukruk is an armoured alien character I created as how to design and build complex characters with rigging and
MODO, ZBrush, MARI, NUKE, an exercise at Vancouver Film School. During this animation in mind, without compromising the original design
Photoshop
Expertise Concepting and process you’ll get an insider’s view of his creation from intent. I’ll also share my approach for keeping clean topology
producing both hard-surface the first spark of inspiration to the final render. By the end of for efficient subdivision, rigging, texturing and rendering.
and organic models
the steps you’ll have a full understanding of how Maya, This project was a great learning experience for me in
MODO, ZBrush, MARI, Photoshop and NUKE can be used to fields such as design, modelling, texturing, rigging, animation,
achieve the final result. rendering and VFX, as well as how to collaborate with other
Lukruk was a very complex project from both a design and disciplines. I hope the little nuggets of experience shared here
technical point of view, so here you can get a closer look at will help and inspire you with your own characters.

Make a concept
Develop who and what your character is

01 Turn inspiration into a concept I imagined a futuristic knight standing on a high

platform. Since he’s extra-terrestrial, I wondered what if an alien race – after their first
contact with mankind – became inspired by human history, especially by the stories about the
Knights Templar and the samurai. If they used their superior technology and created their own
01
group of protectors, what would a member of this group look like? I imagined them as characters
who are spiritual, intelligent and serene – like a group of contemplative listeners – but also as
powerful and unpredictable creatures. They have a monastic lifestyle defined by traditions and
strict rules. Only their strongest hatchlings are selected and Lukruk was one of the chosen. Find energy
and flow
02 Take the design in-depth As the head is the focal
point of the character, I wanted to make sure Lukruk
had a very expressive face, attitude and a great design. I wasn’t
A friend of mine advised
that when turning a
person into a stylised
happy about the first iteration, because he didn’t look as
cartoon character or
powerful as I wanted. I decided to make his head more angled caricature, he focuses on
and sharp along the jaw. The argus monitor lizard was a great capturing the energy,
reference for this. The elongated cartilage tips on Lukruk’s chin rather than the actual
further helped to emphasise the shape of the head. To counter shapes. By this he meant
the mental side, such as
the aggressive features of his design a little, as well as to add the essence of the person,
02
some conflict to the design, I gave him two small ears. as well as the character’s
personality and story.
From the start I was
03 Imagine the body and armour Lukruk’s armour was designed to be functional and
built for protection, but it’s also supposed to have a bit of an elegant and ceremonial
touch that comes through the curved lines and shapes. Because I imagined the model having lots
sure about the energy I
wanted this character to
have. The rather wavy
of detail, I figured a more humanoid silhouette would be better to read. When putting on the movement of lizards and
armour, the longer tail was supposed to be a mechanical extension that could be used as a the elegant motion of
weapon. As a reference for his default facial expression and attitude, I referred to heads of birds snakes can be used to
describe the feeling of the
of prey such as eagles, hawks and falcons. For the hard-surface parts of the armour I specifically form and lines Lukruk’s
collected images of sports motorcycles, CNC machine parts and landing gear. These were body is supposed to have.
03
incredibly useful reference points when approaching design as functionality.

30 O3D3DArtist
Art & Design
 Sci-fi characters

ěũũ("#.ũ343.1(+ũ.$ũ *(-%ũ
5(38ũ,/2ũ(-ũ ũ
ěũũ("#.ũ343.1(+ũ2'.6(-%ũ
3(/2ũ$.1ũ28,,#31(!+ũ
,."#++(-%ũ(-ũ 8
ěũũ("#.ũ343.1(+ũ.-ũ1(%%(-%ũ
8ũ
,(1ũ.-#-

Concept
Inspired by the unpredictable-looking
reptiles such as tegus and iguanas, the
spiritual feel of Gregorian chant, the armour
of the samurai, as well as the Knights
Templar, I decided to create Lukruk, an
armoured alien lizard knight.

3D Art & Design 31

 Character

Work on the model

Begin sculpting in Maya

04 Work on the head Starting out in Maya, I always begin modelling the head because
it’s the most important part of the figure and can determine the look and feel of the
character. For example, a bulky head with a fat neck would make me expect an appropriate body
to support it. I start with the eyes first, placing a polySphere at the position where the eyeball sits
and scaling it to fit. For shaping the eye socket, I usually start with a single polygon and work my
way around the eyeball. When beginning work on a model, start simple, as too many polygons
can make any fixes later on rather difficult and time-consuming.

04

05 Refine the head After blocking in the eyes, I move

over to the mouth, ears, lips, the jaw and the bony
structures around the head. The goal is to create the individual
facial features first and stitch them together at the end. When
they are separate, it’s easier to move them around and adjust
the head proportions. With the eyes filling the spaces in
between, you can already see what the proportions and
volumes are going to be. Above Lukruk’s eyes I created a row of
05
eyebrow-like scales, which help when conveying emotion.

06

06 Model the body When modelling the body I followed the same approach as for the
head. I started modelling, then tweaking very rough representations of the features to
build the correct silhouette, shape and volume. In order to manage the amount of detail in the
original concept art, I broke it down to primary, secondary and tertiary shapes using colour
coding. During the modelling process, I needed to determine the large and small shapes, as well
as where they join or overlap one another without constantly referring back the original concept
work. The chest was one of the more complex pieces, so I wanted it to fit as neatly as possible.
Since everything else was proportionally correct at the end, the chest was very easy to add in.

07 Efficient topology To avoid inefficient geometry, the topology was created to look
nice and smooth, with a maximum of two subdivisions, though most of the pieces
needed only one. The overall rule was to have complex forms and shapes with a decent amount
of polygons for each object. In effect this made life a lot easier for later unwrapping and texturing
07
all the pieces. Always try to stay as clean and efficient as you can and it’ll pay off later on.

32 3D Art & Design

 Sci-fi characters
08
08 The model versus the rig Making sure that Lukruk wasn’t only functional in design
but also in terms of articulation was a challenge, so I gave extra focus to areas that were
most visible and complex. One of the more complicated parts was the waist, as the tail, shield
flaps and legs all come together in this region. To avoid individual armour parts touching or
interpenetrating one another, I used a basic proxy rig while modelling to ensure any changes
made improved deformation as well as design. Instancing geo here also served to improve
iteration times. To make the tail believable and to maximise the range of movement, its inner core
was based on a row of single vertebrae inspired by the chameleon – one of the few lizards that
can completely roll their tails. This dynamic tail is key to Lukruk’s personality and body language.

09

10
09 Sculpt skin details ZBrush was used for detailing the head. As for reference, I chose
images of crocodile skin to create a smooth transition between the separate scales and skin.
Greek turtles have a noisy scale pattern on their neck that I used for inspiration when sculpting the
softer skin parts on the character’s face. I used the Clay brush to block out scales, then while sculpting
in the surface details I kept the design of his head in mind, making sure the details didn’t overwhelm
A clear end goal the overall aesthetic. The Slash2 brush was used to sculpt in some overlapping scales, while the Slash3
I believe it’s important to know the end purpose of your and Dam Standard helped to emphasise the crevices in between. The Inflat and Standard brushes add
creation. In this case the character was a production model more volume to the fleshier areas, while hPolish was used to break some of the hard edges of the
for a short movie. At the end we were going to need many scales. Once happy with the result, I rendered the Displacement map inside ZBrush. Here I chose an
high-res textures, as well as a complex mesh with lots of 8,000 pixel-sized map to have an opportunity for a higher-res detail pass in MARI.
individual pieces that were going to be rigged and
animated, but we also wanted to keep the render times as
short as possible. I thought about what could be done in
advance in order to stay efficient without compromising
quality. For example, adding simple colour shaders while
10 Develop the tones Before jumping into MARI I wanted to make a colour concept for
the head to get a better idea of the possibilities therein. I rendered out basic passes – AO,
Depth, Shadow and two Light passes – using the BPR Render in ZBrush, just as a base to paint on
modelling can help to get a better overview of the shapes in Photoshop. Once they were composited I started to build up the skin, while each group and
and details. It’s also good to sometimes step back and see if
the model still fits the concept art and has the same energy. layer represented a paint step for the skin. As a result I had a clean layer structure that was going
to be rebuilt in MARI to achieve the textured look.

3D Art & Design 33

 Character

Clean up the scene

Prepare and unwrap the geometry for the next stages

Complexity demands organisation

Lukruk has an extremely complex hierarchy. Counting meshes alone, around 1,800 different
objects needed to be named, grouped, organised, unwrapped and laid out efficiently in UV
space. A clean scene is vital to making your life and those of your colleagues easier. Clean
UVs within a good layout are half the texture job and even riggers can benefit. I decided to
unwrap everything in MODO because I was going to bake Cavity, Flat, Diffuse and RGB
maps in it. Switching applications also brought a bit of variation into this tedious process as I
like using a Cintiq when working inside MODO. When it comes to polygon-modelling,
especially, this is one of the most artistic methods I have found.

11 Sort and name the pieces In order keep the amount of potential issues low later on, it’s
important to keep a clean scene. I went through the scene hierarchy and deleted nodes
such as empty groups, duplicates, unused nodes and so on, to make sure there was no excess
left that could cause problems at a later stage. All geo objects were sorted according to the
design hierarchy of the character. For example, everything in the left arm was contained in a
Group node. Individual objects were then grouped separately according to their visibility, size and
shape hierarchy. Each group represented one quadrant in UV space, which was the foundation
for the UV-layout and texturing process.

11

12 Unwrap all the pieces I exported the pieces as

individual OBJ files for UV-generation, plus the MTL file
to keep the assigned colour shaders inside MODO for later
texture-baking. The tools I predominantly used were UV
Unwrap, Relax UV and UV Peeler. Unwrap and Relax are
helpful for generating and refining UVs of shapes that are more
organic and curved. The UV Peeler is a fast tool to unwrap
pipes and hoses. For simpler shapes that are more cubic,
spherical or cylindrical, I used the UV Projection tool as a
12
starting point, as it offers appropriate projection types. The
screenshot shows the finished UVs for a part of the foot.

Collaboration is key
This project was a shared effort between several
people, with Amir Ronen performing rigging and
secondary animation, Colin Giles completing
primary animation and myself responsible for
design, modelling and texturing. We shared and
referenced the main rig as a master file that was
frequently updated with rigging information by
Amir. When Colin, the animator, started blocking
out the animation, I could work on the textures
and shaders. In order to stay on course, we had
regular meetings to discuss our progress, the
next steps in the work, and to discuss any
current issues that might have arisen.
Supplied with your free resources, you’ll find
several video tutorials from Amir (www.
amirronen.com). Who generously provides an
in-depth look at his rigging workflow.

34 3D Art & Design

 Sci-fi characters

Prepare textures
Enhance and refine the character Artist
Showcase
Ben Erdt
I’m a professional character artist working at Guerrilla
Games on next-gen game characters. I’ve been
fascinated by sci-fi universes and movie creatures
since I was a kid. I trained myself in CG art so that I
could make my own.

13
Batty ZBrush, Photoshop (2013)
This is the kind of sculpt I do after work or over the weekend. In
13 Cavity/Convexity maps in MODO Since I was going to texture the actual high-res
character mesh, I used Cavity maps as a guide for painting textures. These were baked
inside MODO using the Occlusion shader. They also helped to bring back some detail into the
this case I wanted to sculpt a bat-like creature based on a drawing
I did some time ago.

Diffuse texture. In MODO I added the Occlusion shader to the Shader Tree and set it to
Concavity with a Max Distance of 60cm and a Spread Angle of 45 degrees. I created another
Render Output, set it to Diffuse Coefficient and made sure the Diffuse Amount inside the base
material was at 100%. The desired texture resolution for the particular object was set inside the
MODO Render Properties. Remember, baking these maps in MODO using the Occlusion shader
takes some time. On your disc you’ll find a step-by-step video guide for baking Cavity maps.

Collector 3ds Max, ZBrush, MODO, Photoshop (2011)

One of the game characters I created for practise. I blocked out
the body in ZBrush, modelled armour in 3ds Max, rigged in Maya
and textured in Photoshop.

14

14 Begin authoring textures After unwrapping the entire character model, I ended up
with almost 50 sets of four to five textures, so it was important to maintain a consistent
workflow. The hidden Plain Diffuse layer was used as a selection guide. The Dirt Effects group
slightly broke up the flat colour among all the textures and added overall dirt and grime to the
character. The MARI group consisted of object-specific layers that were painted in MARI 1.4, Deep Sea Creature XSI, ZBrush, MARI, MODO, Photoshop
(2012/2013)
although this workflow was improved thanks to MARI 2’s full PSD support. Additionally, to make I wanted to texture the deep-sea creature model that appears in
the final character design easier to read, I made his feet more muddy, while reducing this effect my demo reel. It was modelled in XSI, sculpted in ZBrush, textured
higher up his body. This resulted in a light colour gradient from bottom to top. I also focused in MARI and then unwrapped and rendered in MODO.
more on areas that look worn-out and used in order to add a bit more history to his armour.

3D Art & Design 35

 Character

Render and post-production

Finish with tiny details and complete render passes
15
15 Include dirt and
paint chips
Alongside some hand-
painting, I also used
photomanipulation for
texturing Lukruk – something
that increased his sense of
believability. A great source
for photo textures is
www.cgtextures.com. In
order to add paint chips to the
character, I chose a texture
with a high contrast between
the paint and the areas where
it had faded. I created a
selection using Color Range
and used it as a mask on a
solid colour layer. This
enabled more control if I
needed to change the hue,
saturation or brightness. Back
in MODO I then baked an
additional Cavity map with a
lower range, as well as a
Convexity map that was used
to emphasise the edges in the
roughness texture.

16
16 Render out passes The screenshot shows the
passes that an average shot required. My friend and
colleague Andrew Paxson (www.andrewpaxson.com) was in
charge of compositing and adding VFX inside NUKE and he did
an amazing job. I wanted to give Andrew as much control as
possible over the shots, so to give us the opportunity to tweak
the blue parts of the body armour, I took the flat colour textures
that I baked before texturing and created simple RGB maps.
The result was a pass in which R represented the metal types,
G represented pieces that could cause potential artefacts and
noise, while B represented all the blue parts of the armour.
Staying clean and organised like this from the very beginning
saved us a lot of time during lighting, rendering and
composition. There was one case during the first render test in
which I had a bad polygon causing the Subdivision
Approximation to crash mental ray. However, because the
character model had an elegantly structured hierarchy, the bad
piece could be tracked down and quickly fixed.

Final words
Building a complex character such as Lukruk can seem
intimidating, but when you break it down and organise it
into many simple tasks it will feel much more achievable.
Be careful to never lose sight of the original design intent,
no matter how challenging the task is. For instance, I could
only texture Lukruk by splitting him into 50 textures, but I
still made sure that all of those textures were consistent
and re-enforced the original design concept.

36 3D Art & Design

 Gallery

Equilibrium was
created with the idea
that mechanical parts
could form into a living,
breathing life form. I
tried to convey the
innocence and
vulnerability of a young
girl in the piece
Daniel Arnold-Mist
Equilibrium, 2012

Artist info

Daniel Arnold-Mist
Daniel started creating 3D
images when he was just 13
years old
Personal portfolio site
www.liquidminduk.cgsociety.org
Country China
Software used Maya, 3ds Max,
V-Ray

Work in progress…

3D Art & Design 37

 Character

Easy-to-follow guides

Mesh shatter
take you from concept
to the final render
Artist info

Christopher Velez
your
Polyman
characters 2013
Username: Polydude

Personal portfolio site

www.nethub.cghub.com
Country LA, United States
The goal was to create an abstract character constructed from
Software used Maya,
V-Ray, Photoshop
various pieces of geometry. The image illustrates vulnerability
Expertise Photoreal and Christopher Velez spends his time improving his skills as a graphic designer. He also
stylised character creation
performs theatrical acting on the side

I
have always enjoyed looking at sculptures, In this tutorial I will take you through a character, exposing different layers underneath.
especially puzzle and junk sculptures that step-by-step process for creating a mesh We continue by manipulating some of the
are assembled from many parts. This shattered surface using my Polyman character. geometry on our character to help convey
inspired me to find a quick solution to achieving You’ll learn different concepts from mesh randomness and imperfection. We will then
something similar in 3D. So, I am excited to share shattering a surface to a few post-production tips focus on the vacuum effect that is happening on
with you a technique that I use in some of my and tricks. First we’ll discuss how topology will Polyman’s back, by manipulating floating pieces
abstract digital sculptures, which I like to refer to influence the surface of your model. Then we’ll of geometry. Finally, we cover the rendering and
as mesh shattering. cover the peeling away of certain areas of our post-production process.

Concept
For the concept I have chosen
a pose that I think gives a
sense of balance and weight to
the scene. I also wanted it to
add a sense of vulnerability to
the character.

38 O3D3DArtist
Art & Design
Mesh shatter characters

ěũũ!#-#ũăũ+#2ũ$.1ũ.+8,-ũ
ěũ43.1(+ũ2!1##-2'.32
ěũũ6.ũ5("#.ũ343.1(+2ũ
2'.6(-%ũ3'#ũ6.1*Ąũ.6ũ
(-ũ 8Ĕũ 7Ĕũ-"ũ
+#-"#1

3D Art & Design 39

 Character

Getting started
Focus on topology flow

01 Observe the geometry We begin by taking a close look at how topology can affect
the overall appearance of your character. In the top left-hand portion of the
accompanying image, you’ll notice a general face mesh with edge loops defining the facial
features: eyes, nose and mouth. On the top right-hand side, we have another example but with
edge loops that have simple cross sections. Also notice the triangle shape I added on the
forehead to demonstrate how topology will transfer over in the end. The bottom row is the result
after using the topology technique, revealing how edge loops define the flow of geometry.

02 Model setup and detach edges Let’s jump right into making a mesh shattered
model. For this tutorial we’ll be using the Polyman scene that’s provided on the disc.
Import the scene file called Start into Maya, and hide the platform on a separate layer so the only
thing visible is the base mesh. Now switch to component mode and select all the edges that
make up the model. In your modelling menu click on Edit Mesh>Detach Component to detach all
the selected edges, then Mesh>Separate, forcing each face to split apart.

02

01

Adding extra surface detail

While detaching and performing face extrusions, I suggest your model polycount be just
under 7k. You’ll notice that performing an extrusion on thousands of separate pieces of
geometry can be computationally expensive and may even crash Maya. However there are
always workaround solutions. If you want to use a model with a higher density such as 7k or
above, just split the model into smaller parts and tackle each section one at a time. For extra
surface detail, layering a denser model on top of a lower density model will add a cool
randomised look. We’ll cover randomising fully later in the tutorial.

03 Add thickness Now that we have all of the faces separated, we are able to focus on
adding thickness to each face for an improved appearance. Adding thickness creates a
better visual sense of angle and direction. Start by selecting the individual faces that make up our
model and then select Extrude from the Edit Mesh menu. Once the extrude manipulator
appears, select the standard Scale manipulator and begin to scale inwards until the desired
thickness is achieved. There are undoubtedly other ways of adding thickness, but this method
03
often delivers the most interesting results in my opinion.

40 3D Art & Design

 Mesh shatter characters

Mesh shatter and peeling

Start adding the details to your model

04 Scale for a shatter effect Select the entire model

again by going to Modify menu>Center Pivot. This
ensures each individual piece of geometry is scaled locally. Select
and scale the entire model outwards and immediately notice the
results; the topology begins to shatter. We need to find a balance
between the shatter effect and the model’s facial features, so set
the scale XYZ to a value of 1.35 in the channel box.

05 Randomise the surface This is where the fun

begins. We start by taking a creative approach
towards how the surface of the model will look. In the
screenshot provided I began working on the model’s left upper
arm. Select several pieces of geometry with the Lasso tool, then
scale the selection down so it starts to look like floating cubes.
Set the scale XYZ to a value of .68 in the channel box. The goal
here is to create a sense of wear and tear on certain parts of the
04
body. Try scaling your selection in the opposite direction too.

05 06

06 Add Layers Let’s continue with the model’s left arm

as we start adding layers to our damaged section.
Duplicate some of the cubes made earlier and move them out
so they cover up parts of the damaged section. Scale and rotate
the new geometry so they have randomness; repeat these
steps on other areas where you’ve added wear and tear. This
helps create a sense of depth and complexity to the surface.

07 Create a peeling effect Add peeling to make it

seem as if the character is being pulled into a vacuum
above. Start on the back, where we will have the most damage.
Select one half of the back with the Lasso tool, then rotate the
selection away from the centre of the back. Do the same for the
other half of the back so you have a pried-open look.

More on randomising
For Polyman’s face, I preserved the features by scaling
down some of the geometry to add clarity to the eyes,
nose, and lips. I also deleted geometry that was obtrusive
to the model while rearranging other pieces around. By
adding a sense of randomness, your model helps break the
CG look of appearing too perfect. No matter how abstract
your art can be, adding a bit of imperfection will make your
07 scene look more aesthetically pleasing to the viewer.

3D Art & Design 41

 Character

Finalising the model

Add some finishing touches to the scene

08 09

08 Create a vacuum effect Now we’re almost done with our model, it’s time to add
floating pieces of geometry above Polyman’s back. Start by selecting some of the
geometry we peeled back in the previous step and make duplicates. Move the new geometry up
09 Fill empty spaces To finish, we’ll cave in the centre
of our model’s back by selecting the back’s centre and
moving the various pieces down inside the model. Next, scale
so they float just above the back of our model. Do the same for the other side. Continue building your selection down so that it’s smaller, creating the
up on the floating pieces so that you form a pyramid shape. Make sure you randomise the appearance of floating debris inside the hole. You can apply this
floating pieces by scaling, moving, and rotating each piece. to other hollow areas of your model.

10 11

10 Generate floating debris Now we will focus on the final part of the modelling
process. Add some debris to the scene by selecting areas of the model that are
damaged. In this case we’ll focus on the arm again. Select a few random pieces and duplicate
them. Begin to move the pieces around the damaged area and make sure you randomise the
rotation of each piece of debris. Remember that rotating debris in the direction of the vacuum will
help give the image a better sense of direction and momentum.

11 Mesh shatter the platform As for the platform upon which Polyman sits, I added a
bevel to the edges and performed the same steps covered in the first part of the tutorial. I
scaled down the geometry so the pieces appear evenly spaced like tiles. I also added a cube
inside the pieces of geometry to make the platform appear solid. The platform was made small
so that it accommodates Polyman’s body, but it still adds to the scene’s sense of vulnerability.

42 3D Art & Design

 Mesh shatter characters

Showcase
Render and post
Set up your render and head into post Artist
Christopher Velez
12 Set up your lights Now that the model is complete, we can start setting up the lights
for the scene. I used V-Ray for this scene but you can replicate it with a similar setup in
other render engines. I added two area lights; one in the front and a rim light in the back. The first
I was born in Brooklyn, New York, but in 2009 I moved
to California. Art has always been a passion of mine. I
light has a light-grey colour with an intensity value of 20. The second light has a sky-blue colour channel my creative energies into digital art where I
with an intensity value of 16. I tend to keep the lights at a far distance, as I find that adding have limitless freedom to manifest my ideas into
distance to my lights yield better results for the shading in my scene. reality. I’m a freelance digital artist with a focus on
character modelling and I also do theatrical work as
an actor on the side.

Married to Technology
Maya (2013)
This particular character
is based in the future,
where cybernetic
modifications are
mainstream. In this
narrative, she has
a cybernetic
implant addiction.

12

Persian Princess
13 Material settings
As for the materials, I
wanted the surface to
13
Maya (2013)
This simple face study
shows that Christopher’s
resemble something fragile work isn’t just based on
yet sturdy, such as pottery. abstract works, but also
The same applies to the tiles on realistic and
believable models.
on the platform. Assigning a
new material may cause a bit
of memory lag depending on
how much geometry is in the
scene. Set the material colour
to a light grey and Roughness
Amount to .675. Set the
reflection to Blinn and the
Reflect Color to dark grey.
The Use Fresnel checkbox
should be checked and the
Lock Fresnel IOR… should be
unchecked. Change the
Fresnel IOR slider to Z Moments Maya (2013)
something high like 1.9. The Z Moments are a series of sculpts based on characters from the
same settings can be used for popular animated series, Dragon Ball.
the platform material too.

3D Art & Design 43

 Character

Head into postproduction

Touch up your model for a clean final result

14 Render settings Let’s take a look at the primary settings I use for my renders. I have
provided detailed screenshots of my settings with the disc. Start by selecting the V-Ray
tab. Sampler type is set to Adaptive DMC; Antialiasing is set to Lanczos for sharper details;
Adaptive DMC is set to 1 min and 8 max; Threshold is .005. Next we move onto the Indirect
Illumination tab. Turn on Ambient Occlusion; set the Primary bounce to Irradiance Map and
Secondary Bounce to Light Cache. I didn’t use any render passes, but V-Ray makes it very easy
to set up render passes if you choose to use them.

15 Final touch ups Head into Photoshop for some final touch ups. First add some Motion
Blur to the floating debris. Use the Lasso tool to make selections around several floating
pieces of debris, then apply Filter>Blur>Motion Blur. Set the angle to -81 for an almost vertical
blur and a Distance to 17 pixels. I like to randomise the blur in the image; some floating parts
were blurred and some were not, just to give the appearance of random velocities. For more on
touch ups, see the boxout at the bottom of the page.

15

3hours
creation time
Resolution:
14 2,000 x 1,254

More on touch ups

Here are some extra steps I take in Photoshop. I like to
pump up the contrast a bit by going to Image>
Adjustments>Shadows and Highlights, and set the
Shadows Amount down to 17% and the Adjustments
Midtone to +5. I add a bump map to the overall image by
duplicating my main image, going to Filter>Other>
Highpass and setting the Radius to 4 Pixels, then changing
Layer Type of the Highpass layer to Soft Light. Reduce the
Opacity of the Highpass Layer down to 30%.

44 3D Art & Design

 Gallery

Artist info

Adam Martinakis
Adam has been working and
This was created experimenting with CG and
visual media for 13 years
during experiments Personal portfolio site
fragmenting a 3D model www.adamakis.blogspot.co.uk
Country UK
of a body using several Software used 3ds Max,
ZBrush, DAZ3D, V-Ray
different kinds of Work in progress…
topology. I chose a way
that seemed to have an
interesting effect and
painted the pieces in
three different colours to
present a story
Adam Martinakis Fatal, 2013
3D Art & Design 45
 Architectural
visualisation
Create stunning arch-vis artwork from
exteriors to interiors
48 Build a wooden interior 63
Techniques and resources behind
establishing a neo-classical interior

52 Build C4D arch-vis assets

Model a simple antique chair in just a
few simple steps

56 Make atmospheric renders

Create a romantic atmosphere, with
a cool blue sky and warm lights

63 Gallery
Butterfly by
Thiago Queiroz Lima

64 Add post-production touches

Use After Effects to composite scenes
and add post-production touches

67 Gallery
Contemporary Living Room by
Jonathan Beals

64 52

46 3D Art & Design

48

56

67

3D Art & Design 47

 Architectural visualisation

ěũ43.1(+ũ2!1##-2'.32

Easy-to-follow guides
take you from concept Build a wooden
interior
to the final render
Artist info

Discover all of the techniques and resources behind

Christopher Velez establishing a simple neo-classical interior

I
t’s important to have an efficient production-ready arise. Typically professionals can expect to go from a
Username: Polydude
workflow right from the start in order to build concept sketch to a completed image in less than one day.
Personal portfolio site convincing timber in an interior scene. Initially, we will start with a reference image, which is a
www.nethub.cghub.com
During the course of this tutorial we’ll discuss the core part of the arch-vis process. Reference can be found,
Country LA, United States
Software used Maya, processes and helpful tips for modelling, texturing, lighting, taken or drawn, but in an ideal world it should reflect the
V-Ray, Photoshop rendering and post-production. As always, the intention is to colour, style and composition of the intended outcome. This
Expertise Photoreal and create an image in a manner that’s timely and production- is possibly the most important tool for a 3D artists when
stylised character creation
proven, enabling changes along the way should the need trying to achieve photorealism; without a real-life reference

48 3D Art & Design

 Wooden interior

Making a scene
Use a real-life reference to create an arch-vis interior
Simple geometry coupled
with uncomplicated lighting
can often yield great results

it’s hard to understand the nuances of materials and how

colours might blend together.
We’ll be aiming towards a wealthy neo-classical interior,
drawing reference from an image created on behalf of The techniques used
property tycoon Candy & Candy (see below). This image is in the creation of this
image are both
particularly warm and utilises a one-point-perspective simple and
production-ready
composition – something that’s very common in
architectural imagery owing to its strong and simple lines.
So as to experiment with timber materials effectively,
we’ll be using a custom-designed and hand-modelled
armoire cabinet. This features a panelled and curved front,
enabling all manner of configurations. The curved face is
equally useful for achieving graduated lighting and to further
accentuate reflections.
Detailed high-resolution textures are key here, so we
recommend shopping for purpose-made imagery (supplied
with Bump or Specular maps) from the likes of Arroway,
TRU Textures or Turbosquid (www.arroway.de, www.
texturesrus.net and www.turbosquid.com respectively). 02
These will make a big difference and exceed anything just
found randomly.
Another option is www.vray-materials.de, which offers
free preconfigured shaders. However, while this method is
02 Model the individual elements of your scene
The scene has been designed to accentuate the
qualities of timber and so will incorporate a custom-designed
convenient, it affords little understanding of the principles armoire cabinet, complete with a curved front element to create
behind creating shaders. As such, in these steps we’ll be interesting responses to light cast across it. First sketch this and
constructing all our shaders manually in V-Ray, utilising then simply draw it in AutoCAD for dimensional accuracy. Now
various raw texture maps along the way. export a DWG to 3ds Max and begin modelling. The front door
The technical requirements behind this tutorial are elements are created by modelling in a flat plane, then merging
relatively standardised: 3ds Max, V-Ray and Photoshop. the pieces into one editable poly object. Apply a Subdivision
There should be no vast hardware demand, providing the modifier and finally an FFD 4 x 4 x 4 modifier so as to bend the
setup is correct, so an i5 processor with 4GB of RAM is object to shape. This is perhaps unorthodox, but it gives a very
more than sufficient. Also, the principles discussed are quick and believable result.
largely universal and can no doubt be applied to any 3D
package and renderer. 01

01 Find reference images and make sketches

When creating a photorealistic architectural image, it is
paramount that you first find a suitable reference. This should
roughly reflect the intended style, colouring and composition of
your desired outcome and will provide valuable insight into how 01 Find a decent reference
certain materials and lights should look when combined in a image and then combine it
real-life scenario. It is also important to have a rough with simple sketches to set
the groundwork
composition in mind before beginning any image. A simple
sketch is more than sufficient as a guide to work from. Lack of 02 Loosely draw the cabinet in
direction at this stage will stifle the entire image and cause no AutoCAD, model it in 3ds
Max, then apply Subdivision
end of frustration. Put simply, don’t begin until you are nearly and FFD modifiers to bend it
certain of this step’s completion. © Candy and Candy 2013
to shape

3D Art & Design 49

 Architectural visualisation

05

03

04

06
Geometry vs textures
A vast amount can be achieved using textures, from
reflectivity-mapping to full geometric-displacement,
which negates the need for overly complicated models.
Should absolute detail be the name of the game, there
may come a time when modelling individual planks in a
timber floor is the best route, but detail should be relative
to time and distance; don’t spend hours modelling minute
details that can’t be seen or appreciated by the viewer.
07

03 Model the scene With the cabinet done, begin

assembling the elements required for a complete
scene. Start by establishing a VRayPhysicalCamera and
and set the Reflective Glossiness to around 0.85 (higher will
result in more reflectivity). Now add the Bump texture into the
Bump channel and set the value to around 3-5. We won’t be
positioning it in relation to the walls, floor and cabinet. using a Reflectivity map in this case, as the desired material
Remember to always model loosely at first, blocking in larger outcome is very smooth (varnished). Now apply a UVW Map
elements and refining details as the scene develops. For ornate modifier to scale the texture.
profile details, such as the skirting and wall panelling, we can’t
recommend the free Sweep Profile script highly enough. This
brings several useful architectural profiles to your fingertips at
the click of a button (www.3d-kstudio.com). The flower model
06 Add a timber shader Repeat this process for the
timber floor element, combining Diffuse, Specular
and Bump maps in the appropriate channels so as to create a
used can be purchased from CGAxis (www.cgaxis.com). convincing shader. Here we’ve also used a ColourCorrect layer
between the Diffuse map and the texture, enabling the hue, 03 Set a VRay PhysicalCamera,
block in modelled elements
04 Engage V-Ray Now move to V-Ray by selecting it
as the Production Renderer from the Render Settings
dialog box. Subsequently open the Materials palette (keyboard
saturation, brightness and contrast to be tweaked flexibly. This
isn’t essential, but provides an extra layer of control that’s
particularly useful for those who prefer to minimise any time
and then gradually refine
details as the direction of the
scene becomes more clear
shortcut: M) and add a new VRayMtl. Apply this default spent in post-production. Apply a UVW Map modifier again to
04 Activate V-Ray and apply a
material to all objects in the scene. This should act as a clean scale the texture appropriately.
standard grey material to all
base on which to begin constructing shaders and ensures that the scene objects
no mental ray or Standard materials will confuse the rendering
process. 3ds Max’s Compact Material Editor can be used as
opposed to the newer Slate Material Editor, simply on the basis
07 Complete the textures Repeat the process of
creating shaders for any remaining scene elements,
then focus on adding believability where possible by including
05 Find suitable textures and
apply them as maps in a
new VRayMtl
of its ease of use. features such as subtle reflections (every real-life object
reflects light to some extent, so ensure your 3D objects follow 06 Create a new shader for the
timber floor, using Diffuse,

05 Make the cabinet shader Find an appropriate

wood texture, grouped ideally as Diffusion, Specular
and Bump maps. Once textures have been sourced, create a
suit). Not all textures (such as the wallpaper) require detailed
Specular maps, although these can be made by hand with a
degree of Photoshop knowledge. 3ds Max automatically blurs 07
Specular and Bump maps

Repeat the whole process for

the rest of the elements and
new VRayMtl slot in the Material Editor and name this all the textures so as to avoid problems when rendering reduce the Blur values in
appropriately. Apply the Diffuse texture as a map within the animations. To retain sharpness, open up each bitmap channel order to bring sharpness to
Diffuse slot, enable Fresnel Reflections (true-to-life reflections) and set the Blur value to 0.01 (as opposed to 1). all the textures

50 3D Art & Design

 Wooden interior

Simple lighting 11 Patch things up

When in Photoshop,
continue to refine the raw
There are no doubt thousands of ways to light a scene render by correcting errors in
using 3ds Max and V-Ray, from High Dynamic Range the Patches group. This is
Imagery (HDRI) to IES lights – but the key is simplicity. used for chopping out
Understand how your scene should look. Should the lights
elements that simply don’t
be focused? Should they be coloured? Should they appear
in reflections? All of the best images can typically be look right, such as the image
created with less than five lights. and reflections in the photo
frame. We’ve also applied a
crop to the image at this
08
08 Light the scene Once again speed and simplicity
are king, so begin to light the scene using strategically
placed VRayLights (set to Plane mode). Here we’ve only used
stage, having decided upon a
more effective composition.
This is quite simply a black
two lights, one of which is placed directly above the scene’s layer hiding elements that are
elements and another at 45 degrees from the camera, angled no longer required. Rendering
slightly down so as to cast a shadow as if from a nearby at higher resolutions can be
window. The window’s light has a modified Directional value so very helpful once you arrive at
as to cast a harder and more-precise shadow. This creates a this step.
stronger overall feel when compared to soft shadows. Balance
the light’s Power values and tweak the VRayPhysicalCamera’s
Exposure settings where required. 12 Make final
adjustments Now
begin to make large

09 Prepare for rendering Open the Render Settings

dialog box and set the Render parameters as per your
preference. Here we’ve used a rather typical Light Cache/
colour-correction
adjustments, taking direction
from your original reference
Irradiance Map setup, combined with Catmull Rom antialiasing image. Try to match the
and Reinhard colour-mapping, set with a Burn value of 0.85. tonality and materials as best
09
Now add key post-production components to the Render as possible, then
Elements tab, namely VRayWireColor, VRayShadows, subsequently utilise the
VRaySpecular and VRayReflection passes. Set an appropriate VRayWireColor pass to
10 resolution, hit Render and save each pass as a full-resolution select elements and begin to
TIFF or JPEG. add adjustment layers to suit.
Once you’re totally satisfied

10 Begin post-production work in Photoshop

With all the rendering (and inevitable tweaks)
complete, open each of your passes in Photoshop and create a
with the outcome, you can
add more effects to the
Global group, such as
sensible layering/grouping convention. Here we’ve produced a vignettes and overall colour/
stack with crops and helpers at the top, filtering down to global contrast layer adjustments.
effects, colour-corrections, patches, passes and renders at the
bottom. Doing this at the start of every image will inevitably 12
save masses of time and help you identify different stages of
post-production. Now apply passes to the base render to
enhance aspects as you please. Screening the VRayReflection
pass will accentuate highlights, for example.

11

08 Use simple lighting and change directional parameters for

stronger shadows

09 Set up the render and add in key post-production elements as

render passes

10 Open the rendered passes in Photoshop and organise the layer/

group structure

11 Patch up surplus elements and experiment with some cropping to

enhance the result

12 Apply some final colour adjustment layers and tweak the effects in
your Global group

3D Art & Design 51

 Architectural visualisation

ěũ(-+ũ ."#+ē!Č"
ěũ 
ũČũıũ("#.ũ343.1(+
ěũ43.1(+ũ2!1##-2'.32

This is the final render of the antique

chair created by Gustavo

Build C4D arch-vis

assets with NURBS
Learn how to model a simple antique chair in a few simple steps,
Easy-to-follow guides using splines, Sweep NURBS and Subdivision Surface objects
take you from concept

I
to the final render n the following tutorial you’ll discover a simple method for Once we have the approximate shape set up, we can
modelling an antique chair in CINEMA 4D. There are many create cuts around the edges using the Knife tool, increasing
Artist info

overly complex ways of modelling a simple chair like this, but the low subdivision model using a Subdivision Surface object.
after running a few experiments I found a far simpler method, Don’t forget to create these cuts around the low-poly model,
which I’ll share with you. The process that I detail here will because if you import the model into a Subdivision Surface
impart many useful and time-saving approaches in only a few without making these necessary cuts, you’ll effectively lose
simple steps. the model that you want to produce. You can also add more
Usually when we start using CINEMA 4D we can select edges using the Bevel tool.
from many different ways to achieve a similar result. Usually The process of capable arch-vis modelling is one that you
Gustavo Åhlén
we would start with a simple cube and use the Extrude option should really focus on, because mastering and memorising
Personal portfolio site
www.gustavoahlen.com to create the frame of the chair, then its other parts. However, these techniques will open up a huge range of possibilities for
Software used Cinema 4D, this takes more time than is needed. you when it comes to finding customers in new areas of
ZBrush The process I’ve found particularly helpful in the creation of work. In my personal experience, hotels, shopping centres,
Expertise Founder & Creative
Director at Enginetion.
the frame of the chair are using Sweep NURBS and splines to museums and more might buy your 3D furniture designs and
Professional 3D/VFX designer, create the shape. The splines can be applied as rails so that a send them to be manufactured. Any good idea could be
Matte painter for films, games geometric shape such as a circle or rectangle can follow the approved and used in a variety of places. So, follow the
and advertising
shape of the backrest, armrests and legs. It’s important that process carefully and consider how what you learn could be
we can get an approximate shape that we can then manually applied to other projects. Be sure to check out the video
adapt to the final model. tutorial with the disc!

52 3D Art & Design

 Build C4D arch-vis assets

Model an arch-vis chair

Use NURBS to create an antique chair

The frame of the chair

forms the essential shape
of the overall model

This reference image can be used to better

understand the different perspectives of the model

01

01 Shape the frame First we need to create the frame

of the chair using splines and Sweep NURBS. Don’t
forget to use a Frontal view for this. Go to Draw a Freehand
03 Work on
proportions
Select the Top view, then
Spline and select Bezier. Now you need to add the first point in draw a spline using Bezier as
the middle of the X axis and create the other points as you can in the attached image. Use
see in the attached image. Use Adaptive Interpolation Points Adaptive as the
for this. Once you have established the points using the Frontal intermediate points. Now
view, change the view to Left, then drag and drop the points as add a rectangle spline
in the example vertex in the accompanying image. (W=40cm H=40cm) using
the Frontal view. Add Sweep

02 Add Sweep NURBS Next, go to the panel where

you have selected the splines and select Rectangle
(W=42; H=35). Add Sweep NURBS to the scene. Using this
NURBS, drag and drop the
spline previously drawn and
the rectangle spline into
02 method we can convert the chair frame into polygons. Now Sweep NURBS. Now apply
hold left-click on the Subdivision Surface icon (the green icon) the Sweep NURBS with the
and you’ll see a dropdown menu where you need to select splines into a new Symmetry
Sweep from a panel of objects. Once you have added Sweep object. Alter the Tolerance
you need to drag and drop the spline previously made in the until you get a good weld
first step and the rectangle (spline) into it. You should get between the parts on each
something that looks like the accompanying image. side of the X axis.

Recommended types of splines

In this particular project we should use Type=Bezier and Intermediate Points=None. If you try
different types of Intermediate Points you’ll notice that other options will increase the number
of polygons. We should keep the number of polygons as low as possible, so we can
manipulate our piece without problems or make any changes easily. If you use other types of
intermediate points, you’ll increase the number of polygons and the manipulation will be
impossible. This is why I prefer to use the Subdivision Surface object to increase the number
03 of polygons in the final step.

3D Art & Design 53

 Architectural visualisation

05

04 06

04 Model the chair legs Add a cylinder (Radius=10; Height=353; Height segments=4)
and position it below the seat rim previously made, using the accompanying image as
reference. Make this cylinder Editable and, using Extrude and Bevel, get a similar shape to that of
the image. The process used here is better explained in the accompanying video tutorial. Once
you’ve finished the legs, import this piece into a Subdivision Surface object. By applying Bevel you
can add more edges near other edges to keep the original shape without getting a deformation.
Add a Symmetry object to create two legs.

05 Add bevels to the frame Convert the seat frame created in Step 3 to Editable then,
using Edge mode, go to Select>Loop Selection, pick the edges of the seat frame and
deselect the ends (holding Ctrl). Now you can just drag and drop to get a good bevel. Next, using
Loop Selection and Polygon mode, you need to extrude the shape around the surface while
trying to get a low relief.
07

06 Create the chair cushion Now we need to create a spline around the seat frame. I
used the same process as in Step 3, but in this case I used Mirror to duplicate the
selection to the other side of the X axis with the following parameters for the Mirror: Co-ordinate
08

System=World, Mirror Plane=ZY. Once you get the separated splines with the same shape on
each side of the X axis, you need to join the vertex using Join Segment, which will enable you to
close the spline. Add Extrude NURBS, then drag and drop the spline into it. Set Extrude NURBS
Movement (0; 40cm; 0), but don’t use caps.

07 Close up the holes Now convert the previous Extrude NURBS to Editable mode and
use Close Polygon Hole. You also need to close the upper hole (see the accompanying
image). If you close the bottom hole you’ll notice how the Subdivision Surface object added in the
next step will deform the base. You can close the bottom hole with Subdivision Surface
unchecked, then add a bevel in the bottom edges to create two edges. This means that the
Subdivision Surface object can’t deform the cushion.

08 Complete the cushion Add another Subdivision Surface object, then add it to the
Extrude NURBS that we created earlier. Doing this will round the surface. However,
before heading to the next step, uncheck Subdivision Surface. Now, using Loop Selection and
Edge mode, you need to add a bevel. This way the Subdivision Surface object can’t deform the
model. Switch to Polygon mode and select the polygon that you have created using Close
Polygon Hole and move it upwards. Activate a Subdivision Surface object, then use Top view to
help follow the shape of the seat frame.

54 3D Art & Design

 Build C4D arch-vis assets
09 10
09 Model the backrest Using the same process as
for the chair cushion, we need to create a spline
(Type=Bezier, Intermediate Points=None), following the
backrest shape as in the current image, and add an Extrude
NURBS object. Add the spline previously made into it. Use
Extrude NURBS Movement (0; 0; 60cm), don’t use caps and
convert this object to be Editable.

10 Bevel and adjust Close the polygon holes (both the

back and front). Using Loop Selection and Edge mode,
select the edges around the rest and add a bevel around it. It’s
important to keep the original shape during this stage. Apply a
Subdivision Surface object, then add the Extrude NURBS
created previously into it. You’ll likely need to adapt the points
over the backrest, so when you activate the Subdivision Surface
object it will retain the desired shape.

11 Work on the armrest To complete the armrest

element you need to use the same process previously
used for the Sweep NURBS objects. Create a spline following
11 the shape of the armrest, then add a Rectangle spline (W=36;
H=28). Apply another Sweep NURBS and add in all the splines
you previously made. Try to adapt the spline following the
current image. Now add a Subdivision Surface object (Sub.
level=2) to increase the subdivision level of Sweep NURBS.

12 Final details and render Once you’ve completed all

of the previous steps, you can start experimenting with
the ornamental details, as you can see in the current image.
These will bring out greater realism in the final result. By using
Sweep NURBS and Cloner, we can attach new decorations over
the mesh. Try searching for reference for new ideas, and
perhaps try out some alternative ideas to those displayed here.
Experiment! Finish by applying materials and rendering.

Final rendering and

lighting of the scene
In the final rendered scene you can create studio lighting
with a simple open rectangle and two lights in front of the
chair, as well as a camera to focus on the main object.
After this you need to add Ambient Occlusion and Global
Illumination (Render Settings>Effects). Now you need to
assign the materials to each part of the chair. I used a list of
three materials that you can find in CINEMA 4D
(Presets>Prime> Materials>Misc). Studio lighting can be
achieved in a variety of ways, however, so experiment to
find the best results.
12

3D Art & Design 55

 Architectural visualisation

Easy-to-follow guides
take you from concept Make
atmospheric
to the final render
Artist info

Sérgio Merêces

Username: smereces
renders
Gardenian House 2013
Personal portfolio site
www.sergiomereces.com
Country Portugal I wanted to create a romantic atmosphere, with the cool blue of the
Software used 3ds Max, V-Ray,
Forest Pack, Photoshop sky contrasting with the warmth of the lights and resulting in a mix of
Expertise 3D visualisation
production, specifically exterior beautiful colours, making for a stunning hidden sanctuary
and interior renders
Sérgio Merêces is a professional 3D artist who provides 3D visualisations to clients all around the world

56 3D Art & Design

 Atmospheric renders

H
ere I’ll reveal my project workflow and the
steps you need to take to achieve the best
results in your arch-vis renders. I’ll explain the
Concept
workflow and techniques that I use daily when creating For this project – and for all my projects in general – I always
take the time to search for references for my main idea, as ěũ#%#33(.-ũ,."#+2
visualisations. I will also cover important steps such as how ěũ'(-#2#ũ+-3#1-ũ,."#+
this will help with the process of modelling, creating
to set up the lights and environment of the scene using materials and understanding lighting. ěũ1##ũ,."#+
HDRI maps and V-Ray lights. I will also explain the best way ěũ43.1(+ũ2!1##-2'.32
to create textures and materials for the assets, plus the tools
required to create the vegetation.
Finally I’ll explain the post-production process,
particularly how to edit the RAW render image in Photoshop
using tools such as curves, corrections, adjustments and
filters to enhance the final render.

3D Art & Design 57

 Architectural visualisation

Modelling, texturing and materials

Set up your assets for the scene composition

01 Exterior and
interior modelling
We will start by creating the
floor and ceiling of the house
from boxes, then using Edit
Poly to get the right form by
moving the vertices and
adding a chamfer to the
edges using the Quad
Chamfer script available from
www.mariussilaghi.com.
Next we’ll model the window
frames and glass from a box
and with the Edit Poly tool
we’ll inset the two front faces
of the box, then bridge them
to get the main frame form.
After this we need to copy
01
the boxes and adjust them,
subtly moving the vertices.
Repeat this step on a larger 02
scale for the interior walls.

02 Create the
terrain To create
the terrain and design of the
water paths, draw them with
the 3ds Max Freeform tool.
This way we can achieve the
exact look that we want and it
can be completed relatively
quickly. After drawing all the
terrain, convert it to Edit Poly,
then Extrude all the faces to
set the terrain height. Finally,
to add more detail into the
render, we need to apply a
chamfer to the main edges
and then apply the
MeshSmooth tool. For the
03 Textures and materials For the house interior we’ll
create base materials using the VrayMtl, then create
the wood for the walls, floor and ceiling, concrete for the base
Modelling with Ribbon Tools
Foreign Terrain we need to and entrance stairs and metal and glass for the frames. For the 3ds Max comes with many useful tools to facilitate the
use a plane and then convert terrain, create a simple VrayMtl with a grass texture. Also use a modelling process. To create water I use the Freeform and
Strips tools to design the terrain by drawing it in Freehand
it to Edit Poly and model it VrayMtl for the water. To create ripples and waves in the water mode and then create all the elevations with the Push/Pull
with the Freeform Push tool, it’s better to use a Bump material as it creates a more realistic tool. To fix the geometry wireframe I use the retopology
painting all the terrain effect than using VrayDisplacement, which takes much longer tools, allowing me to create new faces quickly. Throughout
elevations. See the to render. However, there are cases where using the modelling process I use tools such as Swift Loop to
create a quick edge loop; the Qslice tool to slice parts of the
accompanying image for VrayDisplacement creates a better effect – use whatever suits models; and the Cut tool to cut model edges.
more detail. your scene and needs best.

03

58 3D Art & Design

 Atmospheric renders

Vegetation and Objects Scattering

Create vegetation and scatter objects using the Forest Pack plug-in

04 Grow the small grass For the terrain we must

create a small grass model and spread it. To create
the small grass group model one blade and, using the Advance
Painter script, copy it to create a small group of grass. Now it’s
time to spread the grass model over the entire terrain. To do
this create a new Forest Pack Object and choose the model of
grass that we just created as the Scatter Object. Use the
Distribute map and Transform settings to define how the grass
will be spread. Check the accompanying image for settings.
04

05 Add water lilies In this example I used several

models of water plants from the Forest Pack preset,
and one from Evermotion Archmodels 124. Create a new Forest
Pack Object and this time, as we want to spread more than one
Object, we need to create five new layers using Add New Item.
After you’ve chosen all the items you wish to be spread, we
need to let Forest Pack know which models to spread more
than others by using the Probability option. By default this is set
to the maximum value of 100%. We need to decrease this
value for the models that we want to have less of a presence in
the scattering. You can use the Paint tool too but it’s a more
time-consuming process.
05

06 Place leaves on the water We will now model

one leaf from a normal plane with four length
segments and seven width segments. Convert it to Edit Poly
and then move the vertices to make a realistic leaf shape. Next,
create a New Forest Pack Object, choosing the leaves as a
Scatter Object and the water plane as a Distribution Object. In
this scene I want the leaves more dense in some areas than in
others, so I will use the Paint tool to do it manually, painting in
the dense areas and erasing in areas where I want less leaves.

06
07 Background bushes and grass For the bushes I
added a model that I already owned and for the grass I
used a lawn grass from a package. Create a new Forest Pack
Object and two separate layers – one for the bushes and the
other for the lawn grass. Next, we need to set the Probability
with a higher value for the lawn grass and a lower value for the
bushes, then set the Transformations to get random values for
scale and rotation. This creates a nice mix between them and
will work perfectly for the background. As the background will
be relatively far from the camera there’s no need to worry
about detail here. Ultimately it’s about the overall effect.

07

3D Art & Design 59

 Architectural visualisation

Add finer details

Create and place the lanterns and interior objects

08 Model the trees I like to use SpeedTree to model my trees. Begin by creating the
main trunk in SpeedTree, followed by the branches. To create an authentic shape we
need to adjust the Skin and Shape parameters. For the branches we’ll need to create two or three
levels of branches to get a more realistic look and to increase the detail for the leaves. The next
step is to create the leaves that will be spread along the last branches level. Do this by creating a
node for the leaves and then creating a new material with the leaves textures and alpha to be
used by the leaves mesh. Finally we need to adjust the density and position of the leaves by
tweaking the Frequency and Placement until you have achieved the desired look.

08

09 Create the Chinese lanterns I used Forest Pack

once again to spread the Chinese lanterns across the
water. I tend use Forest Pack quite often when I need to spread
3D models across a scene because it has the advantage of
making the viewport lighter and can be used even with a huge
number of models. Create a new Forest Pack Object and
choose the Chinese lanterns as Scatter Object and the water as
the Distribution Object, then, using the Paint tool, start to paint
09
the lanterns randomly across the water.

10 Work on interior objects For the house’s interior

objects we will create a new 3ds Max file, merging only
the house so we can manage and compose the interior easily,
and test some compositions until we’re satisfied with the result.
Next we need to copy and merge them into the 3ds Max scene
file using a script called Copitor, available from www.scriptspot.
com/3ds-max/scripts/copitor. With this script you can
transfer objects from one file to another in two clicks.

Leaves in the wind

For the leaves in the wind I created a Particle Flow event,
where I chose for the leaves to be spread by the particles. I
created a wind and gravity effect to influence the spread
and to give it a more realistic effect. I adjusted the Particle
Flow parameters so I could control and manipulate the
scene to look the way I like. Afterwards, I converted the
10 leaves into a mesh using Compound Objects>Mesher.

Sérgio Merêces
Sérgio is CEO and lead 3D artist at Merêces
Arch-Viz, an award-winning 3D
visualisation studio based in Évora,
Lisbon. The studio’s work can be seen
globally in many industries and markets.
The team applies the most advanced Silent House 3ds Max, V-Ray, Forest Pack, Hotel 3ds Max, V-Ray, Forest Pack, Photoshop
techniques in all its works, always striving Photoshop (2013) (2013)
to make unique works of art. You can learn This is a personal project. I wanted to experiment with This is one part of a four-part project for a client. I
more at www.new.sergiomereces.com the V-Ray 3.0 Beta and the Forest Pack plug-in. wanted to set it in dramatic atmosphere.

60 3D Art & Design

 Atmospheric renders

Lighting and environment

Set up the scene’s lights and environment

11 Lighting setup For this scene I used only one main light
for the entire scene illumination, VrayLight’s planes for
the house interior and spherical VrayLights for the Chinese
12 Light the environment For the environment we will be using the HDRI map that we
used for the main illumination. In this case we’ll use the HDRI map for both the light and
the environment. Place a copy of the HDRI map into the 3ds Max Environment Map Slot, then to
lanterns. These lights have a warm temperature to create a nice ease the adjustment process, press Alt+B to open the viewport configuration and set it to Use
contrast between the cool exterior blue of the environment and Environment Background. Using this we can now adjust the background and compose our
the warm yellow colours of the interior. For the main light environment to find the settings that best suit your scene.
create a VrayDome light with a HDRI map – in this case we’ll
use one from www.3docean.net/user/paguthrie. To adjust the 12
light use V-Ray RT and a material override to test and adjust as
necessary. In my case I needed to adjust the HDRI map
settings, as you can see in the accompanying image.

11

13 Render settings For the render settings it’s important

to balance the best results possible with a relatively fast
render time. Personally I like to run a few tests so I can optimise
my render settings, which I prefer to be no more than 12 hours
maximum. My final render resolutions are usually between
4,000px and 6,000px to achieve a nice level of detail and
definition. This particular image was rendered with 5,500x and
300 DPI. You can view the various render settings I used in the
accompanying screenshot. Once you’re done, we’re all set up to
crack on with the post-production in Photoshop.

13

V-Ray RT
To refine your composition I really recommend using V-Ray RT to preview and adjust the
scene illumination. V-Ray RT is a brilliant feature which enables you to adjust the lighting in
Creative Hub 3ds Max, V-Ray, Forest Pack, Photoshop real-time, giving you full control of it and letting you test many possibilities until you’ve found
(2012) the desired illumination for your scene. I use V-Ray RT to create and preview my scene
A 3D visualisation of a creative hub building, designed for an
materials too, and to see how they will interact with the scene illumination. This is an
architectural project.
important step in attaining the best and most realistic look from the materials.

3D Art & Design 61

 Architectural visualisation

Post-production
Bring your render to life in Photoshop

14

14 Import and organise render passes Open the

RAW render file in Photoshop, then open all the render
passes into the RAW render project as layers with the Place
15
tool. To make things easier we’ll create several folders. Create a
folder with the name Masks, where we can put all the masks
that we’ll use to quick select parts of the image. Create a
second folder called Corrections, where we’ll put corrections of
the RAW render image, such as any little mistakes that crop up
in the render and can be easily fixed in post-production. The
third folder we need to make should be called Adjustments,
where we will put our colour adjustments. Finally, create a
folder called Effects where we can place image effects.

15 Corrections In all renders there are small accidental

mistakes that crop up from rendering the scene. In most
cases these are little things that we can’t actually see before we
render the image in full resolution. Thanks to post-production, 16

when they do crop up they can be easily fixed. In this image

there are a few bad lighting calculations and texture errors. To
fix these, create a new Layer in the Corrections folder and use
the Healing Brush Tool, Patch Tool and Clone Stamp Tool to get
16 Adjustments For the adjustments we should create a new Exposure adjustment layer
to fix some of the gamma, as in this instance the RAW render image came out a bit
darker than expected. We also want more light in the shaded areas to create a nice balance in the
rid of as many of the flaws as possible. final image. Next, create a new Curves adjustment layer and adjust the Blue Channel to maintain
a nice balance between the blues and reds. We then need to add a new Levels adjustment layer
to edit the overall levels of the image to create a slightly more dramatic tone.
Sharpen the result
One of the last things I do in post-production is to sharpen
the final image by creating a detail pass. To do this, go to
17 Effects and final adjustments We’ll do the final adjustments in the overall colours of
the image by painting some yellow tones in. Add a layer in the Effects folder and set it to
Overlay, then paint with yellow and orange tones to make a really nice balance between the
Filters>Others>High Pass. This will open a window where yellow lights and the blue colour of the environment. The final touch will be to add some
you can control the strength of the sharpen with the radius
amount. After putting in the desired amount, press OK and environment effects such as fog to give some depth to the image. Use the VrayZDepth pass in
set this layer to Soft Light. Screen mode and add a Levels adjustment layer. On this layer we need to create a Clipping Mask
to control the fog depth and intensity of the VrayZDepth pass.

17

62 3D Art & Design

 Gallery

Artist info

Thiago Queiroz Lima

Username: thilima3d
Personal portfolio site
www.thilima3d.wordpress.com
Country Brazil
Software used
3ds Max, V-Ray, Photoshop

Work in progress…

I was really tired of seeing the same kind of 3D

architectural images: box houses in the middle of
trees or lofts with the same furniture. I thought up
a solution, using the butterfly as my main concept.
It was just thinking outside the box!
Thiago Queiroz Lima Butterfly, 2013
3D Art & Design 63
 Architectural visualisation

ěũũ ũ,/2Ĕũ!.413#28ũ.$ũ
Ċ!#-ũ-"ũ
7(2
ũũũ43.1(+ũăũ+#2Ė
ěũ43.1(+ũ2!1##-2'.32

Add post-
production touches Use After Effects to composite scenes and add
post-production touches to an arch-vis image
Easy-to-follow guides

H
take you from concept ere we are going to cover post-production methods After Effects to spruce up the areas where we want more or
to the final render for an arch-vis image using After Effects. I will start by less illumination. Together with the exposure of the effect, this
explaining how to prepare the 3D scene in 3ds Max and technique can be used to correct the gamma on the image. I’ll
Artist info

then use V-Ray in order to render high-quality passes. These then show you how to use the Z-Depth pass to define an
important render passes will be used later during the environment, create a fog effect and give depth to images.
post-production stage. In After Effects I’ll show you how to Finally, I’ll run through the final touches for bringing the image
approach your project, organise layers, streamline your work its ultimate look, using the Magic Bullet Suite plug-ins, Mojo
and develop the image. and Looks.
With the project set up and organised, I can start with the
necessary steps to apply post work to the image, 01
Sérgio Merêces
demonstrating how you can assist the visual impact and
Personal portfolio site
ũ666ē2#1%(.,#1#!#2ē!., achieve realistic results in post-production. One of the first
Location (2 .-Ĕũ.134%+ steps is to fix any image errors that we can see in the render,
Software used Ċ"2ũ 7Ĕũ142'Ĕũ whether they’re there because we couldn’t see them at the
/##"1##Ĕũı8Ĕũ 4+3(!33#1Ĕũ
'.3.2'./Ĕũ
$3#1ũĂũ#!32
time of rendering, or due to not having time to correct them in
Expertiseũ#-ũ8#12ũ.$ũ the 3D application. Remember, fixing errors in post-production
#7/#1(#-!#ũ(-ũ1!'(3#!341+ũ is always easier.
5(24+(23(.-Ĕũ(-!+4"(-%ũ"#2(%-(-%ũ
-"ũ,."#++(-%ũ.1%-(!ũĊũ Once we’ve taken care of the composition and have an idea
!'1!3#12ũ$.1ũ5("#.%,#2 for the 2D elements that we want to add, we can move to the
lighting adjustments, using blending modes and painting in

64 3D Art & Design

 Post-production
02

01 Render in 3ds Max For more control over your

post-production process, it’s important that you render
out the passes required to assist you. At this stage I always
03 2D elements The
integration of 2D
elements is key to create a
select the render passes shown in the screenshot. The render credible atmosphere. Having
pass Render ID is the most important because we can make formed the idea of the
selections or masks quickly and accurately with it. It’s composition in your head,
important to always save your images in 32-bit format. In my make a quick composite in
projects I always save my render passes in TGA 32-bits or Photoshop to test if it works.
32-bit EXR, because then I can fully control the exposure and I’d make three or more tests
lighting adjustments in post. of different compositions to
see what works best for the

02 Set up in After Effects When you have your

render and all your render passes done, go to After
Effects and create a new project with the resolution you want
idea. Once you have the
desired layout, save the 2D
elements in PNG format to
for your image – in this case we’re using 3,000 x 1,688 pixels. take them into After Effects
Because this project will be a static image, we’ll set the and create a new composition
Animation Time in Duration to 0. Change the Project Depth in to place them all. We will
Project Color Depth from 8 to 32bpc, then import the render apply this composition as a
passes into the project and apply the Color pass. We’ll layer in the project. Apply a
duplicate it and use the Alpha pass in the new layer to remove small correction in the sky,
the background. This means we can add a new sky or make creating a mask in the
some changes to it. We now have the basis on which we can desired zone, as you can see
03
develop the rest of the post-production work. in the screenshot.

04 Use contrast &

gamma This step
is rather important because
here we will fix the lighting
and contrast of the image.
Create a new solid layer and
paint on it with the Blend
Mode set to Overlay. Apply
white all over the areas you
want to be brighter and use
black in those areas you want
darker. Then add the
Exposure effect and set the
parameters shown in the
screenshot. These settings
will reduce the saturation of
04
the image and subtly lighten
the overall feel.

3D Art & Design 65

 Architectural visualisation
05
Light with HDRI maps

I recommend all those who wish to create sunset-lit

images to use HDRI maps. I use these in all my sunset,
sunrise or night-based scenes. I always illuminate my
scenes with HDRI maps because it’s simply the most
realistic way to get good results at the lighting stage. In this
project I’ve only used one 3ds Max V-Ray Dome light with
an HDRI map applied to it. This enables me to manage all
the lighting with the VRayPhysicalCamera parameters.
06
There are many very good HDRI maps to be found online
and you can also make your own, of course. Check out
your disc for free HDRI maps from CGAxis and 3DOcean
(not available for digital readers).

06 Final touches To finish, use the Magic Bullet Suite’s

Mojo and Looks tools in order to add the finishing

05 Use environment
effects In order to
create the whole atmosphere
touches to the image. Create a new composition in the project
named ‘Composition1’, cut and paste all the project layers into it
and apply it as a layer. Now we will add the effects in this layer,
and give depth to the scene, as you can see in the screenshot. Apply Mojo to the layer and
we start by adding the adjust the parameters in order to give it more contrast. Also
Z-Depth pass that was apply Looks to the same layer, using the Looks interface to
rendered in 3ds Max on my begin building up effects. You can use the Contrast adjustment
composition as a layer, then to correct minor issues, while Spot Exposure will illuminate the
add the Invert effect to bottom of the image in order to give it more light. A vignette
reverse the channel. Setting and Chromatic Aberration will provide a realistic and
the blending mode to Screen photographic effect.
maintains the foggy aesthetic.
This kind of effect can be
used to make cool
atmospheric results like fog,
dust, smoke and so on. Now
we can apply the Exposure
effect on the Z-Depth layer,
then control the intensity and
depth of the fog effect by
adjusting the parameters. The
next step is to add the Color
Balance effect on the Color
layer to adjust the parameters
and give a hint of a sunset to
the overall image. Now create
a new layer with the Color
Dodge blending mode and
paint areas where you want a
stronger tone and intensity of
yellowish light. This gives the
image much more colour
consistency with the sunset
environment. Experiment and
adjust parameters to fully
optimise the effect.

66 3D Art & Design

 Gallery

Artist info
Jonathan Beals

Username: thebeals
Personal portfolio site
www.jdbeals.com
Country USA
Software used 3ds Max, iRay,
mental ray, Photoshop, After
Effects

Work in progress…

Since entering the VFX industry, I’ve been

fascinated with artists who can mimic reality
and create photorealistic art. I came across the
‘Foo House’ series by Apollo Architecture and
immediately fell in love with the contrast
between the concrete and wood and how the
light played upon each
Jonathan Beals Contemporary Living Room, 2013
3D Art & Design 67
 Photorealism
If you want to make your digital art as realistic
looking as possible, you’re in the right place
70 Make a realistic human portrait 77
From the initial concept, see how to
make a portrait look like a photo

77 Gallery
Then Flying The Youth by
Leng Shang-Peng

78 Design a stunning dress

Use Marvelous Designer to create
beautiful images of flowing fabric

84 Create hair and fur

Make a truly photoreal and captivating
image with FiberMesh

91 Gallery
Photorealistic CGI by
Set Visions Imaging Studios

92 Bring still-life objects to life

Construct a realistic still-life scene using
photorealistic techniques

78

84

91

68 3D Art & Design

92

3D Art & Design 69

 Photorealism

Concept
I didn’t use any concepts
but rather relied on
real-world reference that I
gathered as inspiration
and visual aid to create
my final artwork.

ěũ43.1(+ũ2!1##-2'.32

70 3D Art & Design

 Realistic human portrait
Easy-to-follow guides
take you from concept
to the final render

Make a realistic
Artist info

Dan Roarty

Username: droarty
human portrait
Personal portfolio site
www.danroarty.com
Location Redwood City, CA, USA
Freckles in a Blanket 2013
Software used Maya, Mudbox,
V-Ray, Shave and a Haircut, This image is a realistic portrait of a girl wrapping herself in a
Photoshop, Knald
Expertise An artist specialising in blanket, looking into the camera at the viewer
realistic 3D heads and portraits
Dan Roarty is a lead-character artist at Crystal Dynamics/Square Enix

O
ver the next few pages I’ll outline the texturing within Mudbox using stencils and also show you how I created the eyes and the
steps and techniques I used to create projections. I’ll introduce a free piece of scanning texture used. When we have everything modelled
this image. First, we’ll gather some software I used on my phone for capturing 3D and textured we’ll create realistic V-Ray shaders
reference and define the project to ensure a data and using it in the final image. We’ll cover for the skin, hair and eyes, and I’ll show you the
quicker and more organised workflow. We’ll cover creating and styling hair using Shave and a Haircut lighting setup and render settings I used for the
modelling in Maya, as well as sculpting and and the workflow for converting to Maya Hair. I’ll final piece.

Define your concept

The level of planning you do can make or break your project
01
01 Decide on an idea When working on their own
projects, some 3D artists have a tendency to create a
concept on the fly, rather than spending quality time deciding
02 Gather reference I gathered a lot of online
references for hair colours, facial expressions and
specific poses of the face that looked appealing. I didn’t intend
or defining exactly what the final image will look like. Without a for my model to fully match the references I found, but set out
clear goal, it’s impossible to know what assets need to be to use them as inspiration and a general guideline to follow.
created and when. At times this can lead to incomplete Another big reference and help for me was my wife. I took a
projects and/or an unrefined final image due to lack of few photos of her looking into the camera as a reference of the
planning. With this in mind, I decided to create a realistic general look and angle that I wanted to achieve. It was also very
red-headed female wrapped in a red blanket and looking into helpful to see how the blanket would look in the picture and
the camera. Thanks to this process I knew which assets I how I wanted to make it.
needed, such as a face, hair, eyelashes and blanket. So, with the
idea defined, let’s gather some reference in order to better 02
guide us through the process.

03 Block out meshes and UVs With your reference gathered, start blocking out the
base meshes in Maya prior to sculpting. I had a basic head mesh that I had previously
created, which I am able to reuse when creating a new project. I spent time ensuring that the
head reacted well to sculpting and also deformed properly when adding facial expressions. Now
lay out the UVs to ensure no overlapping and proper distribution of UV space among the
polygons. You can test how the UVs respond by applying a basic checker texture in Maya. We
can worry about creating the blanket later on, as we’ll be creating it using a 3D scanning
approach. With your head mesh ready with UVs, let’s export that and the blocked-out blanket 03
geo as an OBJ to bring into Mudbox.

Break it down to bite-size pieces

It took me quite a bit of time to build a head I was happy with and with nice UVs. Don’t feel
pressured to rush into creating a large, complete project right away. At times, I think 3D
artists take on far too much as a first project, and end up sacrificing quality in order to
complete a piece. Begin small and work on easier projects while concentrating on
fundamentals. Practise anatomy and hone in on your skills in between bigger projects.

3D Art & Design 71

 Photorealism

Sculpt the head in Mudbox

Start forming and texturing the head

04 Add expression and pose With our base head

exported, open it up in Mudbox. I spent quite a bit of
time ensuring all the proportions of the face were correct and
04

looked at as much reference as possible. As we’re creating this

character from the imagination, it will be important to ensure it
looks natural and all proportions are accurate. Get the face in a
proper neutral state where the expression is blank but natural
when looking ahead. Place a bone in Mudbox on a sculpt layer,
rotate the head to the angle that matches the reference and
continue refining the shape. Add the expression you want to
convey in the face, largely using the Grab and Sculpt brushes.
At this stage we’re ready to add all the pores and wrinkles.

05 Create pores and wrinkles When creating the pores and fine wrinkles, we will need
to subdivide the model to a high enough division to accept the sculpting information. I
subdivided mine up by five. Next, create a layer called Pores and select the Pores stencil that
comes with the Stencil brush in Mudbox. From here you can use the Sculpt brush and start
applying it to the entire head. With the pores complete, create a new layer called Wrinkles and,
using the Knife brush, start defining the lips, the areas under the eyes and other areas that may
have visible wrinkles. For the final touch, make a Bumps layer and use more of the provided
stencils in Mudbox, applying each to the model to get the look we’re aiming for.

06 Scan and retopologise the blanket For my blanket I wanted to try something new,
so I used the free 123D Catch software by Autodesk on my smartphone. I asked my
wife to wrap the blanket around her in the proper pose, and I used the software to capture 360
degree images of her before using the program to create a very basic 3D mesh. It worked
extremely well. From here I retopologised the model, scaled it and further sculpted and
manipulated the geometry to fit with my head mesh in Mudbox.

05

06

72 3D Art & Design

 Realistic human portrait

Sculpt back and forth

When sculpting the head in Mudbox it’s okay to go back
and forth throughout the process. When you’re at the
rendering stage, you might notice that the shape of the face
or expression isn’t quite reacting the way you were
intending with a skin shader applied. As such, I usually bring
my sculpted head into Maya, apply a skin shader and try
different lighting conditions to see how natural it feels. If I
notice areas in the face that aren’t reacting correctly to
lights and shadows, this usually means there is further
refinement to be made to the expression or shape.

07 Texture the face For texturing I relied on photo

reference for projections. I began using the Paint brush
and blocking out very basic colours to show where I wanted blush,
lip colour and eye shadow present. From there I created new layers
and used photos of Oldriska from 3D.sk as texture reference for
projecting her skin and freckles. I wasn’t too concerned with the
top of the head, ears or back of the head, as I knew none of these
would show in the final render. I wanted to ensure that the freckles
would pop as well, so I used the Burn brush afterwards to really
punch them out when it came to rendering. I also projected a faint
colour for the eyebrows to use later as a template with Shave and a
Haircut. When done, save the finished texture as Colour, bring it
into Photoshop and create a slightly desaturated blue with less
07
contrast. Save this out as Subsurface.

08 Use Knald for reflection For my Reflection and Gloss maps I used a newer piece of
software called Knald (www.knaldtech.com). Among other features, Knald enables you
to take maps and generate them into others. To start, save out a hi-res version of your head from
Mudbox and then a lower subdivision (2 or 3 should be fine) without the pores and wrinkles
present. Next, use XNormal or Mudbox to extract a normal map at 4,096 x 4,096. With the
map baked, let’s open Knald and play with the settings so there is enough detail present in the
maps. We can then save out both an AO map and a Concavity map at 16-bit TIF format.

08

09

09 Complete the Reflection maps Open Photoshop

and start to build the Reflection and Gloss maps. First,
take your colour texture, desaturate it and bring it in as a new
layer. Now import the AO and Concavity maps you exported in
the last step and set them both as multipliers on top. I tend not
to add too much contrast in these maps as it will not enable the
reflection to react to the skin as intended. Next paint on some
additional layers for areas such as lips and eyelids to punch the
amount of reflection with a higher value and save out the map
as the Reflection map. Next, open up a duplicate of the map,
blur it and paint dark areas where you’d like to have broader
reflection and brighter areas for hotter reflection. Save this out
as a Gloss map. Examples of both are provided.

3D Art & Design 73

 Photorealism

Add realism to the hair and eyes

Create the hair with Shave and a Haircut and the eyes with realistic textures

Styling
the hair
When styling the hair it’s
important to take your
time. I’ll occasionally do
quick test renders to see
how the hairs are
clumping together to get
an idea of how it may look
later on. I created two
separate passes of hair
for this project – one for
the majority of the hair
and the other for
individual strands that fall
in front of her face. By
creating additional hair
selections you can control
10 Style the hair To create the hair, import the lowest
subdivision of your head mesh into Maya and extract a
cap for the area you want to spawn hairs from. Duplicate the
them later on with Maya head mesh then select the faces that will act as our hair cap
Hair. I tend to turn off the and use the Extract tool in Maya. Select the newly extracted
Shave and a Haircut Hair
mesh and using Shave and a Haircut select Create New Hair.
Visibility tab and just
focus on the strands, as With our new hair created, let’s update the collision mesh by
that’s what I will selecting the hair and shift-selecting our original head using
ultimately be using. For Shave>Edit Current>Update Collision Mesh. I used the Shave
the eyelashes I use the Brush tool by hand to style the hair to achieve my desired look.
same approach as that of
the eyebrows. Once happy with the results, use Convert>Guides to Curves
10
and then save them out for later to be used with Maya Hair.

11 Shape the eyebrows For the eyebrows, import your low-res head mesh into a new
scene and apply your Colour texture to it to act as a guide for styling the eyebrows. We’ll
want to create curves to act as our eyebrows later that sit on top of the head mesh. To do this,
let’s make our head live by selecting it and then the magnet icon in Maya. With our head live, use
the EP Curve tool and draw the specific eyebrows you want. Take your time with this and ensure
that parts are elevated above the mesh by moving the curves after you’ve created them. With
11
the adjustments made we can save these out as an eyebrow group for later.

12 Model and texture the eyes When modelling the eyes I use a simple approach that
has always worked for me. I break down the eyes into three separate parts: the lens,
sclera and pupil. For the lens, I create a sphere and ensure there is a bulge outwards such as in a
12

real eyeball. For the sclera it is a concave sphere with the centre hole cut out for the pupil. The
pupil itself is just a standard plane that sits directly behind the sclera and is completely black. The
sclera is the only piece that I texture with a colour map that you can find provided with this book.
Next I will show you the specific materials I use.

74 3D Art & Design

 Realistic human portrait

Shaders and lighting

Shader creation for the skin, hair, eyes and lighting
13
13 Lighting setup Bring in the highest-res OBJs of the
head and blanket. Ensure Visible In Reflections and
Visible in Refractions is selected for both meshes under your
Render Stats Attribute. For lighting create a V-Ray Dome light
and select an HDR image to be placed into the Dome Tex slot
at maximum res. For my HDR image I used Newport Loft,
available for free from hdrlabs.com. Set your Subdivs to around
24 for better sampling, and play with Intensity to find what
works best for you. Create three basic poly planes and apply a
basic lambert of a colour you want to illuminate your mesh. I
created an off white colour for mine. Place them in areas where
you would like to see some illumination on the character.

14 Render setup
Some of the major
takeaways for my render
14

settings are my Adaptive

DMC settings, each of which
I cranked to 6. This makes a
huge difference in the quality
of my hair rendering. For the
Reflection Texture I used the
HDR image from my dome
light and set the GI Texture to
a warm colour rather than
texture. For the global
illumination ensure it’s turned
on to take advantage of the
planes you created. For

15 Layered skin shader with reflection For the skin

shader, start by creating the layered V-Ray shader VRay
Blend Mtl. Create the skin material by selecting a V-Ray Fast
Primary Bounce I used Brute
Force and for Secondary
Bounce I used Light Cache. I
SSS2 shader and plugging it in to Base Material. Plug your used a multipliers setting of .7
Colour texture in to the Overall Color slot and your Subsurface but you can experiment.
texture into the Sub-Surface Color slot. I ended up using a dark
red for my scatter colour and changing the Scatter radius to
.650 (depending on the size of your head mesh). Next let’s
create the reflection shader by creating a VRayMtl and slotting
16 Create the eye
shader For my eye
shader I created a VRayMtl
it into Coat Material 0. Input your Reflection map into your for my lens mesh and a
Amount under reflection, and your Gloss map into your VRayFastSSS2 for my sclera. I
Highlight Glossiness slot. You can see the specific settings I want my lens to have only
had for my shader in the screenshots supplied with the disc. reflection and refraction and
my sclera to have a bit of
15 subsurface scattering. I
turned my Reflection Color
on the lens to just above black
and set the amount at .740.
Adjust both depending on the
amount of reflection you
want in your final image. I
turned Refraction Color to
white and set it to 1.0.
Refraction IOR was best set
to 1.60. I also had a slight
fractal bump map on my lens
but it was barely noticeable.
For my sclera I put my eye
16
texture into the Overall Color.

3D Art & Design 75

 Photorealism

Render and composite

Add the final elements and bring the image together

17 The hair shader When rendering the hair I found it best to take the curves we created
earlier and apply them to a Maya Hair system. First let’s create our hair system by
creating a basic polygon cube and selecting Hair>Create Hair. Next, select your curves group and
18 Render Elements With everything ready to go, start
rendering with V-Ray. I prefer to turn on Multiple
Render Elements under Render Settings. Here you are able to
go to Hair>Assign Hair System>hairSystemShape1. This will apply your hair system to your hair choose what layers you prefer to render separately and
curves. Play with the specific settings for how you want the hair to look. To select a V-Ray Hair therefore have the ability to adjust in composition later on.
shader, select your hair system, go to the top and select Attributes>V-Ray>Hair Shader. Now You can see the multiple layers I decided to render separately
select VRayMtlHair3. You can see the settings used in the screenshots supplied with the book. in the screenshot supplied with the book.

17 18

Don’t forget the lashes and brows

A couple of things I haven’t touched on are the eyelashes and the eyebrows. For the
eyebrows I used the same method for creating the hair. I started with the curves, created a
new hair system and then applied it. For the shader, I relied on a duplicate of the hair shader
with darker values. For the eyebrows I created a new hair system but used the default
material that comes with Maya Hair at almost a black value with slight transparency. Hair
takes time to render correctly so don’t get discouraged if it’s not perfect right away.

19 Quick Photoshop comp All the hard work you’ve

put in has finally paid off and now it’s time to play
around with some colours and values. At this stage I did some
minor touch-ups and colour adjustments. One thing I found
useful to do is to play with some overlays using your render
elements. There is no right and wrong way of touching up your
photos. Ideally you will have done most of your hard work
before rendering, so you won’t have to worry about adding too
much at this stage. For this project, I darkened the lashes and
19
played with the levels to find the exact look I was after.

76 3D Art & Design

 I integrated MARI and Photoshop to Gallery
paint all my textures. All shader
Easy-to-follow guides
take you from concept parameters were based on the real world
to the final render (Model Size, Scattering Distance, Skin
Artist info
Refractive Index, Fresnel Reflection). I
then fine-tuned each parameter to get
the final effect on the image
Leng Shang-Peng Then Flying The Youth

Leng Shang-Peng
Website www.power3d.cc
Country China
Software used Maya, Silo,
ZBrush, MARI, Photoshop,
NUKE
Bio Once a CG supervisor at 37
Digital in Shanghai, Leng now
works in 3D tool development

I used a V-Ray light

and Global Illumination
for the final lighting.
Rendering the whole
image took up more
than 50GB of memory.
For any large-resolution
rendering piece you
must ensure there is
enough memory on
your computer,
otherwise the render
will be slow
3D Art & Design 77
 Photorealism

Design a
stunning dress
Lady in Red 2013
Use Marvelous Designer to
create beautiful images of
flowing fabric
Sander Boerefijn is a 3D artist with a
background in mechanical engineering

Easy-to-follow guides
take you from concept
to the final render
Artist info

I
n this tutorial we are going to create a dynamic dress design with the software
Sander Boerefijn program Marvelous Designer (MD). Marvelous Designer is primarily focused on cloth
simulation and is mostly used for the creation of clothing and furniture pieces. It has even
been adopted by studios as prominent as Weta, for use on the production of The Hobbit.
Username: SanderB Throughout each step of the dress creation in this tutorial I will cover the basic Concept
Personal portfolio site functionality of MD right through to some of the more in-depth aspects. We will discuss
I wanted to create a
www.3dvance.nl everything from creating the patterns and the sewing and draping of the fabric to the beautiful, dynamic and
Location The Netherlands
physical properties and animations. At the end of the tutorial I will also reveal the scene flowing dress – one that
Software used Marvelous
Designer, 3ds Max, Adobe setup in 3ds Max for the final render. The great thing about MD is that you don’t need a almost seems to have a
Photoshop fashion design background to follow this tutorial or to create any kind of clothing. There’s a life of its own. The image
Expertise 3D generalist, was inspired by a variety
lot of creative fun to the process! Be sure to check the disc for some accompanying video of fashion photography.
architectural visualisation
files that will go into more detail on several of the steps discussed.

78 3D Art & Design

 Design a stunning dress

Avatars and patterns

ěũũ 15#+.42ũ#2(%-#1ũ
ē9/!ũăũ+#
The basics of using Marvelous Designer
ěũũ43.1(+ũ5("#.2
ěũũ43.1(+ũ2!1##-2'.32

01 Explaining avatars In real life fashion design, designers use a mannequin or dress
form to drape and size their designs. In MD this mannequin is named an Avatar, which
can be seen in the left 3D window. You can pose this avatar in MD using X-ray joints, but for this
tutorial we are going to use a custom avatar. We need an avatar in both a T-pose and a final pose,
both of which we will both import as an .obj. We want to start with a T-pose avatar so it is easier
to drape the initial clothes before animating. However, before we import it we want to make a
backup of the bounding volumes (A-BV) and arrangement points (A-POINT) of the default
avatar. This way we can reuse them with our imported avatar later on.

01

02
02 The pattern window Before we start making the
2D patterns for the undergarment we want to enable
the grid and then set the grid distance by going to the pattern
window properties (F12). This and the unfold option will help
us when making symmetrical patterns such as this dress.
Start with the freehand polygon tool and create only a half of
one side of the pattern. Use the background shape of the
avatar to get an approximate idea of the correct size. Now
right-click on the edge over which you want to unfold and
then copy and paste the result to create the backside. Using
this workflow you can adjust the pattern relatively quickly
until you are satisfied.

03 Add hems Next

we will add some
hems to the top and sides of
the dress. With the Add
Point/Split Line tool,
03
right-click on the edges on
the sides of the hem. Now
you can fill in a distance to
create a point on both sides. Using the Marvelous
Disable the grid for now so
you can use the Internal Line
Designer avatar
tool to connect these two I used the avatar that ships with MD, adjusting it and
points. Edit the curvature of posing in 3ds Max, before exporting back into MD. To get
this avatar or the catwalk animation that comes with it into
this line to match the outer your 3D package you need to open the .avt file that is
line. Make a new polygon out located in the Marvelous Designer install directory (C:\
of the four points and adjust Program Files\MD 3 Personal\Model\Avatar\AVT). You
the curvature to match it with can open this file using an unpacking software like Winrar
or Winzip and in it you can find the Collada (.dae) file and
the others so that we have a the textures.
piece with the same shape.

3D Art & Design 79

 Photorealism

Pattern sewing, draping and adjusting

Time to start working in 3D
04
04 Sew the undergarment Before we can drape the garment we must first use the
sewing tools to connect all the pieces together, as you can see in the image. Ensure that
the crossed line is always on the same side of the sewing line on both pieces. If the seam is
reversed you can fix it by right-clicking on the sewing line with the Edit Seam Line tool and pick
Reverse. Sometimes it’s hard to see if the seam is correct but the 3D window will help.

05 Drape, simulate and adjust in real time When we press the sync button the
patterns will appear in the 3D window but need to be set in place. For this we are going to
load the bounding volumes and arrangement points we saved earlier. Now we can select a pattern
piece and click on the arrangement point where it should be placed. Once the pieces are in place
and all seams are correct it’s time to start simulating. Once the clothing has settled we can identify
if there are some places that need to be fixed by either changing the 2D patterns or by dragging the
clothing around in the 3D window in real time. Another helpful tool here is to use the Strain Map
mode, which indicates which parts of the clothing are too tight or too loose according to colour.

06 Start on the bottom half Start by creating a polygon circle by left-clicking in the 2D
pattern window. Fill in a radius of 3500.00mm and delete the right half so a semicircle
remains. Split the straight line into seven uniform lines and split the resulting middle line again in
two uniform lines. Grab the point last created and hold Shift to drag it to the left. While holding
down the Shift and left button, click on the right button to get a popup to enter a value. Fill in the
amount of half the length of the seven lines’ length. Now create a new circle using the line length
and place it right in the middle of the three middle points. Now we can use this circle to trace
around and then delete it. On one of the outer sides we want to leave out a part of this pattern so
the edge of this layer will try to curl a little. Copy, mirror and then paste this part to create a full circle
05 again. Sew it together, and then onto the lower edge of the undershirt. Simulate it falling into place.

06

Particle Distance
It’s a bit of an odd term for what it defines, but the Particle Distance value determines how
dense your mesh will be. This value is the average distance between vertices in millimetres. So
a lower Particle Distance will increase the amount of vertices and with it the detail. Be careful
that you don’t go too low. Set the Particle Distance to your desired mesh complexity.

07 Wrap the top section Before we wrap the top part of the dress, we want to create two
internal lines in the undershirt to attach the draping layers and the drapes of fabric. Freeze
what we have created so it won’t interfere. Create a rectangular pattern piece and sew it to the line
we just created on one end. In the settings for the piece set the layer property to a value higher than
the underskirt so it will fall on top of it. Now we can use pins, created by holding W down in the 3D
window while clicking, to pin the long part into place. We need to move the pins around to make
new pins while wrapping the piece around the body. At the end remove all pins, except maybe a
few important ones, and simulate it into place. You might need to drag the pattern piece around to
07
get it placed as you wish. Lower the Particle Distance to give it some more detail.

80 3D Art & Design

 Design a stunning dress

Material and physical properties

Animate the dress layers and physical properties of the materials

08 Material physical properties When we’ve

simulated the pieces up to this stage, they have all
reacted in a manner dictated by a set of physical material
properties. The undershirt will have stretched out when the
draping part was pulling it down, for example. You can change
these settings to simulate different kinds of cloth materials, using
MD’s useful list of presets. These presets provide a helpful
starting point for changing material physical properties. For the
undershirt we will apply a simple cotton preset as we won’t see a
great deal of this garment in the final render. However, for the
more visible parts of the dress we will use the satin preset. Lower
the values of both the Bending-Weft and Warp to 2 so they will
bend a lot easier. Also lower the Density so the fabric will be
lighter when simulating. Change Stretch-Weft to 26 and
08 09
Stretch-Warp to 23.

10

09 Animate layer one Now that we have set the physical properties we can start animating
the draping layers using the Wind option. Activate this from the Environment dropdown.
Select the Wind icon and in the properties set the Strength to a slightly lower amount and enable a
Frequency that causes the wind to start and stop over set periods of time. Set the wind to Spherical
and place it just underneath the avatar’s feet, causing it to blow the dress upwards. Click the Simulate
button and see if you are happy with the result. If not, tweak the settings. You can also use the mouse
to drag the dress in the direction you want it to go. Once you are satisfied with the way it looks, stop
the simulation. At this point we want to disable the wind again and start increasing the Particle
Distance of the dress to create more detail.

10 Start on the other draping layers Once the first layer is done save the file and
export what you have to .obj. Create a new avatar that includes the first draping layer.
This way it doesn’t need to simulate the first layer again and our second layer won’t go beneath it
when we simulate it. We can create the next layer quite simply by using the pattern of the first
layer once we have imported the new avatar. This routine will be duplicated for the third layer
once the second layer is completed.

11 Animate the second layer For the next layer we will follow a similar routine as we did
when animating as the first. However, we need to raise the wind settings so the upper
layers will float higher. We also need to add pins to the end of the fingers to prevent the next
layers from going through the hands. This part of the creation process can require some
tweaking of settings and pulling of cloth to achieve the desired result. If you’re using the advanced
version of MD you can just let it simulate and then afterwards select the specific frame you like
11
best. Lower the Particle Distance to make the simulation go slower and create more detail.

3D Art & Design 81

 Photorealism

Finalising and rendering

Finish the dress in MD and render using a V-Ray
13

12

12 Finalise the dress Once you are finished animating a third and final layer, you will need to
complete the top part of the dress. In the avatar properties set the Skin Offset to 1 and make
sure the wrap part is on a higher layer value than the third draping layer. Simulate the wrap part so it
will fall over the third layer. Next, add the other layers and switch the avatar to the first posed one.

13 Export to 3ds Max To export the dress we need to order the 2D patterns nicely in a
squared shape, because it’s from this that we will generate the UVs. In the export .obj
options use unified UV coordinates to export the cleanly laid out UVs. Use the same scale as you
will use in 3ds Max. In my case I am using cm. After importing the dress into 3ds Max, add
ProOptimizer modifiers to the separate parts to lower the vertex count if needed.

14 Set up the V-Ray material By adding an unwrap

UVW modifier and extracting a UV template from the
dress we can create an Opacity map and a Bump map. The
Bump map is used to create some additional wrinkles and in
my case was created from textures of crumpled paper. The
Opacity map adds the illusion of thickness to the material
edges. These maps were then used in a VRay2sidedMtl, which
gives the dress a very thin, see-through feeling. The falloff map
is a map I use a lot in my cloth materials. Most of the time I will
turn off the trace reflection option. The skin material is a
14
VRayFastSSS2 and the hair material is a VRayHairMtl.

82 3D Art & Design

 Design a stunning dress
15

32hours
creation time
Resolution:
6,000 x 3,500

Importing personal patterns

For this image I created the clothing patterns myself, but
you can also use some patterns from external sources
should you wish to. Since MD doesn’t have an import
function for pattern files a little workaround is needed.
First convert your pattern source to a supported image file.
Create a new material and apply the image to the texture
slot. Next, create a rectangular pattern piece and apply the
material. Now you can make the new pattern pieces by
tracing over the image with the Create Polygon tool. Don’t
forget to freeze or delete the pattern piece with your image
on it afterwards.

15 Scene and render setup For this scene I used a

studio setup ,with a key light with a slight blue tone
and a VRaySoftbox texture in it, shining in the direction of the
avatar’s face. I added a main fill light from the opposite
direction and three fill lights on the sides with colours varying
between a light blue and light yellow tone. In the environment
slot I used a 360-degree spherical image of a warehouse
that creates realistic reflections to add to the depth of the
piece. The avatar’s hair was achieved using the Ornatrix hair
plug-in from Ephere with the VRayOrnatrixMod modifier on
top of it.

3D Art & Design 83

 Photorealism

Easy-to-follow guides
take you from concept

Create hair
to the final render
Artist info

Yasin Hasanian
Personal portfolio site
www.superhero.cgsociety.org
Country Iran
and fur
Cold as Lava 2013
Software used Maya, ZBrush,
MARI, Shave and a Haircut,
Arnold, Photoshop
Expertise Yasin specialises in FiberMesh is just one of ZBrush’s many powerful options, so here
texturing, shading and lighting
with a solid knowledge in we’ll use it to make a truly photoreal and captivating image
modelling and sculpting
Yasin Hasanian is a freelance CG artist working professionally in the film and game industries

T
hese steps will take you through how to Obviously, ZBrush is a program that’s opens up a whole host of options for producing
create decent-looking fur with the help of constantly evolving with each release. It’s become brilliant results beyond just hair and fur.
ZBrush. We’ll do this by utilising a game-changer for a lot of us by offering Regarding other tools that will be used for this
FiberMesh in different stages of the pipeline, optimum methods of accomplishing tasks that are project, Maya will be used to set up the scene,
using it in conjunction with other software, otherwise much more tedious. One of these Shave and a Haircut for the fur, MARI for textures,
including plug-ins such as Shave and a Haircut. game-changing features is FiberMesh, which Arnold for rendering and Photoshop for post.

Build a concept Some facts

about tigers
Dissect the fur structure and analyse photo references
Tigers can be up to three

01 Gather reference images Either you are working

on something completely imaginary, or you’re working
on something that exists in real life. For the former, it’s not
metres in length and
weigh as much as 330
kilograms. One of the
most interesting
necessary to find reference images, as they may not actually characteristics about
exist and so you are basically sculpting and evolving your ideas them is their striped
in a program such as ZBrush. For the latter, it’s recommended pattern, which is unique
that you gather as many photo references as you can. Good from one to another, just
like human fingerprints.
photo references are those that help you throughout most of Their pattern acts as
the pipeline stages. For instance, they will hint as to whether camouflage, helping them
the fur in area X is generally soft, tough, short, long and so on. to successfully hide in the
wild from predators. Also,
01 02 not all tigers are orange in
02 An overview of sculpting The final Ztool for this
tiger has been made available with the tutorial, so we
won’t go over the minute details of how it was sculpted here. In
colour. Due to a mutation
in colouring, there are
tigers that feature black
general, the head was made in ZBrush from scratch with stripes and blue eyes.
DynaMesh. One thing that you have to make sure of, however, One significant advantage
that the tiger has is their
is that if you start your modelling completely inside ZBrush and eyesight. They can see
plan to send the meshes later to another program, such as just as well as humans
Maya, check your mesh sizes as soon as you have blocked out during the day. However,
the general shapes in Maya. This means you won’t have to alter at night, they are able to
see six-times what a
the sizes later when it comes to rendering, as this tends to human can.
become problematic in complex scenes.

03 Analyse the references After finalising the sculpt, take some renders into ZBrush to
come up with a sense of how to lay the fur on the tiger and what is required for each
part of it. As you can see in the accompanying image, there are five main hair types and seven
hair systems. Each of these hair types require a different hair system, so the whole project can be
easily managed. For example, the cyan area has a distinct sharp type of fur, while the green area
03
has a soft-looking fur type.

84 O3D3DArtist
Art & Design
 Hair and fur

ěũũ(%#1Ĵē93+ũ-"ũ
2!1##-2'.32ũ.$ũ2(-Ħ2ũ
2#33(-%2

Concept
The main inspiration
for this work was
nature! Tigers are one
of my favourite
animals and two of
their visually striking
characteristics are
their eyes and fur.

3D Art & Design 85

 Photorealism

Lay the foundations

Pre-visualise fur in order to get quick feedback

04

04 Begin your pre-vis Often you are required to give quick feedback to your client or to
make sure you’re on the right track, which means visualising how the work, in this case
the fur, is roughly going to look. Before the introduction of FiberMesh, there was no easy way of
going about this. For example, completing this phase directly with Shave and a Haircut could
definitely take a lot more time and was prone to errors. Now you can quickly apply some fur with
FiberMesh, groom it and render it through BPR to evaluate its different aspects.

05 Optimise FiberMesh In situations where there are multiple fur types or sections on
a character, there are essentially two ways of organising FiberMesh elements to
separate them. The first method is to make polygroups on your sculpt for certain regions that
have different types of fur and then, when the FiberMesh is created, the fibres actually respect
the polygroups, so you can further manipulate them separately. The second method is just
making the FiberMesh sections one by one, which may not be as fast as the previous method but
05
enables you to focus better on certain areas.

07

06
07 Groom the fibres into a pattern While grooming
the fur in this stage, the most important factors you
need to be aware of are the direction of the fur, its coverage
(thickness), the length and the density covering a given region.

06 Start generating fibres Based on the outlines we came up with in Step 5, we can go
over the model and mask each region in turn to grow fibres from. Usually, on the
FiberMesh Modifiers tab, it’s best to zero out any sort of randomisation at the start to avoid an
These key factors will really help you sell the piece to the client.
The Groom brushes, such as GroomHairLong and
GroomHairLengthen, handle the direction and length, while
unwanted result and to build up the look manually. Since we’re still on the prototyping stage, the density and coverage should be taken into account in the
idea is to work as fast as possible without worrying too much about details. After accepting each FiberMesh Modifiers tab. GroomHairLengthen is pretty much
of the FiberMesh sections, we can groom them with GroomHairLong. Although the fur isn’t a Move brush, so you can use this brush to manipulate the hairs
actually long, this brush is perfect because it enables more control along each strand’s length. or make them longer as desired.

86 3D Art & Design

 Hair and fur

08 Prepare the mesh After you’re finished creating all FiberMesh elements, proceed to
make a series of adjustments on the sculpt to make the head read better, then retopologise it
with the ever-useful ZRemesher by holding Opt/Alt to preserve asymmetry. This step could have been
completed sooner in the process, but due to a high probability of model changes in the prototyping
stage, it’s much better to complete it at this step. Please note that you can simply skip this step if you
are following the tutorial using the tiger ZTL file provided.

09 Move to the texturing stage At this point you can begin building up the textures.
Most of the time it’s good to paint textures in a 3D texture-painting application, such as
MARI or BodyPaint, in conjunction with Photoshop. Send 2 subdiv levels of the tiger head, both
low and high, to MARI (or your program of choice). Work on the low level to block out the
colouring and switch to a higher level of subdivision to work in the finer details. All the textures
08
here were hand painted in layers, leaving enough freedom in case of a texture alteration.

09

10 MARI performance tips If you are experiencing

any lag or difficulties when working with MARI, try
decreasing the Frame Buffer size. Moreover, in normal texture
painting you don’t really need to paint in 16- or 32-bit buffers,
so try switching to a lower Depth Buffer in order to free more
resources on your graphic card. Caching layers have a red sign
that indicates they are static and not in use any more. You can
always un-cache them later in case you need them.

Volume illusion
If you are working on a furry creature based on reference
images, always be mindful of the volume and shape of
what you are sculpting. For example, if you match your
sculpt completely to a reference image on the silhouette
and later decide to add fur to it, you will most probably get
incorrect results. This is because you’re adding another
layer of silhouette on top of the character, making it look
fuller. Sometimes you might be able to cope with this later
10 by an overall negative inflation.

3D Art & Design 87

 Photorealism

Further refine the fur

Make FiberMesh ready for Shave and a Haircut

11 Continue building FiberMesh elements The

second version of FiberMesh regions can be created with
Shave and a Haircut in mind. Here the accurate appearance of
the fibres is imperative. In contrast to the previous type of fur,
here we want to exhibit less density and use more precise
grooming. For the previous version of fur, most of the time was
spent in the FiberMesh Modifiers tab, whereas here the time is
spent on grooming. The guideline for this round is still the Step
5 figure, with perhaps some minor alterations after the
prototyping stage.

12 Layer the fibres Some areas on the tiger fur have a

layered FiberMesh set up for higher fidelity in the look
development, as well as to have Shave work as expected.
Without this, Shave doesn’t really know how distinct the two
fur types should be, resulting in a uniform interpolation
between the strands. An example of this layered style took
place on the tiger chin – one layer with a dense, short
FiberMesh and one with less-dense and longer elements.

13 Adjust the grooming brushes As you are

grooming the fibres while some other objects are visible
in ZBrush, you might have noticed the fibres do not fully
respect your brush movements and end up with a scraggly
look. This is because grooming brushes have the Front Collision
attribute active by default. To disable this, go to the Brush
palette>FiberMesh and zero out Front Collision Tolerance. This
slider keeps away the fibres to penetrate into the object’s
surface. In fact, Preserve Length in this tab tells ZBrush to treat
any mesh the way a FiberMesh is treated, provided that it is
11
enabled on a brush.

12

Shave and a Haircut tools

S&H offers a bunch of brushes to comb the hair, but here
the only brushes I used were Translate, Scale and, where
the strand moved inside the geometry, Stand. Aside from
the brushes, the Shave selections came in handy when I
wanted to accurately comb the fur, by switching to
Component mode, clicking on a selection mode and
manipulating the guides one by one or as a chunk. Still, a
major drawback of S&H is the limited Undo feature. As
such, if you plan to do a major combing, save your scene! 13

88 3D Art & Design

 Hair and fur

Fit the pieces together

Prepare the tiger fur ready for rendering

14

14 Set up the scene in Maya Now send all the parts to

Maya to prepare them for rendering and export out all
the FiberMesh elements to serve as guide curves for Shave. In
order to convert the fibres to curves, you can use this script:
tinyurl.com/TDAHairAndFurScript. This is why the second
version of FiberMesh elements were quite sparse. For each
section of the fur, extract a root geometry from the low-res
head, grow Shave out of it, then comb the Shave with the
curves created out of each corresponding FiberMesh.
Additionally, you can make Control maps, such as Density, to
further refine the Shave nodes.

15 Add whiskers Regarding the whiskers, you can totally

rely on FiberMesh from the start. Toggle the Brush Size
to Dynamic mode, move to the highest subdiv level and begin
masking strong little dots on the places you want the whiskers
to grow out of. Next, preview them and, back in the Modifiers
tab, tune it in a way that only one strand per dot grows.
Afterwards, give them a Profile of 6 and Segment of 10 because
we’re going to render them directly and not convert them to
guide curves. To finish this step, groom and export these
15
elements to Maya.

Yasin Hasanian
Yasin was born in Iran in 1990. He has had a
great passion for both the art and science
behind CG since he was 14 years old. He
holds a bachelor’s degree in Computer
Science and has been working as a
freelance CG artist in the industry for over Miss Mutation Maya, ZBrush, mental ray, Say Cheese to the World Maya, ZBrush, Shave
five years. Yasin has created work for major Photoshop (2011) and a Haircut, mental ray, Photoshop (2010)
clients, but he still wishes to one day land a Maya was used for the base mesh and scene setup, Sculpting and most of the textures were made and
job at a big studio. ZBrush for sculpting and Photoshop for texturing. hand-painted using just the mouse.

3D Art & Design 89

 Photorealism

Render and post work

Modify your shaders and set up the final render in Arnold

16 Set up the shaders Here the Arnold aiHair shader

was used for the fur. The same colour texture that was
painted in MARI was assigned to both the Root and Tip slots,
however, each was controlled with Remap HSV nodes. Set the
Ambient Diffuse to 1 so that the shader takes the lighting
direction into account. Although the physical range for Indirect
Diffuse is from 0 to 1, a value of 2 has been used here for
artistic purposes. Only the primary specularity was used with
low Strength and Glossiness to maintain the shader’s energy.

17 Begin the render There’s only one Skydome light in

the scene with an enhanced HDRI texture, meaning
both the Diffuse and Reflection were altered. All the Shave
nodes had a Min Pixel Width of 0.3, with the default Arnold
Transparency Depth of 10, which is enough to provide a soft
look in certain areas. A stand-in of the high-res tiger head with
pre- and post-render scripts was used to get switched with the
low-res version for rendering, which saved time by skipping the
displacement tessellation process.

Arnold and fur rendering

Arnold is an amazing renderer that’s optimised and
powerful. Arnold works well with fur/hair and its IPR even
supports node initialisation. For example, you can simply
launch IPR and create your shaders, assign them to Shave,
tweak them and it just renders it simultaneously. To control
fur rendering artefacts, you can either decrease them by
increasing the AA samples or Min Pixel Width. However,
increasing MPW too much increases render time, as the
16 renderer has to shoot more transparency rays.

17

90 3D Art & Design

 Gallery
This image was created in 3ds Max, then
Artist info
lit and rendered in V-Ray. It was then taken
into Photoshop for basic post-production with
combined vignette, flare effects and creative
depth of field to add some atmosphere
Set Visions Set Visions Imaging Studios Photorealistic CGI, 2013
Imaging Studios
Set Visions produces
photorealistic CGI under the
exclusive brand PIX
Personal portfolio site
www.setvisionspix.co.uk
Country UK
Software used 3ds Max, V-Ray,
Photoshop

3D Art & Design 91

 Photorealism

Bring still-life
objects
Afternoon Siesta
to life 2013
Construct a realistic still-life scene using photorealistic
techniques and the careful study of light and shade
Sofian Moumene works as a lead CG artist at Carioca Studio in Bucharest, Romania

ěũ4++ũ
$3#1-..-ũ(#23ũ
2!#-#
ěũ,.!'(-Ĵ)4,3#ē
ěũ43.1(+ũ2!1##-2'.32

92 3D Art & Design

 Still life

T
he purpose of this tutorial is to create a
photorealistic still life scene from scratch. We’ll
start with modelling the various organic and
Easy-to-follow guides
take you from concept non-organic objects and then move on to building detailed
to the final render shaders. We’ll be using 3ds Max for the base modelling and
ZBrush for detailing and texturing. Marvelous Designer will

Artist info
also have a role in simulating the drape in the scene, and
finally Photoshop will be used for some post-production
touches. We’ll be using ZBrush’s Spotlight to texture the
various fruit in the scene, which is a great tool for eliminating
any troublesome UV seams.

Sofian Moumene Concept

Personal portfolio site
www.sofianmoumene.com For this project I
Country Romania wanted to illustrate a
Software used 3ds Max, V-Ray, feeling of relaxation
Marvelous Designer, Photoshop and warmth –
Expertise 3D and advertising something that you get
illustration for a variety of on a vacation or a long,
commercial projects lazy afternoon.

3D Art & Design 93

 Photorealism

Sketch out your composition

Organise your essential starting materials

01 Search for
reference images
You can use Google to find
02 Complete rough sketches With all reference
images gathered, a clearer image of the whole scene
should start to take shape in your mind. You can then start to
The Marvelous
Designer drape
free images, or Shutterstock make some very rough sketches in Photoshop based on your
for high-res and high-quality early ideas. Also try making some basic shapes in 3ds Max and The drape you see here was simulated in Marvelous
Designer, which is great for creating different types of cloth.
photographs. These will aid quickly arrange them to figure out the composition. This is also I exported the objects that I knew would interact with the
you when it comes to the step where you can create your camera and decide on the drape from 3ds Max in DAE format. This included the
detailing the shaders as viewing angle. Of course, many things will likely end up background wall, the table and the ceramic wine pot. Next I
accurately as possible. This is different in your final render, but this stage will still help you to created a rectangular sheet of cloth in Marvelous Designer,
at 200 x 80cm. I clicked on the red Sync button to
an important step as you establish some broad, foundational ideas. synchronise the 2D viewport with the 3D one. This way, if I
need to be clear about the modified something in 2D it would be visible in the 3D
direction for the materials version as well. I went to Environment>Gizmo>World
you’re using, as well as the Co-ordinates Gizmo in the top toolbar to change the object
overall design of your assets. axis in a more controllable mode. Next I moved the cloth
above the imported objects and pushed the Simulate
button to see the cloth falling on the objects. I started
01 pulling and arranging the cloth as I wanted in order to get a
pleasant shape, then selected the cloth and lowered the
value for Particle Distance from the right-side Properties
menu to get a more detailed mesh for the cloth. I exported
the resulting mesh as an OBJ and imported it in ZBrush to
retopologise it with the QRemesher, because the mesh
from Marvelous Designer had jagged edges. I quickly
applied QRemesher with the Same option ticked so I got a
similar number of polygons after the remeshing. I also made
02 a UV with UV Master and exported the mesh in 3ds Max.

Start modelling
Use a variety of tools to model objects

03

03 Work on the room Use real-world scale for the

whole scene and build just two walls – one in the front
to form the background of the still-life subject and the one on
the left, where you can also put a small window to use as the
main light source. The floor isn’t visible in the final image, but
it’s important to include it to obtain more natural light bounces.

04 Shape the wicker basket For this element, start

by creating a cylinder (Radius: 18cm, 24 sides and 5
segments). Add an editable poly, select the vertical edge loops,
as well as one horizontal edge loop, and convert them into
splines. Make a shape as the cross-section of the wicker strand
and use the splines you just added as a path with Loft active
(Create>Compound>Objects>Loft) to define the volume of the
wicker. Make another copy, vary both the wicker strands and
then clone them to create the walls of the wicker basket and
the bottom part. Select the walls and the bottom region and
apply an FFD (2x2x2) modifier to give them a conical shape.
You can create the basket’s handles by adding a Twist modifier
to three splines, then a Bend modifier. Finally, edit these with
04
and Edit Spline modifier to fit the rest of the wicker basket.

94 3D Art & Design

 Still life

05 Make the figs For the figs you should use the references you gathered in step one,
paying close attention to them as you model. Start drawing the contour on the reference
image with the Line tool and use the Lathe option to convert it into a 3D shape. Next, apply an
Edit Poly to refine the shape and make two versions of the fig – one cut in half and one whole.
After this step is done, just grab the unwrapped UVs and export them as OBJs to ZBrush to
create the tiny details. These can be used later back in 3ds Max as Displacement maps. In
ZBrush you can use the Spotlight tool to project a photo of half a fig on the 3D model and then
05
apply the Mask by Intensity option from the Masking dropdown menu. To finish this step, grab
the Inflate brush, paint the unmasked areas of the fig and replicate all the details from the
06 reference. Clear the mask and sculpt in some extra details where needed.

06 Add the pear To model the pear, you can use the same workflow as for the figs: first
drawing half of the pear in 3ds Max with the Line tool, then applying the Lathe option to
convert it into a 3D shape. Again, you can refine the fruit’s shape with an Edit Poly modifier and at the
end make it into a UV. In ZBrush, paint the textures (which can also be used for the shading process)
and then use the masking tools in ZBrush to achieve the finer details and shape.

07
07 Bring in the plates and cutlery To model these
objects you can apply basic techniques with the Knife
tool, using an Edit Poly to model the rough shape, then
importing it into ZBrush. Add the details such as little bumps
and scratches, as well as the decorative patterns from a vector
pattern (made in Illustrator, for instance). For the plate, you can
apply the same process, sculpting in the details by hand. For
the cup, grab the Line tool to draw the outline and then use
Lathe and Edit Poly to refine the shape a little. Find some
decorative vector patterns from Shutterstock that you like and
then paint the UVs in Photoshop to get a Displacement map.

08 08 Create the grapes Modelling the grapes is more

complex due to the amount of objects. First block out
the shape of the cluster of grapes by making a sphere and
cloning it out as instances, so when you modify one, the change
applies to all of them at once. Once you get a shape you like,
add a Noise modifier with very subtle intensity and a Squeeze
modifier to stretch the spheres a little. You can also start to
make the leaves from a photograph. There are different
approaches to this, with the simpler method being to use an
Alpha map of the leaf and texturing a plane with the Noise
modifier. However, as this is a relatively close-up shot you need
to really convey the thickness of the leaf. So, draw the outline of
the leaf with the Line tool in 3ds Max and then apply an Edit
Poly modifier. You can also insert a Subdivide modifier to
09 increase the poly count without encountering any of the
problems that TurboSmooth causes when you use it with an
uneven poly flow.

09 Finish the grapes element Next, apply the Shell

and TurboSmooth modifiers, as well as the Bend and
Noise modifiers to add more details. Next, model the stems by
10 drawing them with the Line tool in 3ds Max and giving them a
slight amount of thickness. Export the mesh to ZBrush and use
DynaMesh to make it into a single continuous mesh. Sculpt all
the details you wish and create the UVs with UV Master.
Export it back to 3ds Max.

Modelling with layers 10 Model the glass and ceramic wine canister
For this, you’ll once again need to use the Line tool for
the outlining, along with the Lathe modifier. Add another Edit
One of the reasons I love 3ds Max is because I can model almost anything using modifiers as Poly modifier to model the handle and the neck of the canister.
layers. I always keep as many of the modifiers as I can use without collapsing them to Apply the TurboSmooth and Noise modifiers to add some
editable polys. This way I have access to the layered history of the steps I used to reach a small imperfections to the glass. For the ceramic canister, use
certain shape. Of course, it’s not great to have too many modifiers in the stack, because you the same method but at the end add a Displacement map to
can end up using too much RAM, but keeping the vital ones can be a great help.
mimic the horizontal marks left on the clay.

3D Art & Design 95

 Photorealism

Light the scene

Experiment with HDRIs and settings

11 Lighting test It’s a good idea to complete some different experiments with various
HDRIs for the reflections and lighting. As we know that the main light source is coming
from the little window in the upper-left corner that was made during the modelling process, we
11

just have to find the right HDRI for the reflections.

12 Define the light sources Try to avoid filling the scene with lights and use as few light
sources in the scene as you can. In this example, only two V-Ray Dome lights coupled
with an HDRI and a V-Ray rectangular light have been used. When you’ve decided on which
HDRI to insert, just create a V-Ray light source and select the Dome type, load the slot with the
chosen HDRI and then experiment with the Intensity multiplier. As we don’t need the light from
the HDRI, just the reflections, we can deselect Affect Diffuse from the option’s rollout. Now
move to the exterior of the room, just outside of the small window on the left of the scene, and
try making a plane-type V-Ray light as the main light source. Experiment a little with the position
and intensity to quickly find a satisfying result that looks natural.

Create drops with MultiScatter

Later, near the end of the creation process, I decided to make the fruits a little wet to give
them a more lifelike appearance. To do this I simply used the MultiScatter plug-in for 3ds
Max to scatter some water droplets on the fruits in the basket. I first modelled a few simple
drop shapes, using some photos I found on Google, then I used four MultiScatter objects to
distribute them on the various fruits. 12

Texture and shade

Use ZBrush’s powerful texturing tools

13 Texture the figs and pears This is the first stage where we’ll take advantage of
ZBrush’s Spotlight tool for texturing. First download some high-res photos of cut-in-half
figs and whole figs, then start projecting with Spotlight onto the 3D model. Here the 3D fig
model was subdivided into around three-million polys to obtain a crisp texture, because the
resolution of the PolyPaint depended on the number of polygons. The best part of the Spotlight
feature for texturing is that you don’t have to worry about the seams. Next, transfer PolyPaint to
the texture by going to the Texture Map rollout and using New from the PolyPaint option. You
can texture the pear using a similar process. The shader for the cut-in-half fig here was VraySSS2
along with the potato preset, because it’s close to the result we want to achieve once the scale
has been altered. For the pear, a simple VrayMtl was added with the Translucency set to Hard
Wax, with some Bump and Translucency maps made from the Diffuse map in ZBrush.

13

96 3D Art & Design

 Still life

14 Make the grape shader To create the shader for the grapes, you can simply apply a
V-Ray Blend material and combine it with two VRayMtls. The first layer you can see here
was matte, opaque and with a lighter colour. The second layer has a more glossy and translucent
finish. Try mixing this with a dust-like texture map, which you can download from Shutterstock,
or another resource site. 

15 Build the shader for the glass container To finish the glass container, first make a
simple glass material. Here a VRayMtl was used with a little greenish fog (Fog Multiplier
= 0.15) and for the dusty effect another dust texture was loaded into the Refract slot with a value
of 20. You can also add a Bump texture to obtain some finer details, enhancing the imperfections
of the glass. The ceramic canister material can be achieved by creating another VrayMtl and
setting the Diffuse and Fog colours to reddish hues, with the Fog Multiplier set to 0.1.

16 Finish with a metal shader Make one version of a metal shader and then modify it
according to each object (plate, cup, knife). Extract lots of ambient occlusion and Cavity
maps from ZBrush using the Multimap Exporter, then use these to blend the different materials
with the help of a V-Ray Blend material to create the dirt in the small details on the plate and
14
knife. Try mixing several textures in a Composite map to finish the cup.

15

16

Work in reverse
In my opinion the best way to learn something is to reverse-engineer it. Supplied with this
book you have the full scene at your disposal, so study the shaders, textures and the models.
If you have any questions please don’t hesitate to contact me.

3D Art & Design 97

 Environment
The range of environments that can be created
in 3D is vast, and waiting to be explored
100 Build a city in 60 minutes 106
Create massive 3D urban environments
for your 3D productions

106 Use modular textures

Create a fast and efficient real-time
environment using modular textures

112 Create a videogame environment

Visualise a concept and bring the white
box into higher-res models for pre-vis

118 Build matte paintings

Look at how Vue can be used to create
matte paintings for film or TV

124 Create underwater renders

A scary diving moment involving a huge
school of fish and a great white shark

131 Gallery
Seedy City by
Zhang Chen

124

131

98 3D Art & Design

118

3D Art & Design 99

 Environment

Easy-to-follow guides
take you from concept

Build a city in
to the final render
Artist info

Brooks Patrick

Username: bwpatrick

Personal portfolio site

60 minutes
CityEngine Sci-Fi City 2013
www.facebook.com/cityengine
Country USA
Software used Esri CityEngine Learn how to easily create massive 3D urban
Expertise Specialist in the
modelling of urban environments environments for your 3D productions
for planning and urban design
Brooks Patrick is a 3D solutions engineer at Esri

F
or those who haven’t been introduced, been successfully used in movies such as Cars 2, hundred-thousand housing units by hand!
CityEngine is a standalone software Total Recall and Man of Steel. CityEngine’s procedural nature gives you the
product that provides professional users in Matthias Buehler from the CityEngine power to literally make infinite building variations,
architecture, urban planning, entertainment, development team explains that creating realistic with full control over all the parameters of each
GIS and general 3D content-creation with a 3D models has been a notoriously labour- building, including the level of detail.
unique conceptual design and modelling intensive process, typically involving teams of set Once your 3D city model is complete, you can
solution for creating 3D cities. designers and multiple specialised software export to your 3D package or game engine of
Esri CityEngine is the tool of choice for the programs. Think of the all the tedious work choice, using one of many industry-standard file
creation of massive 3D city models for movies, invested in movies to model cities such as formats. We make sure that with our technology,
TV and games. CityEngine’s procedural tools have London, Paris, Tokyo or New York, making several you can work in a fast and cost-effective way.

Get to know CityEngine

Create massive cities in minutes

01 Dynamic city layouts This guide provides an

introduction to those new to CityEngine. First you will
need to download your free 30-day CityEngine trial: (www.esri.
Master the
city tools
com/cityengine), which gives you the fully functional
CityEngine Advanced Version, with full export capabilities, that CityEngine is a powerful
product that can build
runs on Windows, Mac OSX and Linux. CityEngine ships with amazing cityscapes.
the comprehensive City Wizard, which will walk you through Digging into any new
the process of creating your first procedural city. Once you software can be tricky at
become familiar with the controls, it’s easy to lose yourself in first, however, the more
you use it, the easier the
the possibilities of city generation. CityEngine offers a library of functionality becomes.
example projects that will help you on your way, ranging from The CityEngine Resource
New York City 2259, to historical examples of a medieval city, Center is a great place to
or even an archaeological reconstruction of Pompeii. start as you take on the
learning curve and get
used to the application.
02 Get to grips with procedural modelling
When it comes to procedural modelling, rules offer
unlimited possibilities to drive the process and rapidly generate
There are example cities,
tutorial videos and data
templates available,
massive cityscapes. These rules are defined in CityEngine’s which will help you to
better understand and
proprietary CGA Shape Grammar, enabling the creation of use the software.
01
complex building forms. You can edit and refine the CGA code
as needed. The generic nature of CGA provides enough room
for a wide variety of designs. Continuous detailing of code
fragments will give you the true advantage of procedural
modelling – complete freedom of the full city layout, parallel to
02
a guaranteed high-quality mesh.

100 3D
O 3DArtist
Art & Design
 Build a city

Tutorial screenshots of
the various settings used

3D Art & Design 101

 Environment

Generate a city
Set up your first procedural scene with the City Wizard

03 Get started When you open CityEngine for the first

time, you will be prompted with the Welcome Wizard. Procedural
This dialog will let you configure 3D navigation settings and modelling
gain an overview of the learning and support resources. You
will also be asked to download example projects, so simply
and GIS
click to start the download. We’ll be using these again later on CityEngine is not only
03
in the process. capable of creating
realistic urban designs
entirely within the tool,
but also based on
imported real-world GIS
datasets. Try generating a
few cities using the City
Wizard, and for those
who want to dive in,
download and open some
of the example cities from
the CityEngine Resource
Center or import
real-world GIS data from
04 05 a city near you.

04 City Wizard This part of the Welcome Wizard will

support you in the creation of your first 3D city
model. You will first want to name your city and specify its size
(small, medium or huge) depending on what machine you are
running. You will then be asked to name your first scene and
specify the terrain settings, but let’s keep these terrain defaults
for now. Next, you need to choose a street pattern to
automatically grow your city. Since this tutorial’s theme is how
to generate a sci-fi city, let’s choose the Spiral pattern.

05 Choose your city style The final step of the City

Wizard will prompt you with a choice between four
different CGA rule sets that will drive the procedural creation of
your city model. Choose the Textured City style and click
Finish. The Wizard will now generate your city automatically.

06 Essential skills It may take a couple of seconds for

CityEngine to generate your city, but you can monitor
this process as well as your computer’s memory allocation in
the bottom-right corner of the layout. For more information on
getting started, you can view Tutorial 01: Essential Skills, which
06
can be found on the CityEngine Resource Center.

07
07 Explore the city
Rule parameters
The initial buildings vary in
height and structure type, but
can be set specifically by
selecting one or multiple
buildings and altering an
attribute in the Inspector. Select
one building, frame it (F key)
and change its Height
parameter in the Inspector
window. You can do this to any
building across the terrain and
find the specific look that you
are aiming for.

102 3D Art & Design

 Build a city

Concept
The procedural methodology enables your city models to
remain flexible throughout the whole production process,
giving you the option to even separate tasks among
different artists (layout, mass modelling, façade, roof
library creation and so on).

3D Art & Design 103

 Environment

Refine your sci-fi metropolis

Assign rules and develop a futuristic theme
08
08 Open the example NYC 2259 Remember those example projects you chose to
download in the Welcome Wizard? Now it’s time to assign rules from the sci-fi project
named ‘Example_NYC2259_2012’ in the city we just created. If you missed downloading the
example during the Wizard, you can access it (as well as many others) from the example gallery
on the CityEngine Resource Center. Once you have downloaded the example, you can begin
importing the project (File>Import>Project>Existing Projects into Workspace). The example will
appear in the form of an archive file, so choose this option and select the file from your download
location. You can now click Finish.

09 Assign some sci-fi rules Once you have imported

the Example_NYC2259_2012 project folder, it should
be visible in the Navigator, which is located by default on the
left side of the CityEngine interface. Expand this project folder
and locate the Rules folder to see all of the rules associated with
this example. Select some Lot shapes of one block in the
viewport, then just drag and drop the buildings.cga rule onto
the parcels. If needed, toggle the Model Visibility switch at the
top of the viewport (or hit F12) to better see the parcel and
street shapes.

09

10

10 Adjust the sci-fi building height When the new rule is assigned, the model will
automatically regenerate those typical buildings as sci-fi structures, but you may need to
adjust their height. This can be achieved using the same process from Step 7 – by selecting a
building and altering the parameters of choice in the Inspector. You will also notice that this
specific rule has different attributes than the Textured City rule set, including the ability to adjust
the level of detail for a higher polygon count.

104 3D Art & Design

 Build a city

Add extra elements 1hour

creation time
Include futuristic flying cars and export your city Resolution:
4,000 x 3,200
11 Populate the city with flying cars Now that we
have applied the new building rule, it’s time to populate
the scene with flying cars. Select a few street shapes, then
drag-and-drop the flyingCars.cga rule onto them. Flying cars
will now start to appear in the scene, making it seem much
more vibrant and appealing.

12 Export your 3D city model In the previous steps

you have learned how to assign rules to procedurally
model a city, even in a variety of predefined city themes. Now
you can save an image snapshot directly out of CityEngine by
clicking the star (Bookmarks menu) in the top-right of the
viewport. CityEngine is also able to export to traditional 3D
formats – including KML, COLLADA, OBJ, 3DS, DXF and
Autodesk – making it easy to bring the model into your
preferred rendering software. Select your area of interest and
go to File>Export Models.

11

Bring more detail

A very useful attribute in this sci-fi rule is the Level of Detail
setting, which can be assigned per-generated model. This
ensures that larger sci-fi cities can continue to generate on
computers that don’t necessarily have a lot of RAM, or that
almost invisible details are not even generated on buildings
located far in the distance. This saves a lot of precious
space and optimises your scene. 12

3D Art & Design 105

 Environment

Use
modular
textures
The Watchtower 2013
Easy-to-follow guides
take you from concept Create a fast and efficient real-time
to the final render environment using modular textures
Artist info

Tor Frick is a 3D generalist, working with everything from

characters and environments to shaders and pipeline design

O
ver the following pages I’ll showcase my workflow
for developing a modular environment. First I’ll go
through the basics, such as blocking out the
environment, analysing what to build and how, and how to
Tor Frick design and create modular textures. I’m using MODO and
Personal portfolio site UDK here, but the workflows aren’t application-specific. I will
www.torfrick.com also briefly show how my modelling pipeline looks, both for
Country Sweden modular textures as well as high-poly objects. Later I will
Software used MODO,
Photoshop, UDK discuss how to assemble the various elements, finalise the
Expertise 3D generalist, with a scene, and how to use shaders and detail meshes to mask the
focus on hard-surface modular nature of the scene. The end result is a scene that is
modelling and environments
easy to create and can be readily expanded upon.

What is a modular texture?

A modular texture is a texture consisting of several different
elements and shapes, which can be combined together and
reused for a number of different assets. Modular textures are
most commonly used for sets of environment assets that do
not require custom detail, but still need more structure than a
tiling texture, such as architectural elements. The more variety
you can create with a modular texture the more powerful it is.
They are most commonly used in real-time game development.

106 3D Art & Design

 Modular textures

Build the blockout

ěũũ
ũ3.6#1ũ22#3ũ
ũăũ+# Block it out and break it down
ěũũ
ũ2#+#!3(.-ũ.$ũ,."4+1ũ
3#7341#2ũ$.1ũ8.41ũ/1.)#!3
ěũũ43.1(+ũ2!1##-2'.32

01 Start with a basic shape I start by blocking out the rough, basic shapes of the model.
There is quite a lot of iteration required on these main shapes until I get something that I
think will read well. Once I have the main shapes locked down I do a blockout of some of the
mid-sized shapes to see what kinds of details I’ll need, such as the antenna and entranceway. At
this stage it’s also a good idea to get a mental picture of what will be needed from the textures, as
it really helps to plan ahead when producing a modular asset like this.

01

02
02 Divide modular
and unique areas
With the blockout of the
model complete, I visualise
what materials I’ll be using
and which need to be unique.
For the base of the tower I
want to use concrete, while
the rest is metal. The majority
of the tower will be easy to
create with metal trims and
plates, but some of the more
complex areas will benefit
from unique textures, such as
the antenna and the rounded
shapes, as well as the door for
the entrance. I colour-code

03 Analyse what’s needed in modular textures

Some of the materials have a clear separation, such as
the concrete foundation, which is an isolated part, so having the
the segments in the blockout
for an easy overview.

concrete all by itself will work fine. For the metal surfaces it’s a
bit harder. The trick is to distribute the details and features
correctly between the textures, so that you can get as much as
possible out of them. The metal plates for the shell are also
easy to separate into their own texture. Here I’ll divide the rest
of the metal into two textures, one for metal plating and one for
trims and details. The details will be baked down from
high-poly to low-poly, but since I will be using them with a
03
modular texture, I’ll put them on the same sheet.

Quick hard-surface sculpting

In order to make the textures pop a bit more, without introducing too many additional steps
in my pipeline, I use MODO’s sculpting tools to add basic detail sculpting on the hard-
surface textures. They don’t offer the same power as ZBrush or 3D-Coat, but since the tools
are non-destructive and work on top of your Sub-D mesh, they’re great for a quick damage
pass. I use it a lot for features such as complex machinery, where you don’t want to bring the
whole thing over to a sculpting app but still want to achieve a slightly worn surface.

3D Art & Design 107

 Environment

Create the main tower

Construct the various elements and assemble in UDK

04 Work on the modular textures For every

modular texture, I make sure that I have all the areas I
need covered by that texture and lay it out efficiently. For the
05 Start building
unique elements
The main reason for including
concrete I just divide it up into different square pieces that will discrete, unique elements is
work with most shapes. The same goes for the metal plates. they should be areas that
For the hull I make sure that I fit a small trim in for the side can’t be made quite so well
profile of the plates. It’s important to tile the trim/detail texture with modular textures. As
efficiently, since trims tend to be quite long. It’s also a good idea such I make sure I use shapes
to reserve some extra space in case you feel you need to and details that require
include additional detail later on. custom textures. I use a lot of
floaters when I build
04 high-polys and even try to get
away with as little subdivision
modelling as I can. I mirror all
the unique stuff as
aggressively as possible,
since I plan to cover that with
vertex painting later, saving
05
both time and texture space.

06 Texture the unique pieces The textures for the

unique elements are basic, so I keep them relatively
clean, since I’m adding extra paint and rust in the shader. I want
07

a uniform and generic metal to be then mirrored and reused, so

I use the same template for all the unique textures. I mostly add
some small text and decals and add the texture as a base for an
07 Map the modular textures When creating a modular environment, a lot of the time
will be spent making UVs. I try to save as much time as possible. For instance I might
unwrap one of the plates first, then as I copy them out to create the hull I’ll shift the UVs around
overlay normal on top of the baked one. so that I get good usage out of the texture. The same goes for circular shapes. I unwrap one circle
first and cut that into pieces, so all the UVs are almost finished from the start.
06

08 Assemble the tower Now I assemble the main

tower in pieces per material, instead of one object with
multiple materials, so that it will be faster to update the individual
parts without breaking the rest. Additional small parts that can be
used several times are added as separate objects, so I can reuse
08
them with greater ease and give them unique shaders.

108 3D Art & Design

 Modular textures

Artist
Showcase
Tor Frick
I am 28-year-old 3D artist, currently residing in
Uppsala, Sweden. I am a 3D generalist, as there is
no single aspect of making 3D that I enjoy more
than the other, which is why I have chosen to do a
bit of everything.

09

09 Make the shaders I opt for the relatively expensive shaders, going for looks and speed
more than pure performance. For all the metal, I go with a pretty complex vertex-blending
material, where I can scrub off the paint and add rust with two separate channels. No specific blend
masks are needed for this, I just use the Diffuse texture, since the ambient occlusion and rust in them
work as a natural blend. The last thing to be added is a cracked Normal/Blend texture to make the
paint-flaking more realistic. I also experiment with the parameters as much as I can so that the shader
can be more controllable with instancing.

Baroque mansion, MODO, UDK, Photoshop (2013)

Part of a real-time baroque mansion I made in UDK. The
project started out as a really small baroque corner and then
just exploded into an entire scene.

Modular Lab MODO, UDK, Photoshop (2011)

This is a still shot from an experiment in modular textures,
where I created an entire environment using only one or two
256 x 512 textures.

10

10 Paint the tower Since all the materials are based on the same master shader, I can
easily tweak colours in real-time until I find something I’m happy with. After tweaking the
colours I start using the vertex painting to scrape off paint in natural locations and add rust in
areas where you might expect it. Make sure you paint all the mirrored surfaces to distinguish
them as much as possible from one another. You can get away with a lot of very aggressive reuse
of textures as long as you hide it with clever UVs and mask the different variants.

Modularity for speed

Don’t be afraid to overuse modular textures. It’s better to use them too much and then
Datacore MODO, UDK, Photoshop (2011)
replace the parts that don’t work with unique ones, rather than wasting effort building unique
This is a small environment I built using mainly modular
assets that later turn out to be a waste of time because they were not actually that important.
textures in UDK. I wanted to go for a cleaner look. In total the
Using modular textures is fast and the time you save enables you to give more attention the
scene uses 69,000 triangles.
big picture and the parts that actually matter.

3D Art & Design 109

 Environment

Add finer details

Decals, vegetation and rock will add realism

11 Add decals This scene is pretty sparse when it comes

to decals, since the majority of the environment is
organic. I’ll use two decals, one giant number for the tower and
a simple rust decal, which can also be created using shaders if
you’re willing to put in the work. Decals are a great way to bring
in a sense of narrative to the environment, but in a practical
sense they can also be used to cover up unnatural-looking
surfaces or hide repetition.

12 Build out the ground area For the next stage I can
use my blockout as a base and clean it up a bit, redoing
the topology so I have an even distribution of polygons to help
with the vertex-blending of the ground textures. Without a
good topology it’s all too easy to end up with unnatural-looking
blends if you’re not careful. For this reason I made a quick
photo-based rock texture and one for the grass as well, which I
can then blend between. The same texture can be used to
create a cliff mesh and a couple of quick rocks. Since I’m
applying the same texture this ensures that the transition
between the rocks and the ground goes relatively unnoticed,
which is essential for a believable outcome.

Can’t build it? Hide it!

Very often there is one part of a model or a texture that I’m
not really happy with and I don’t know how to fix it. Rather
than dwelling on this and wasting a lot of time trying to find
11
a way to make impossible shapes meet, I simply try to hide
it from view. If you can take something that doesn’t look
good and then save time by hiding it behind something you
12
know already works, why not do it? In the world of game
development it’s quite often the case that you will need to
include things within a strict time limit, so everything that
can free up more time for overall quality is great. Of course,
if you have the time, feel free to do things the hard way.

13

13 Include vegetation I want to enforce the feeling that

the tower has been abandoned, so a bit of overgrowth
always works. For this I’ll make two or three simple ivy meshes
and add them to places where it would make sense for them to
grow, such as crevices and surfaces where they could find
some grip. I also built some very basic Alpha planes for grass
and some dry weeds that can be used to further entrench the
tower in the scene, as well as to fill in some of the transition
areas between the rock and ground.

110 3D Art & Design

 Modular textures

Polish the scene

Make the final tweaks and backdrop to the scene

14 Use a simple lighting setup The lighting setup for this scene is very simple, as I only
want to use a single dominant light for the Sun and then some basic spotlights to
highlight some of the shapes and ivy clusters. In general, I feel that the larger the scene is, the
less there is a requirement for attention to detail. This means I can focus more on the big picture,
rather than spend a lot of time ensuring every single small asset really pops.

15 Add a backdrop When constructing the backdrop I make sure to have a separate
window open at all times with a view of the shot of the tower. This way I can make certain
that I work on the right elements and don’t get bogged down working on details that won’t even
be seen in the final render. In this case I’ll reuse the same rocks that I used up-close and just
14
rotate and scale them until I get the shapes I need. The only new asset I’ll make here is a simple
sloped ground mesh, for covering larger pieces of ground. You can recycle the same tower a
couple of times in the background, to give a greater sense of scale and distance. 15

16 Move to post-production This scene is very light on effects and post-processing. I

used colour correction to make the colours pop a bit more. The original plan was to have
it more sombre and desaturated, but the saturated look works much better. All the clouds are
created by making simple billboards with an Additive Cloud texture placed on them.

16

17

With hindsight
Taking a look at the scene afterwards, there are several things that I realise should have been
done differently. For example, several of the elements of the unique assets could have either
17 Final scene tweaks In this final stage I take a step
back from the scene and make final tweaks. I always do
a few more passes than I think is needed. It’s easy to feel like
taken more advantage of the unique texture space, or have been converted to modular you are done with something, but taking a few extra rounds of
parts. Also, I could have either merged my two metal plate/trim textures, or extracted more polish is almost always worth it. Here I added some more ivy
unique parts out of them. Like I’ve previously mentioned, never be afraid to go too modular,
meshes, tweaked the lighting a bit, and ironed out some of the
you never know what it could do for your scene!
rough edges on the rocks by adding more detailed meshes.

3D Art & Design 111

 Environment

Easy-to-follow guides
take you from concept
to the final render Create a
videogame
Artist info

Stefano Tsai

Username: StefanoTsai
environment
Personal portfolio site
www.stefanotsai.idv.tw
Location UK
Engine room 2013
Software used 3ds Max
Expertise Speed modelling to Use 3ds Max to quickly visualise a concept and bring the
quickly visualise ideas,
particularly with regards to white box into higher-resolution models for pre-vis
mechanical design
Stefano Tsai works for The Creative Assembly, SEGA as a senior artist

W
hen creating videogame focus on modelling, forming the various
environments, developers need assets and quickly presenting them
to quickly take concept ideas effectively in the 3ds Max viewport.
and present them as 3D pre-visualisations The final steps of the process will cover
to ensure they will work in the game. lighting in the viewport. 3ds Max’s viewport
Here we’re going to show you how to lighting and ambient occlusion support
quickly nail down such videogame saves time and makes for a more efficient
environment ideas, from the concept workflow. This is a straightforward stage for
through to the initial 3D pre-vis models. In a 3D pre-vis artist, but also simple enough
the first steps we will show how to analyse for a novice to understand. We’ll conclude
the scene, then plan out what to model and with some ideas on how you can pitch your
how best to model it. Following this, we will environment based on this design.

112 3D Art & Design

 Videogame environment

Concept
Here you can see a quick concept of an engine room. This
ěũũ
ũăũ+#2ũ$.1ũ#!'ũ23#/ũ room was built into the rock to give the inhabitants some
.$ũ3#$-.Ħ2ũ/1.!#22 natural protection. They used lasers to cut through the rock,
then installed the reactor engine inside, which is why the
walls are so flat.

3D Art & Design 113

 Environment

Analyse the scene

Artist
Showcase
Stefano Tsai
I’m a concept designer and 3D artist. I have been
Apply colours and basic shapes

Blocking out
working in the game and entertainment industries
since 2001. Even though I have spent over ten years
the key
working within the CG arena, I still feel so excited elements
about future projects. I am current working on a new
alien IP at The Creative Assembly Ltd, SEGA. At this stage we’re going
to create the first blocking
of the models and form all
the key elements. With
01 the first blocking models
in, we can start to
visualise the proper size
01 Establish a design Don’t rush to building something
straight away, just spend time talking to game
designers to see what the original idea and story behind this
and shape, then put them
into the correct place. The
most important thing for
level is. The most important thing is to find out which areas the this stage is to keep things
as simple as possible.
player is going to interact with most. This could be with NPCs We’ll only create the basic
(non-playable characters), just the machines or even a shapes so that people can
computer’s monitor. instantly understand what
their function is. For
example, if at this stage I
want a plumbing system,
Corridor Study 3ds Max (2013) I won’t build gate valves,
I used simple shapes and materials to form one section of corridor. miniature valves,
This is the realistic version, as I tried to match the real world as draincocks, jumpers,
much as possible. nuts, washers and so on.
I’ll only model basic pipes
using 3ds Max’s Spline
tool and turn on Enable In
Viewport from Rendering,
then play around with the
02 Radial setting.

02 Colour the white box Use different colours to aid with your thinking when separating
the primary, secondary, path and background elements. The primary and secondary
elements are key for the environment, so everything else should be built around and support
them. Here we’ve used yellow as the primary elements, orange as the secondary and blue as the
path for the players. This is the quickest way to understand the space of the environment and the
relationships of its various components.

03
Corridor Study 3ds Max (2013)
Here is another study of a mechanical corridor, but this time in a
more steampunk style.

03 Clear up all the important spots for NPCs or story There’s nothing more
important than understanding story and how players react with NPCs in the
environment. It’s hugely beneficial to place these into the 3D scene so you can see all of their
Car Study 3ds Max (2013) relationships with the space surrounding them. Here you can see we’ve put down all of the NPC
A study of a futuristic car’s front panel, here tested with two locations, game save spots and medical kits, in order to figure out where the focal point for the
different tones. scene will be. This information is going to really help when lighting later, as it can be used to lead
players into the specific areas you want them to go.

114 3D Art & Design

 Videogame environment

Add simple details

Make the scene light and efficient. Don’t overcomplicate it

04

04 Use primary elements There are many similar

objects in the scene, so here we used an instance copy
for those models to save time. For example, when you’ve finished
one model, you can duplicate it multiple times in the scene. When
you modify one model, you can see all of its children change at the
same time. This is the easiest way to read their relationships with 05

one another, as well as with the space around them. For this stage,
it’s best to break the basic shapes and make them more compact.
The key is to integrate all of the basic shapes you get into the scene,
so people can’t tell that you’re using simple boxes and spheres.
05 Build up the secondary elements Treat this stage as a chance to extend the
primary elements. Imagine that you’re the engineer and need to work out the way to
connect these heavy-duty machines. You can pick a few features from the primary elements and
then use the same language to put them together. This can be approached structure-wise or you
can extend some support units. For example, if you have a big engine unit, you should have a
generator and batteries to make the system work. After you’ve built all the elements you want,
you can model the structure to support their weight and place them into the environment. Here
we’ve used a steel hull to hold them all together.

06

06 Build a pathway Focusing more on the actual

gameplay for a moment, you need to ensure that the
pathway is large enough for players to walk through. In
contemporary games, this can mean two or three people
walking through together. Normally this is wider if you compare
it to a real-world scale. Here we can use the information taken
from Step 3 to form the path through the engines. For creating
these platforms, try using the Shell Modifier to minimise the
work required. Don’t spend too much time on modelling every
metal frame, simply model from a plane and let 3ds Max 07

generate the thickness for you. You don’t need to do everything

by hand, just find the easiest way to save time.
07 Use the viewport lights to review Different colours and tones can tell a story all by
themselves. For example, here bold yellows, reds and blues have been used for lighting.
Place yellow Omni lights from the top and middle to light up the primary elements. Try to think
Choose a focal area about the real source of the lights. Here we can presume there are some lamp stands near the
machines to help workers and strong lights from the ceiling to illuminate those ceiling machines.
This is a pre-vis scene, meaning you can’t include However, remember that if you show everything, you’re really showing nothing, so don’t be afraid
absolutely all of the details that you want. You only have
limited time, so you have to pick certain elements that are of darkness. You need a contrast of lights. Here we’ve left some darkness between the main
most in need of further development, especially in the focal platform and the lower deck, then placed some red Omni lights. Red gives off a dangerous mood,
areas. Again, you need to keep it as simple as possible, as we want players to feel as though there’s something even more dangerous than mechanical
while at the same time keeping all of the modifier stack, if machines down there. This is the best way to give players another layer that they can see and
you adjust the models. You can add another editable poly feel but can’t reach. Blue is often used because it has a more artificial feeling. This could make
on the top of your stack list, as this way, if things go wrong,
you can always go back to a previous step. players feel as though there are computer monitors nearby. This hints to the player that there is a
control room above this area, for instance, which could be their next destination.

3D Art & Design 115

 Environment

Refine the environment

Focus on the smaller points and evoke realism

08 Refine the background In this particular scene,

the engine room is sitting inside a bed of rock. To
make this believable, we need to have real material on the
surface, not just flat and box-like surfaces. Here we’ve used the
Soft Selection options with an editable poly to quickly adjust
the faces to look more like rock.

09 Finish the blocking details With the viewport

Omni lights and the background set up, we now need
to finish blocking out our models. Here we’ve added some huge
pillars along the rocky surface to establish some buffer
between the rock and steel materials. This can be done after
completing the first pass of lights, because it’s now clear where
the darker areas are going to be and, even more importantly,
where the player will be moving. We don’t need to put too
much detail into the areas containing darkness, or spend time
on something that the player can’t see or reach. Darkness can
08
be a great tool in this regard.

09 10

11

10 Develop the main reactor If you keep all of the

basic shapes from the original cylinder, it will make the
object very boring. As such, take the top area and open four
doors for the cooling system’s pipes to release some hot air, but
still keep the side of the cylinder simple, only adding some
frames to make it a bit more heavy-duty. Keeping the middle
surface clean is a big benefit later. You don’t want to overdo the
models, as the clean space can be used for decals and textures.

11 Shape the turbine In the same vein as the previous

step, you need to stick to the main shape of the object
and develop around it. Here you should complete a bit of
research on jet engines and select a few features to include in
your design. Some turbine blades have been added along the
surface, with some inset to provide depth for the objects.

116 3D Art & Design

 Videogame environment

Use lighting effectively

Add final details to the scene to make it come alive
12
12 Further detail and scale We have built a lot of large and impressive machinery here,
but this is potentially a bad thing, as the player may not feel as though they belong. The
main reason for this is because they don’t have anything to interact with in the environment. You
need to provide some small details to make players feel as if they’re part of the world; they need
to feel they can interact with the environment and feel that the scene was built around a human
being. You could include a computer monitor or lamps; anything that the player feels they can
reach out and touch. This helps promote a sense of scale and sells the reality of the scene.

13

13 Make viewport lighting dramatic Drawing the

scenery is just like using charcoal pencil, as you can
wipe away some details for better balance. After bringing in
more details for the models, we need to return to adjust the
lights. Don’t be scared to add more contrast, because if you
give the same value to everything, your final scene will be
boring. Tweak the Omni lights to create a more exciting mood.

14 Insert decals Now follow the concept to add some

neon lights to enrich the scene and bring more colour
into it. Include more Omni lights around the big blue neon
structures in the background to let its blue colours fade into the
scene. The same goes for the yellow neon lights on the
reactors, which help to bring in more of a sci-fi mood.

Play around
14
15 Post-production in Photoshop 3ds Max currently
can’t handle visual effects in the viewport, so for these
and show you’ll have to use a script to grab a high-resolution screenshot.
You can then bring this into Photoshop to add some further
opportunity details such as smoke, volume light and other small details.
to your lead These are the final touches for the scene and can bring the
image a bit closer to the final look, as well as help sell the mood.
Don’t forget to show If possible, you can normally use another angle to give a team
multiple versions of your leader a little more perspective.
concept. You can easily
drop another version of a
reactor, a turbine
generator or even play
around with the
15

8 hours
creation time
composition of the
elements. Most of the Resolution:
time your first version
isn’t your best, so you’ll 6,000 x 3,840
need to adjust it. It can
make huge difference
spending just another 30
minutes moving things
around or changing their
proportions. You don’t
need to model your scene
precisely. This process is
more about putting down
ideas and developing
them with your lead than
making the perfect
models. It’s meant to be
quick and rough; more
like a pencil sketch than
the final inking.

3D Art & Design 117

 Environment

Easy-to-follow guides
take you from concept Build matte
paintings
to the final render
Artist info

Dark Hoffman Neverland 2013

Personal portfolio site
www.darkgrafix.com Here we will look at how Vue and its
Location USA
Software used Vue Infinite, various options can be used to create
Photoshop
Expertise Concept art, matte paintings for film or TV
designing props, environments,
matte paintings
Dark Hoffman has been a digital matte painter in the
film and commercial industries for over a decade

118 3D Art & Design

 Matte paintings

I
’ve used all different types of software for my digital
Concept matte paintings, but Vue is the go-to program that I
use when I can’t find the right landscape, mountains,
The image of Neverland is skies or foliage.
ěũũ43.1(+ũ2!1##-2'.32 an iconic one used in film,
television and on the In the following steps I will explain how I use Vue for
stage. The purpose of this matte painting projects in the film and commercial
exercise is to create a look industries. I’ll reveal how to create terrains, a calm ocean
for the island that can be and an extended landscape including Metablobs.
used as a matte painting
for a blockbuster or a TV The outcome could be used as an establishing shot of an
series. First we will create island to show where the next scene takes place. Usually
the look of the island in these are pretty shots. As such, after the scene is rendered
Vue and then take it into fully in Vue, I will take it into Photoshop CS6 and tweak the
Photoshop for post- colour, make a Levels adjustment and add any small
production work.
finishing touches to make it ready for the final shot.

3D Art & Design 119

 Environment

Design the layout

Establish the broad shapes of the new island

01 Create a new world The first thing that is needed to

create an environment is an idea. Usually the
production designer gives the art team the proper artwork to
01

show what the world looks like. Colour, layout and a theme are
very important to start a matte painting. Also we are normally
given a plate or a frame of the film to start with, but sometimes
we have to start from scratch, which is what we will be doing
here in Vue. We’ll lay in terrains from the Vue Library and edit
them so they represent islands and mountains.

02 Find reference To make an impressive island you

need to know what it looks like in terms of shape,
texture and colour. You also need to decide what the look of the
environment is, which includes atmosphere, time of day and
where the ice will be placed in the scene. In our case, the script 02
calls for ‘an overall aerial view of Neverland as we fly towards it
through the clouds. It’s a tropical-looking island with coves,
waterfalls and a snow-covered mountain in the centre. Captain
Hook’s ship sits silently in Cannibal Cove.’ We can start
creating the terrains by bringing in set terrains from the Vue
library. We will then start laying out the central mountain and
background with the terrain tools.

03 Set up the building blocks To help with the render

time and to work with the scene in an easier fashion,
we will open a default scene before we add an atmosphere. In
our scene we’ll start by bringing in a simple procedural terrain.

03

120 3D Art & Design

 Matte paintings

Set the scene

Unlock the beauty of Vue’s evolution

04 Work with Hyperblobs The next stage in the new

version of Vue is building with Hyperblobs and
Hyperterrains. You can stack these in any formation and any
configuration of simple geometrical shapes and then simply select
the Hyperblob tool. Pick the right Bump map and you have yourself
a realistic-looking rock structure. Any material can be added for an
even more photorealistic feel. You don’t have to follow this shape
exactly to build the rocks, but make as close a match as you can.

05 Group the shapes With the structure of the rock

built, hit the Hyperblob button. This will group your
shapes and create a blobby shape, one that looks closer to an
actual rock shape. Double-click the Material Editor and hit the
Bump tab. Select the ball and another box will open, which offers
options relating to which Bump map you can use. Here we’ll use
Complex Sedimentary Layers, although you can try all the options
if you wish. With your Bump map applied, change the scale so the
04
arch appears larger and closer to the scale we need.

06 Shape the mountain Start adding terrain from the Terrain Library. When you’ve
brought one into your scene, double-click the object and a new window will open where
you can edit the terrain model. Click the left-hand buttons, jumping from dunes to mountains
05

until you get a good shape. In the EcoSystem editor I also added three default plants – date palm
(25% Presence) old eucalyptus tree (25% Presence) and a coconut tree (50% Presence).

07 Apply materials The centre of Neverland has a mountain with snow that goes down
to grass and rock using a mixed material. You can play with which material you want to
be used most. Here I made one island that I then used for various areas of the main island,
meaning I didn’t have to populate the main terrain with an EcoSystem and slow down the render.
I named this island ‘Neverland Cove’ and saved it as a separate scene so when I need it, I can
merge it into the main scene. Be sure to delete the Sun and ground that won’t be needed.

06

07

Painting with plants

If there are certain areas that need to be populated with plants, you can just paint them in.
Painting the flora will create a separate layer so it won’t be attached to the terrain. Hit the Eco
Paint button and choose which plants you want to paint. You can change the size of the brush
and the scale of the plants to suit. The brush paints directly on any areas of the terrain, no
matter its shape. This is a great way to add exactly what you need, where you need it.

3D Art & Design 121

 Environment

The ocean and beach

Build up a tranquil ocean and a sandy shore
08
08 Include MetaWater and scenic beaches
Adding an ocean is just a click away in Vue and
editing it is now even easier. Essentially we want shoreline
foam, so start by selecting a MetaWater with Foam. To add a
beach, select a simple terrain and flatten it, then add a sand
material and make sure it comes through just enough to form a
beach before sinking into the water. You can adjust how much
of a beach you want, but in this example we just want the most
sandy areas around the cove. Duplicate this and add it around
the perimeter of the island where you want to see beaches and
have them trail out into the ocean, so there’s some overlap. This
may take a bit of adjusting, but the results are fantastic and will
really help to sell the realism of the scene.

09 Add planets This is probably the easiest part of the

process and a lot of fun. On the left of the interface, in
the menu bar, you’ll find a thumbnail of Saturn. This is the
Planets button. Click it and a planet will be added into the sky
automatically. To edit this, go to the top-right of the screen and
you’ll find a drop-down menu from which to choose the planet
you want, as well as its phase, colour and brightness.

10 Render time Now we’re into the final stage of the

Vue process. To keep the render time down to a
minimum, turn off anything you don’t really need, as well as
anything that can be completed in post-production using
Photoshop. Try to maintain a good balance between keeping
render times down, while hanging on to everything you need.

09

10

122 3D Art & Design

 Matte paintings

Final touches
Input the finishing elements and move to post-production

11
12

11 Add Hook’s ship For the pirate ship I added a model of

The Golden Hind from Poser. You can use any ship from
Cornucopia 3D or Poser, or any 3D model you have access to. I 12 Take the image into Photoshop Now you can bring in the image rendered out from Vue
and do some retouching in Photoshop. First let’s play with the levels to bring some of the blacks
brought it in with textures and baked it, so it was compressed, then in and help with the contrast. Desaturate the image a little, about 25%, then add a bit of colour back,
placed it in the cove. This is the only element in the scene that mainly the blues. Following this, add a dark vignette around the image to add some depth and make it
wasn’t created in Vue. It can also be added later with Photoshop. feel like it’s out in another universe. All this will help sell the fantasy element of your image.

13
40hours
creation time
13 Place more elements Since we can see planets
surrounding the island, we also need stars in the sky,
which will be reflected in the ocean. I have a personal collection
Resolution:
5,500 x 3,700
of starfields, but you can grab plenty for free online or paint
them in yourself. I recommend you download a few and try
them out. Pop them in the file and screen, then add a layer
mask and paint out any stars over the island and the planets.

14 Final details For the final touches I added foam

around the shores, as well as little waves. Vue can add
these to the scene for you, but since our camera angle is so
high it would be hard to see. I also added some small waterfalls
around the island as a nice finishing touch. There you have it: a
14
realistic image of Neverland.

3D Art & Design 123

 Environment

Iliya has supplied

settings for his particle
flow, materials, lighting
and renders

124 3D Art & Design

 Underwater renders

Create
underwater Easy-to-follow guides
take you from concept
to the final render

Artist info
renders
Diving with sharks 2013 Iliya Atanasov
Professional portfolio site
www.pixelhunters.com
I wanted to re-create a scary diving moment, Country UAE
Software used ZBrush, Maya,
when you are surrounded by a huge school of fish V-Ray, Digital Fusion
Expertise I am a CG generalist
and a great white shark bursts through them with a strong sense of the
details that make a difference
Iliya Atanasov is studio director and lead artist at Pixelhunters

H
ere we’ll be working through the process of
creating a dramatic underwater scene, complete
with a ferocious shark at the centre. We will
Concept
explore tips for modelling all elements in the scene using The idea behind the image is to
ZBrush and Maya. We will then focus on texturing these re-create one of those moments
elements, as well as how to create an effective light rig using when a diver can get too close
to a huge monster: the great
V-Ray in Maya. Next we’ll render a number of different white shark. I wanted to achieve
passes, including Color, Specular, Reflection, MaterialID and a scary, photoreal look, but at
Z-Depth. These elements will then be assembled with the the same time to have this
final image to add life to it. To finish, we’ll play with different dreamy stylised feel, as it’s also
heart-stopping moment to swim
colour corrections to achieve the final look. The modelling of
with a shark.
the shark was completed by Lei Zheng.

3D Art & Design 125

 Environment

Establish your base mesh

Move between Maya and ZBrush via GoZ to build an initial shape

01 Sculpt the body Divide the base mesh several times,

using the Move tool and Clay brush to define the shape.
Move from low-res subdivisions to higher-res subdivisions to
01

change the larger forms to smaller details. Try to keep the

sculpt on the lower subdivisions as much as possible.

Why model the whole shark?

The main aim of the image is to re-create the feeling of
being right next to a huge monster, which will be stronger if
we’re closer to the mouth of the shark. Developing the full
body with the tail will help if we decide to change the
composition and can see more of the model’s interior.

02 Detail the face We should pay a lot of attention to

the face, especially the mouth area and teeth, as they
will be the largest parts in the final work. Make the brush
slightly smaller than the previous steps to help add very fine
details. Go a step down to the low-res level and reshape the tail
a little to conform to the final render pose.

03 Model small fish and other elements The idea

here is to model a low-poly object with a few
variations on the tail and textures, which can later be used for 02
emitting instancers. Take a sphere with eight spans, extrude/
move the vertexes back to form the body and make a couple of
variations to be smoothed later. As we’re dealing with a huge
amount of fish, multiplying them would take too long, so add an
emitter to produce particles that will follow a path. Later you
can add in your fish models to take the place of the particles.
For the corals, use a low-poly sphere primitive from Maya with
30 spans and delete the half that won’t be visible to save
memory. In the Polygons tab, go to the Sculpt Geometry tool to
reshape the rock. You can also make loopable textures in
Photoshop from real coral images. Include other types of corals
and creatures, use the General Editors>Visor>Paint Effects and
convert them to polygon objects.

ZBrush tips
There are a whole range of ways
you can customise your ZBrush
interface. Gaining access to the
most commonly used tools will
greatly speed up your workflow.
When working with Move tools or
some large-radius brushes, the
various Automask settings are very
useful to keep brushes on the area
you want affected, you can find
these settings in the brush’s Auto
01 Use ZBrush’s Move and Clay
Masking menu. Also, assigning the brushes to define the body of
brushes you use frequently to the shark
numbers on your keyboard is a
02 The face is the most vital
must. Simply hit Cmd/Ctrl+Opt/ aspect of the model
Alt on any brush tool, then press
03 the shortcut you want it be. 03 Apply Maya primitives to
model the various elements

126 3D Art & Design

 Underwater renders

Work with dynamic elements

Explore Maya’s Dynamics section to bring your scene to life

04 Create curve flow Go to the Effects section in Maya’s Dynamics menu and click Create
Curve Flow. Basically this is directional emitter, where the particles are following a path from
point A to point B. With five sections of control curves, you will be able to make certain points with
wider diameters – where the fish can fill the space more – as well as some areas where the stream is
thinner, so smaller amounts can pass. Set Control Resolution to -6, Manipulator Resolution to -2 and
Emission Rate to - 250 to provide enough particles for the stream to look busy. You can play back and
04
stop at around the 50th frame, so enough fish are populated all along the path of the curve.

05
05 Replace the particles with an instancer Go to
Particles>Instancer (Replacement), select the
low-poly fish objects, as well as the particle shape, and press
Create to replace the particles with the fish models. Set the
Maya Playback Speed to play every frame and press the Play
button to see how the fish will swim through the curve. Now
you can change the diameters of the circles, as well as alter the
positions of the vertex points of the curve to form the desired
path for the composition of the image. With the Flow_
particleShape node selected, go to the General Options and set
Position to World Position. Then under Rotation Options, set
Aim Direction to Velocity. This will secure the school of fish are
facing correctly along the path.

06 Model the plankton in the water A tiny but very

essential part of any underwater scene is the clouds
of plankton that inhabit the oceans. Even on the most clear
days, there are lots of small particles swimming around. You
can use the Particles menu under Dynamics to create an Omni
emitter with a particle cloud shape (Clouds (s/w)). Make the
Radius pretty small, around 0.003. With a rate of 100
particles/sec, you can playback and stop at the 50th frame
again to leave enough plankton particles for the scene.

06
04 Maya’s Dynamics menu is
very intuitive when it comes
to populating a scene
05 Play with general options of
the Instancer to achieve
quality results
06 By using Omni emitters, we
can distribute plankton
particles in the scene
07 Maya Paint Effects is just one
way you can create realistic
bubble elements

07 Make an ocean surface and bubbles One of the

best modules in Maya is Fluids. Go to the Visor tab
and choose one of the Ocean Examples, such as Normal Sea.
This will create an Infinity plane with an Ocean shader attached
to it. You can tweak the settings for the Scale to fit the
measurements of your scene. Later you will need to render the
ocean as a separate pass using mental ray or Maya software,
as V-Ray does not currently support the Maya Ocean shader.
There are two simple ways to produce the bubbles. Firstly, you
can create a Directional emitter with a Spread of 0.058 and a
Rate of 100 particles/sec. For the shape, use Sprites with a
Scale of 0.1. You can use a simple bubble image with an Alpha
channel for the bubble. The second method is to use Maya
Paint Effects, where there are presets for making bubbles in
07
General Editors>Visor>Underwater>Bubbles.

3D Art & Design 127

 Environment

Apply materials and shaders

Prepare the basic and advanced materials for your scene
08
08 Texture the shark model Choose a VRayBlendMtl for the shark and the
VRayFastSSS for the base material. It isn’t compulsory to use an SSS, however, so feel
free to experiment with basic VRayMtls, along with Maya Blinn material. We’re simply using the
VRayFastSSS because it produces the best results for this scene. We’ll intentionally make the
scatter colour very red, as this colour will begin to disappear underwater within a depth of five
meters. If we go to ten meters, there will be no red at all, but because this is an artwork we can
still give it a slightly red tint. For the Diffuse Color we’ll use the texture as it is with a Diffuse
Amount of 0.4, but for Sub-surface Color you can make the texture much more saturated with
clearly saturated reddish parts around the tummy and the mouth. Also try setting the Scatter
Radius to 2.5 to exaggerate the effect a little more.

09 Add Bump and Reflection maps Use a normal Bump map to establish some
thinner details. For the second layer of the VRayBlendMtl, use a VRayMtl with a black
diffuse colour, a small amount of Reflection and a Highlight Glossiness of around 0.846. This
layer will help replicate the feel of wet tissue. For the teeth material, again use an SS shader with a
marble preset, some reddish tint for Sub-surface Color and an orange Scatter tone. Make sure
the Scatter Radius is pretty small. The eyes are easy to create, as they are just reflected black
surfaces with a high reflectivity applied to them.

09

10 Detail the small fish and corals For these parts

we’ll use just a VRayMtl, not an SSS, to save rendering
time. Make the Bump setting big, such as 0.8 or 1, and add
slight reflectivity. Use the same type of material for the corals,
without the reflectivity, and apply a very high Bump setting, as
the corals will be far away, making the bump nearly invisible.

08 You can try other methods, 09 Use Bump and Reflection

but VRayBlendMtls on the maps to get the wet effect
shark model, as well as the
VRayFastSSS for the base 10 Texture the other elements, 10
material, works best such as the school of fish

128 3D Art & Design

 Underwater renders

Light and render

Prepare lights and render passes

11 Move to V-Ray Applying a V-Ray dome light can achieve an overall bluish atmosphere.
Setting a low Intensity of 0.4 can give the feeling of the dimmed underwater mood we’re
looking for. Add two rectangular lights to brighten the shark with bluish tints and a white tint
along both sides of the nose. The top one (Intensity 40) simulates the Sun light over the ocean
surface, while the left one is more like a video light (Intensity 20). Light the corals behind with
four more lights. This makes the tips of the corals much brighter than the lower parts.

12 Adjust the lights Now turn on Global Illumination, putting the Primary bounces on the
Irradiance map and Secondary bounces on the Light cash with relatively small
subdivisions (500). This is because, in this scene, we don’t need such well-defined thin shadows.
If you want to use real caustics, with rays emitting from the rectangular Sun light, this next part
will be tricky. The huge numbers of small fish involved would be almost impossible to render, but
11
there’s a trick. Make a Directional light and project a Procedural Noise texture on the Color
attribute. Now apply a white Lambert material to all the objects to achieve the caustic pass,
which we’ll later use for compositing. Make sure you render the ocean separately from the rest of 12
the scene, as V-Ray currently doesn’t support the Maya Ocean shader.

13 Begin render passes First add extra dirt texture to

simulate some Ambient Occlusion, building up some
shaded areas around the teeth base and inside the mouth. Next
is the Reflection pass, which you can use to make areas around
the gums and inside the mouth more reflective later. The
Specular pass serves much the same purpose as the Relection
13
pass, but will exaggerate the specular levels further.

14
14 Make the final passes The RenderID pass is needed to gain access to the different
parts of the image. If you want to make certain rocks or the teeth a slightly different tint,
you can Chroma Key it, create an Alpha and use this for correcting these parts in the compositing
stage. The Z-Depth pass is used for achieving volume or depth in your image. This will make
some parts of the scene seem foggy, which is very normal for underwater scenes. Visibility in
water normally only extends to between 5 to 20 meters.

15 Apply Sun rays To produce the rays coming from the Sun light, you can use a separate
pass with Spot lights. For the Color and Intensity attributes, apply a Procedural Noise
texture and turn on the Fog setting. Also activate Shadow for some added atmosphere.

15

11 Start setting up the lighting

with a V-Ray dome light and
rectangular lights
12 Effective use of Global
Illumination is the best way to
achieve the right atmosphere
13 A bit of dirt texture will
produce ambient occlusion in
the teeth and mouth area
14 You can produce authentic
depth of field with the
Z-Depth render pass
15 A Spot light can be used to
produce the Sun rays

3D Art & Design 129

 Environment

Unify the scene

Composite your layers and elements Artist
Showcase
Iliya Atanasov
Iliya is a 36-year-old Bulgarian CG artist. He has worked
for TV channels, advertisement agencies, film
companies and videogame studios. He currently works
for Pixelhunters, which is based in Dubai, UAE. The
studio mainly focuses on videogame cinematics.

16

Smilyan House Maya, Digital Fusion (2008)

16 Merge the layers We’ll be using Digital Fusion to composite the different passes.
Create a blue gradient that’s darker at the bottom and brighter at the top. Next add the
water surface and apply a Polygon Mask to it, with large Feather setting. Add the shark, fish and
This scene required some hard-surface modelling, as well as
Maya’s Paint Effects for the lush greenery. IBL lighting in mental ray
was also applied to finish things off.
corals, then multiply these with an AO pass on top and a small Transparency setting. Compose
the caustic layer on Hard Light mode and use Luma keys to take the darks from the Alpha out of
the image. Merge the Reflection pass on-screen with a 0.3 Blend and the Specular pass on 0.3.

Alien Swamp Maya, Digital Fusion (2009)

Another hard-surface modelling piece that again used Maya’s Paint
Effects for the foliage. Lighting has been used in unique ways to
achieve a simultaneously drab and homely scene.

17

17 Use the Z-Depth pass Apply the Channel Boolean tool to add the Z-Depth to the Z
channel, so that you can use the Fog and Depth Blur tools. Apply a couple of fog layers
with different tones to match the lower dark layers of the water. Compose the rays on Screen
mode, as well as all the particles and bubbles. Apply Directional Blur on some parts of the fish
and the tail of the shark to simulate movement. Every diver wants to swim in clear blue water, but
most of the time the ocean is very murky and greenish, so experiment with different levels of
shadow and green contrast to achieve the final look.

More dimensions
16 Move to Digital Fusion to
Maldives Maya, Digital Fusion (2012)
Always try to use a combination of 2D and 3D programs. This made great use of displacement materials and Maya’s Paint
composite the piece together
Hugely beneficial changes can be made in the compositing Effects. The wooden structures and the epic clouds combine to
17 Tweaking the water to be stage, rather than trying to achieve everything in 3D. It can produce a pleasing composition.
murky and forboding will also save time when fine-tuning the results.
finish off the result nicely

130 3D Art & Design

 Gallery
Artist info

Zhang Chen
Zhang is a 27-year-old student
from China. He has been a 3D
lecturer in Beijing for six years
Personal portfolio site
www.cg-zhang.cghub.com
Country China
Software used Maya, ZBrush,
mental ray, Photoshop

Work in progress…

This work took me one month to complete. I

attempted to portray the desolation and loneliness
of a city without the usual dark shades and tones. I
adopted the bright sunlight of noon, which
enables the rich colours to shine through
Zhang Chen Seedy City, 2011
3D Art & Design131
 Vehicle
From complete cars to fighter jets, have some
fun with 3D vehicles
134 Model a complete car
A showroom render of a classic roadster:
141
the AC Cobra

141 Gallery
No More Gas by
Colie Wertz

142 Build a vehicle game asset

Create and prepare a vehicle object for
use in a videogame production

149 Gallery
AEG27Cern 05 by
Luigi Memola

150 Create Bunkspeed renders

Showcase a sci-fi vehicle concept in the
shortest possible timeframe

154 Model a fighter jet

Build a dramatic scene with a realistic
fighter jet being pursued by the enemy

161 Gallery
Waldo by
Moran Tennenbaum 134

154 149

132 3D Art & Design

142 150

3D Art & Design 133

 Vehicle

Easy-to-follow guides
take you from concept Model a
complete car
to the final render
Artist info

Aldo Vicente
’67 Shelby Cobra 2014
A showroom render of a classic roadster: the AC Cobra,
Username: AldoVicenteCG popularly known in the United States as the Shelby Cobra
Personal portfolio site Aldo Vicente works as a 3D artist in marketing and freelance
AldoVicenteCG.Wordpress.com
Country USA
Software used Maya, mental
ray, Photoshop
Expertise Modelling, textures,
materials and lighting

Concept
I wanted to make an
image that looked like a
car advertisement,
striving for a certain
level of realism balanced
against a focus on visual
interest and appeal.

134 3D Art & Design

 Model a complete car

ěũ731ĴććĈĴ4++ ."#+
ěũ43.1(+ũ2!1##-2'.32

I
f you want to learn the process of modelling a complete car in Maya, building realistic
materials and lighting, then rendering out passes to composite and enhance a final
image, you’re in the right place. Here we’ll look at how to deconstruct and then
reconstruct a complex vehicle body quickly and accurately. This super-efficient modelling
method, combined with mental ray’s powerful ––––tools, make this pipeline surprisingly fast
and very fun. We’ll be using the ’67 Shelby Cobra as our subject, but most of this tutorial can
be applied to any type of car, so feel free to follow along with your favourite model. We’ll be
covering a lot of features and methods, but won’t always have room to cover every detail in
each one, so I always encourage exploring and experimenting beyond the bounds of the
steps. The tools and techniques used here are some of my favourite, but I’ve seen a lot of
different methods used to achieve some great results. Remember that it’s all about
exploring, learning and having fun to find the workflow that makes the most sense to you.

3D Art & Design 135

 Vehicle

Make preparations About line

structure
Gather blueprints and set up the scene Structural lines can be
seen on any modern car
body design, though they
01 Collect reference We’ll need two types of reference,
starting with the turnarounds for the image planes,
which will be our primary guides when creating our model. A
02 Start your scene in Maya To avoid distortion
when making image planes, the dimensions of each
plane should be based on the resolution of the image it will be
can sometimes have
different names. When
modelling any car it’s
great source for free vehicle turnarounds at excellent quality is displaying. Once the textures have been applied and the image important to begin by
The-Blueprints.com. Next we need an abundance of photo planes are lined up, just resize them to a real-world scale. You establishing these lines
because they act as
reference from many different angles and in a variety of can find the real-world dimensions of this Cobra online, so set markers for the body’s
environments. In the modelling phase, these will be helpful to Maya’s working units to inches and create a box with the same key plane changes. By
fill in the perspective details that aren’t clear in the turnarounds. dimensions as the car. Rescale the references to match the establishing this
These images will also give us a great understanding of the values of the box and create a display layer for the planes. framework early on, we
material properties of our surfaces. are left only with the
simple task of polishing
the spaces in between.
01 For an apprentice
modeller, I should note
that the Shelby Cobra is a
racer with a pretty unique
line structure and
particularly complex
curvature. I’ve found
more conventional car
bodies are typically much
02 easier to practise with.

03

03 Establish structure and the car design Before

we start modelling, we’ll identify the important lines
that make up the body’s design. These will give us a great
starting framework for the body geometry. We’ll be working
with all the lines noted in the adjoining image, but let’s define
some of them here. The A-line runs over the car, tracing the
silhouette from the front of the hood to the back of the trunk.
The belt line runs from the headlight, through the bottom edge
of the windows. The base line runs along the bottom of the car,
from the very front to the very back.

Begin modelling 04

Establish the main forms of the car

04 Line up in the Side view Create a box with two vertical subdivisions, two horizontal
subdivisions and one depth subdivision. In the Side view, the bottom edge of the box
should line up with the base line of the car; the middle edge loop should line up with the shoulder
line and the top edge should line up with the belt line of the car. At this stage simplicity is key, so
strive to establish these lines accurately, but with a conservative use of vertical edge loops. Once
these edges have been created in the Side view, just line them up in the Top view.

05 Adjust for the Top view In the Top view, the centre edge loop is the A-line, so delete
the right-side geometry and work on the left side. Create an edge loop to make the hood
and the deck crease and then make another edge loop to establish the rim line. We can line these
edges up in the Top view as we create them. Make an edge for the windshield line and one for the
deck line. With the lines in the Top view done, finish lining up in the Side and Perspective views.

05

136 3D Art & Design

 Model a complete car

06 Tweak the front and back For the front, we can use the edge loop from the shoulder
line to make the top lip of the grill, then we’ll create a second edge loop to create the lower lip
of the grill. This new edge loop will also serve as the rocker panel line on the sides of the body, as well
as the bottom rim of the trunk in the back of the body. Finally we can address the subtle taper that we
see in the Front view, where the car is wider at the shoulder line and more narrow at the base.

07 Model the
Shelby’s fenders
For the next step we need to
create a cylinder with 20 axis
subdivisions. Rotate it onto its
side, delete the cap faces and
rescale it to line up in the Side
view. If we delete the eight
faces along the bottom of the
cylinder, we’re left with a
12-face arc. Line this up in the
06
Top view, using the two ends to
establish the inner and outer
rims of the fender. Next we
need to merge the fender by
adding four strategically placed
edge loops to the body
geometry and deleting any
unnecessary faces. Repeat this
07
process for the rear fender.

08 Refine the body With the key lines established, we can continue adding edge loops
to help refine the geometry. First make any edges needed to extract the doors, hood,
trunk and grills. Next we can focus on rounding out the surfaces between key lines. Maya’s
Sculpt Geometry tool is our friend in this step, but remember to keep it simple. Ultimately we’ll be
smoothing our geometry to get the finished roundness, so we only need enough mesh density to
get our model looking polished in Preview Smooth mode.
08

09 Work on the interior The interior is comprised of a

few large components and lots of small ones, all of
which can be simplified and re-created with standard box
Notes on
modelling
modelling. Modelling these components can be quick and easy
if we first reduce them to their basic forms. Analyse some of Feel free to mix and
the larger components such as the cabin box, which is match parts to create
your favourite
essentially the tub in which the interior sits. The dashboard is a combination. Here I opted
rounded cube tucked under the windshield line and the seats for bigger, more stylish
are broad L shapes with soft edges. Start simple and remember fenders from a reference
that Preview Smooth can help provide that last level of polish. image. The Cobra has a
one-piece body, but most
modern car bodies are
comprised of separate
panels. If you’re following
along with one such car
type, I would highly
recommend breaking
these panels up into
separate meshes: front
bumper, bonnet, front
wing, quarter panel and
09 so on. Also, name
everything and keep the
display layers, hypershade

10 Add exterior accessories External elements can

be a subtle but important factor in producing realism.
Compared with the interior, these are more varied and
and outliner organised.
Remember to delete any
unneeded history and
click Freeze
numerous, but still very simple, so the same advice goes. Transformations on
Consider each piece, start with simple forms and use Geometry whenever
smoothing to get that final crispness. Because some of these appropriate. This should
pieces can be subtle, it’s useful to comb through the reference keep the scene neat,
crash-free and far easier
images to make a list of all the accessories. Incorporating these to work with.
10
small details can be critical in achieving realism.

3D Art & Design 137

 Vehicle

Work on materials and lighting

Use image-based lighting and form realistic car paint

11 Set up image-based lighting (123ũ%et a suitable

ũ(,%#ũIJũ3'#1#ũ1#ũ/+#-38ũ.$ũ%1#3ũ-"ũ$1##ũ2.41!#2ũ
$.1ũ 2ũ.-+(-#ēũ.1ũ3'#ũ5#'(!+#ũ+(%'3(-%Ĕũ-ũ.5#1++ũ"1*#1ũ
 ũ6(3'ũũ$#6ũ!1(2/ũ'.32/.32ũ6.1*2ũ6#++Ĕũ2.ũ3.ũ2#3ũ3'(2ũ4/ũ6#ũ
2(,/+8ũ-##"ũ3.ũ%.ũ(-3.ũ.41ũ,#-3+ũ18ũ#-"#1ũ2#33(-%2Ĕũ4-"#1ũ
3'#ũ -"(1#!3ũ(%'3(-%ũ3 Ĕũ(-ũ3'#ũ-5(1.-,#-3ũ"1./".6-ũ,#-4ēũ
+(!*ũ3'#ũ1#3#ũ 433.-ũ-#73ũ3.ũ ,%#ũ2#"ũ(%'3(-%ũ3.ũ!1#3#ũ
3'#ũ ũ2/'#1#ũ1.4-"ũ3'#ũ2!#-#ēũ.6Ĕũ6(3'ũ3'#ũ ũ2/'#1#ũ
2#+#!3#"Ĕũ6#ũ!-ũ%.ũ(-3.ũ3'#ũ
331( 43#ũ"(3.1ũ3 Ĕũ,.5#ũ.5#1ũ3.ũ
3'#ũ +'/#ũ-."#ũ-"ũ/+4%ũ(-ũ.41ũũ(,%#ē

12 Make the car paint material Mental ray comes

with a powerful and simple shader designed to simulate
1#+ũ!1ũ/(-3Ėũ,(Ĵ!1Ĵ/(-3Ĵ/'#-Ĵ7Ĵ/22#2ēũ'#ũ22#2ũ
variant is best to use because it’s compatible with mental ray’s
Passes system, which may be handy later. This shader can
11
(-23-3+8ũ%(5#ũ42ũ2.,#ũ#7!#++#-3ũ1#24+32Ĕũ 43ũ ũ+2.ũ1#!.,,#-"ũ
#7/#1(,#-3(-%ũ6(3'ũ(32ũ,-8ũ$4-!3(.-2ũ3.ũ2##ũ6'3ũ.3'#1ũ+..*2ũ
!-ũ #ũ/1."4!#"ēũ#3ũ3'#ũ2#Ĵ!.+.1Ĕũ3'#-ũ42#ũ"%#Ĵ.+.1ũ-"ũ
(3Ĵ.+.1ũ3.ũ2(,4+3#ũ'.6ũ!1ũ/(-3ũ3.-#2ũ!'-%#ũ2ũ3'#ũ241$!#ũ
curves away from the eye, or into direct light. The Bias value
"(!33#2ũ'.6ũ#!'ũ!.+.41ũ +#-"2ũ(-3.ũ3'#ũ2#Ĵ!.+.1ē

13 Use other materials .r this image, almost all the

.3'#1ũ,3#1(+2ũ6#1#ũ!1#3#"ũ42(-%ũ3'#ũ5#123(+#ũ,(Ĵ
,3#1(+Ĵ7Ĵ/22#2Ĕũ6'(!'ũ(2ũ"#2(%-#"ũ3.ũ1#+(23(!++8ũ2(,4+3#ũ
5(134++8ũ-8ũ241$!#ũ38/#ēũ 3ũ!.,#2ũ6(3'ũũ!.-2("#1 +#ũ+( 118ũ
.$ũ331( 43#ũ/1#2#32ũ3.ũ04(!*+8ũ%#3ũ-ũ#7!#++#-3ũ2313(-%ũ/.(-3ũ$.1ũ
any material. This shader is also worth studying in real depth,
but for this tutorial we can only cover some of its basic
$4-!3(.-2ēũ313(-%ũ6(3'ũ3'#ũ5(+ +#ũ/1#2#32Ĕũ)423ũ")423ũ-"ũ
36#*ũ3'#ũ331( 43#2ũ3.ũ24(3ũ8.41ũ-##"2ũ-"ũ'.-#ũ(-ũ.-ũ3'#ũ
12
/#1$#!3ũ+..*ũ$.1ũ#!'ũ241$!#ē

13

How to get the best

materials and lighting
Materials and lighting are disciplines in their own right, so
we can’t cover all of their intricacies in great depth here.
Below are just some of the most important concepts that
should be mentioned:
ěũũ#3ũ4/ũ.+.1ũ -%#,#-3ũ-"ũũ(-#1ũ.1*Ą.6ũ3.ũ
ensure that Maya is reading textures correctly, as well as
displaying renders accurately.
ěũũ -!1#2#ũ3'#ũ(-+ũ3'#1ũ4+(38ũ(-ũ3'#ũ -"(1#!3ũ(%'3(-%ũ
tab by increasing the filter. The Accuracy and Point
#-2(38ũ!.-31.+ũũ82ũ-"ũũ.(-32Ĕũ/1."4!(-%ũ'(%'#1ũ
lighting detail and fidelity.
ěũũ.ũ/1./#1+8ũ1#-"#1ũ,(Ĵ,3#1(+Ĵ7Ĵ/22#2ũ,3#1(+2Ĕũ8.4ũ
should enable Output Maya 2009 Passes in the
Advanced dropdown menu.

138 3D Art & Design

 Model a complete car

Start the render

Adjust your render settings and make initial passes

14 Tweak the quality settings 3Ħ2ũ!.,,.-ũ/1!3(2#ũ3.ũ

")423ũă-+ũ2#33(-%2ũ-#1ũ3'#ũ#-"ũ.$ũũ6.1*Ą.6Ĕũ
#!42#ũ3'#8ũ!-ũ2(%-(ă!-3+8ũ(-!1#2#ũ3'#ũ1#-"#1ũ3(,#ũ6'#-ũ
8.4ũ!.,#ũ3.ũă-(2'(-%ēũ.4ũ!-ũ!+#1ũ4/ũ+(2(-%ũ13#$!32ũ(-ũ
#-"#1ũ#33(-%2š4+(38š831!#ĵ!-+(-#ũ4+(38ēũ#+#!3ũ
-!1#2#ũ 7ũ,/+#ũ#5#+Ĕũ+.6#1ũ3'#ũ
-3(ı+(2(-%ũ.-3123ũ
-"ũ")423ũ3'#ũ 4+3(ı(7#+ũ(+3#1ēũ.4Ħ++ũ+2.ũ6-3ũ3.ũ1#-"#1ũ(-ũũ
+(-#1ũ$.1,3ũ3.ũ/1.5("#ũ3'#ũ,.23ũ!.-31.+ũ$.1ũ/.23ı/1."4!3(.-ũ
#-'-!#,#-32ũ-"ũ!.,/.2(3(-%ēũ#3ũ3'#ũ#$4+3ũ43/43ũ1.ă+#ũ
2ũ(-#1ũ2ũ-"ũ42#ũ-ũũ.1ũ/#-ũ ,%#ũ$.1,3ēũ -ũ
3'#ũ4+(38ũ3 Ĕũ%.ũ3.ũ1,# 4Ă#1š38/#ũ-"ũ42#ũ3'#ũ

ũĸ+.3ĹũČũ7ũĊĉũ(3ēũ

15 The Beauty passes .,,.-+8ũ8.4ũ6.4+"ũ+8#1ũ

1#Ą#!3(.-2Ĕũ2'".62ũ-"ũ1#$1!3(.-2ũ.-ũ3'#ũ(Ă42#ũ
/22Ĕũ 43ũ$.1ũ3'(2ũ(,%#ũ3'#ũ#438ũ/22ũ'2ũ ##-ũ42#"ũ2ũ3'#ũ
2#ũ$.1ũ3'#ũă-+ũ!.,/.2(3#ēũ#-"#1ũ.43ũ3'#ũ$4++ũ(,%#ũ6(3'ũ++ũ
3'#ũ5(24+ũ(-$.1,3(.-ũ.-ũũ2(-%+#ũ+8#1ēũ 3Ħ2ũ+2.ũ%.."ũ3.ũ!1#3#ũ
(2.+3#"ũ#438ũ/22#2ũ$.1ũ3'#ũ!1Ĕũ3'#ũĄ..1ũ-"ũ3'#ũ
!*%1.4-"ēũ'(2ũ!-ũ/1.5("#ũ(-"(5("4+ũ!.-31.+ũ.5#1ũ#!'ũ
. )#!3ũ(-ũ3'#ũ#438ũ/22ēũ.ũ3'(2ũ 8ũ2(,/+8ũ341-(-%ũ.Ăũ1(,18ũ
(2( (+(38ũ(-ũ3'#ũ'/#ũ#-"#1ũ332ũ$.1ũ4--##"#"ũ. )#!32ē

16 The Reflection passes '#ũ ."8Ĕũ!'1.,#ũ

!!#22.1(#2ũ-"ũ6'(3#ũ,1 +#ũĄ..1ũ1#ũ++ũ'#5(+8ũ
1#Ą#!3(5#ũ241$!#2Ĕũ2.ũ8.4ũ-##"ũ3.ũ!1#3#ũũ51(#38ũ.$ũ#Ą#!3(.-ũ
/22#2ũ3.ũ6.1*ũ6(3'ũ(-ũ3'#ũ!.,/.2(3(-%ũ/'2#ēũ
%(-Ĕũ(3Ħ2ũ42#$4+ũ
3.ũ'5#ũũ$4++ũ!#-#ũ/22Ĕũ(-ũ""(3(.-ũ3.ũ 2.+3(.-ũ/22#2ũ.$ũ3'#ũ
14
!1ũ-"ũ3'#ũĄ..1ēũ$ũ!.412#Ĕũ(3Ħ2ũ #23ũ3.ũ42#ũ,#-3+ũ18Ħ2ũ22#2ũ
2823#,ũ$.1ũ3'(2Ĕũ 43ũ(-ũ2.,#ũ!2#2ũ(3ũ!-ũ #ũ"5-3%#.42ũ3.ũ
!1#3#ũũ"(Ă#1#-3ũ*(-"ũ.$ũ#Ą#!3(.-ũ/22ēũ18ũ!.3(-%ũ3'#ũ31%#3ũ 15
%#.,#318ũ6(3'ũũ$4++8ũ1#Ą#!3(5#ũ1!'(3#!341+ũ,3#1(+ũ3.ũ/1.5("#ũ Make the
-ũ+3#1-3#ũ38/#ũ.$ũ#Ą#!3(.-ũ/22ũ6(3'ũ5#18ũ231.-%Ĕũ$4++8ũ right passes
./04#ũ1#Ą#!3(.-2ē
ěũũ'#ũ22#2ũ3 ũ#- +#2ũ
16 42ũ3.ũ1#-"#1ũ,-8ũ
/22#2ũ3ũ.-!#ēũ 3Ħ2ũ
04(!*Ĕũ#28ũ3.ũ42#ũ-"ũ
#731#,#+8ũ'#+/$4+ēũ 3Ħ2ũ
!/ +#ũ.$ũ,4!'ũ,.1#ũ
3'-ũ6#ũ!-ũ!.5#1ũ'#1#Ĕũ
2.ũ(3Ħ2ũ"#ă-(3#+8ũ6.13'ũ
+..*(-%ũ(-3.ũ-"ũ
#7/#1(,#-3(-%ũ6(3'ē
ěũũ ũ+2.ũ1#!.,,#-"ũ
234"8(-%ũ"(Ă#1#-3ũ
/1!3(!+ũ,#3'."2ũ$.1ũ
!1#3(-%ũ51(.42ũ/22#2ēũ
*#ũ3#232ũ3.ũ!1#3#ũ
4-(04#ũ.1ũ!423.,ũ38/#2ũ
.$ũ/22#2ũ3'3ũ!-ũ #ũũ
+.3ũ.$ũ$4-Ĕũ/1.5("#ũ
+3#1-3#ũ5(24+ũ!.-31.+2ũ
-"ũ%(5#ũ8.41ũ(,%#ũũ
4-(04#ũ5(24+ũ#+#,#-3ē
ěũũ#ũ!-ũ!./8ũ3'#ũ#/3'ũ
/22ũ(-ũ'.3.2'./ũ(-3.ũ
-ũ
+/'ũ!'--#+ũ-"ũ
3'#-ũ42#ũ(3ũ6(3'ũ3'#ũ#-2ũ
+41ũă+3#1ũ3.ũ04(!*+8ũ
!1#3#ũũ1#+(23(!ũ"#/3'ũ
.$ũă#+"ũ#Ă#!3ē

Ċũ
13ũĜũ#2(%-ũũ139
 Vehicle

Composite the passes

Bring all your render passes together in Photoshop

17 Experiment with passes For this project we’re

compositing on top of a Beauty pass, so we don’t
absolutely need all the different types of passes, but a complete
set can be a powerful tool for enhancing the visual quality of
the final image. Having an abundance of different passes to
work with is always advisable, even if some end up being
discarded. This way you can experiment with a bunch of
different passes, adjustments and blending modes to see what
looks the most effective. In addition to the Beauty and
Reflection passes, Specular, Shadow, Ambient Occlusion,
Depth and Diffuse passes have also been used here.

17

19

18

18 Composite your passes Open Photoshop and copy

your various passes into a single PSD. Take the full
Beauty pass, place it at the bottom of the layer stack, then put
the Diffuse pass directly over the Beauty with a low opacity
setting. You can adjust its opacity or apply a layer mask to
control how the elements in the Beauty pass show through.
Finally, take all the other passes and set their blending mode to
Linear Dodge (Add). Now that you have our PSD set up to
provide full control over each element of the image, you can
start enhancing everything.

19 Make final touches At this point we’re still striving

for realism, but with a focus on visual interest and
appeal, such as a photograph enhanced for advertising. You
can use your artistic judgement to fine-tune the colours and
other elements accordingly. Focus on increasing the value
range to put duller darks with brighter whites and boost the
saturation on the body. Also, try creating subtle photographic
effects such as a vignette and a shallow depth of field. This is
your final pass at enhancing the image, so take your time and
experiment with different blends and passes.

140 3D Art & Design

 Gallery
This image was created as a demo
showcasing the use of a variety of different
software packages and photographic
techniques to develop a keyframe concept
art piece. I shot a background at a concrete
plant with the notion of adding myself and
a bike within it
Colie Wertz No More Gas, 2012

Artist info
Colie Wertz
Colie has worked with film VFX
for 17 years and is developing
content with 3D Systems, Inc.
Personal portfolio site
www.coliewertz.com
Country USA
Software used Maya, V-Ray,
MARI, Photoshop, Procreate

Work in progress…

3D Art & Design141

 Vehicle

Easy-to-follow guides
take you from concept
Build a vehicle
game asset
to the final render
Artist info

Ready to roll out 2013

Rainer Duda
Learn how to create and prepare a vehicle object
Username: Rainerd for use in a videogame production
Personal portfolio site Rainer Duda is a freelance 3D generalist focused on creating assets for high-quality videogames
www.rainer-d.de
Country Germany

I
n this tutorial we’ll cover the creation of a static game Following these steps we’ll build a high-poly version of the
Software used 3ds Max,
Photoshop, xNormal asset, which can then be easily used in the Unreal asset from the existing low-poly geometry. We can then
Expertise Rainer specialises in Development Kit without the need to constantly jump project these details onto the low-poly asset via a Normal
creating assets for videogames,
including building maps and back and forth between programs. As a base we’ll be using map, as is the usual convention in contemporary videogame
texturing the final result a blueprint, which will be modified later on with some new development. We’ll use xNormal for generating the Normal
components to create a more exciting asset. map (www.xnormal.net), rather than just the 3ds Max tools.
It’s important to establish a working low-poly mesh that This is simply to illustrate that there are alternatives to the
includes a valid geometry representation but also at least built-in toolsets.
two UV channels, a proper collision geometry and a LOD In addition to all the above, the low-poly asset will also be
object. We will discuss each of these elements in turn equipped with a Normal map and custom-painted textures
throughout the tutorial. to give it a final look.

Start the mesh

Use a grid and blueprint to begin

01 Prepare the scene Before the actual modelling process starts, prepare the scene with
three grids, which will hold textures that cover the different views of the car from the
blueprint (2001 Dodge Charger blueprint from www.the-blueprints.com). We’ll need these
guides to constrain all the proportions while constructing the vehicle chassis. This scene setup is
actually quite simple. First we can just place a grid in the scene, scale it to a rectangle and unwrap
01
it via a UVW Map modifier along with a planar projection. An Unwrap UVW modifier enables us
to open the UV Editor and place the patches correctly. Duplicate this grid twice and match the
UV sets to their corresponding places on the blueprint. For the front, scale the grid to a box and
then place all of the sets next to one another in the correct order.

02 Get the grid working Now, with the template build container set, it’s time to model
the left side of the car. For this, use a grid with just a few subdivisions, convert it to an
editable poly and place it parallel to the reference plane. Start moving the outer points along the
silhouette and the middle points to fit the front and side shape. If the resolution of the grid is too
low, then some more edges in between using the Connect Edges function in the Modifier panel
will help. Take care that even in the silhouette you have a nice edge flow. For later use, it’s best to
place some cuts on the position where the door is situated. After selecting the edges from the
cut, it’s necessary to delete the polygons between the chamfer edges, as well as to extrude the
02
open edges just a bit inside.

03 Add depth After the silhouette is looking good, it’s time to start adding some depth.
To do this, select the corner edges on the outside of the silhouette and extrude them
towards the centre of the grid. As a result, a rough half body should now be visible. The new
available points next to the centre of the grid need to be scaled up a little bit, as the car is bigger
in the middle than on the outside. Now we can go one step further and add some more edge
rows and chamfer to refine the depth of the car. This occurs by moving the points along the
03
silhouette according to the top reference plane on the ground.

142 O3D3DArtist
Art & Design
 Build a vehicle game asset

Rainer has supplied

low- and high-poly solid
and wireframe assets, as
well as screenshots to
help with the steps

Concept
If you don’t have any self-produced concept
art or designs, blueprints are another great
option. For a quick and easy start we will take a
blueprint of a 2001 Dodge Charger from www.
the-blueprints.com.

3D Art & Design 143

 Vehicle

Bring in more features

Finish the initial mesh and start adding more elements

04

04 Complete the body To finish the body, return to the Mirror function in 3ds Max and
use it to duplicate the first half of the mesh to the opposite side as a copy. Both of these
pieces will need to be attached afterwards. Only the open edges at the middle of the car should
05 Add more elements Some vital elements are still
missing, such as the wheels, exhaust pipes and engine
parts. We’re also missing exciting additions, such as a ram and
be left after this. A quick and easy fix is to continuously select two vertices and weld them guns on top of the car. These missing bits are easy to model as
together. With this flow path, we can adjust some points after welding to fit the silhouette even we are still on the low-poly model. The details will come from
better than before. It’s important that we have a nice edge flow on the car, as this will help us the high-poly later on, so keep the geometries quite simple at
later on at the high-poly-modelling stage. this stage. Model just one wheel and one turbine out of a
cylinder with a few insets and edge loops with different scaling.
Note that all the props are being modelled in the centre of the
3D space, which makes things easier, especially when we
switch to the high-poly stage. The ram is more or less just an
extruded box, including a chamfer and a bit of vertex-scaling.
The same procedure goes for the ends of the exhaust pipe, plus
the mirrors, while the axle consists of a simple cylinder.

06 Time for the big guns The last sub-asset what we

need is a medium-sized Gatling gun. We don’t really
need special reference for this piece, as we can imagine it
however we like. The first thing to do is model a base structure
out of a cylinder. After converting it to an editable poly, an inset
and afterwards a poly-extrude will complete the base. The gun
itself consists of one cylinder with a chamfer and ten simple
05 cylinders arranged in a circle. Ammunition comes from a
cylinder that stays vertical to the gun next to a funnel, which is
06
quite simply a modified box. The holding structure is an
extruded box braced by a half-cylinder that can be extruded
along the normals. Again, note that we are using only simple
geometries at this stage.

Creating your low-poly

There are several ways to build low-poly objects, but this
tutorial uses only simple geometry to build the low-poly
object, basically so that everything runs a lot smoother
than otherwise. Some of the more-detailed pieces of the
vehicle consist of a collection of small pieces that are
stitched together. Of course, it’s entirely possible to create
a high-poly and then build the low-poly around it without
any stitching of geometry at all. Imagine some of the Epic
Games assets from titles such as Gears of War and Unreal
Tournament, which were built out of a single mesh and still
maintained quite a lot detail. Aside from this, stitched
objects involve even more attention during the Normal
map generation process.

144 3D Art & Design

 Build a vehicle game asset

Finalise the sculpt

Begin unwrapping to project details and paint the object

07 Build space to paint Before we start painting or even

exporting elements of the car, we need to unwrap the 08 Let there be light The UDK gives users the
possibility to use their own UV channels for the Light Choice of
low-poly with a UVW Map modifier. Be careful and overlap as
many parts as you can, as it makes sense to save space and scale
map, which is useful as we can decide for ourselves what kind
of polygons get more light information than the others. The
coverage
them up later. Both the front and back wheels can be overlapped only difference in comparison to the first UV channel is that The newest version of
and of course all the gun barrels, plus the remaining parts that are there can’t be any overlap on the Light map. By increasing the Epic Games’ UDK
symmetrical to one another. Don’t try to flip UV sets, because this map’s channel to 2 and clicking on Abandon, a new UV supports collision
can produce Normal map errors that we’ll have to fix later in channel will be created. It’s necessary to reorganise all geometries that are
stitched together as well
Photoshop by inverting the Green channels. meaningful polygon units so that there is no overlap at all. a prefix like UCX instead
of MCDCX. The latter
07 08 was used when Unreal
Engine 3 was released,
but it still works fine. If
there is no time to spend
building a collision
geometry, or it’s simply
not that important, there
is another solution you
can opt for. UDK offers
built-in tools to create
collisions inside the static
mesh viewer. Users can
choose between simple
collisions up to convex
shapes. The same built-in
feature actually exists for
the creation of a UV
channel for a Light map.
This shows how there’s
no need to struggle with
3ds Max if you’re
unfamiliar with it.

09 Set up a collision model At this point the low-poly

isn’t ready for the game world, as it needs information
relating to where the player is unable to move through or over.
09

To provide this information, build nine simple boxes that cover

almost the whole car, including the guns. It’s important that the
boxes aren’t touching or intersecting with one another. After all
the boxes are in place, it’s time to attach them and give the new
piece a name with the prefix ‘MCDCX’ followed by the name of
your low-poly. After exporting the low-poly, complete with the
collision geometry, the UDK will interpret the second geometry
during the import as collision data.

10 Export and map-generation xNormal takes low-

and high-poly meshes as separate inputs, which is
good for us because at the moment we only have the low-poly 10
version. What’s important at this stage is that we don’t export
the whole vehicle at once, but rather its various units. To do
this, split the complete vehicle into parts to be on the safe side
and to obtain proper maps. At the end you’ll have the body, one
wheel from the front, one from the rear, both mirrors, the
bumper, one turbine and an exhaust pipe. All of these parts
need to be centred in the 3D world. In addition we’ll keep
another version of the complete low-poly vehicle as well.
Unfortunately we have to explode the turret to make it work
properly. This means all the parts that aren’t overlapping need
to be detachable and separated from the main mesh and as
such can be exported separately. Furthermore, low-poly parts
for the map-generation need to have just one smoothing group.

3D Art & Design 145

 Vehicle

Make the high-poly mesh

It’s time to slice the low-poly and build high-poly parts

Maximise
on polygons
For this vehicle we built a
high-poly directly in 3ds
Max. However, you’ll find
there are certain
limitations when working
with a large number of
polygons, as if the
number is too high or 11
your machine too slow, it
simply won’t be possible
to work in a simple
manner without any
crashes and distractions.
11 Time for metal Our low-poly has a straightforward
edge flow and that’s key to building high-poly metal
plates. Now it’s time to slice the low-poly mesh vertically and
As a consequence, users horizontally. The outcome must be single-sided plates and
can switch the whole
stripes that need to be extruded and closed. A chamfer on the
high- and low-poly-
building process to outer edges will produce some nice gaps. Next to this some
ZBrush. Pixologic has more details should be added such as a back light, openings at
implemented a highly the back, grilled windows, rivets, door handles, a front grill and
useful function to render so on. To save time, it’s possible to skip the modelling of the
decent Normal maps, but
alternatively it’s also window grills, because this detail can be added on the Normal
possible just to create a map through a texture transformation. There are many
high-poly mesh and use it programs that convert pictures to a Normal map, such as
in combination with CrazyBump or the NVIDIA Normal map tools for Photoshop.
xNormal and a low-poly
However, you need to be aware that an unwrap is never 100
object from 3ds Max to
achieve the same result. per cent stretch-free and the spot to replace texture pieces
must be very carefully selected.

12 Include ornaments for the final touch Add some details around the car to make it
feel a little more unique. To build these ornaments, add a few boxes in a row and
subdivide them to build very rough formations such as leaves and square-shaped flowers.
Adaptations can be made by moving vertices and splitting edges, plus extruding along paths.
After that it’s easy to duplicate them in one direction as much as is needed to build longer chains.
When the various ornament chains are complete, a Path Deform modifier combined with a path
can be used to adjust the positions of each chain. The overall resolution must be increased by
12
using a Turbo Smooth modifier.

13 Prepare to export
the vehicle mesh
Next merge the high-poly
objects, such as the rivets,
together with the objects
underneath, and the
ornaments with their base
meshes. Basically we need to
attach all the high-poly pieces
together that belong to their
corresponding low-poly
meshes. Before you start the
actual exporting, you must
take care that the meaningful
units of the high-poly meshes
match their positions with the
corresponding low-poly
objects. Just try to avoid
offsets and you’ll be ready to
13
export all of them.

146 3D Art & Design

 Build a vehicle game asset

Create the maps

Take your meshes into xNormal

14 Import all the meshes Open xNormal and import

all the high- and low-poly meshes into their respective
slots by clicking the Import button. A proper naming
15 Name of the game We can now begin setting up xNormal. Move to the Baking
options to create Normal and Occlusion maps by activating the relevant flags. To increase
the quality of the projected details, it’s recommended to set the Antialiasing to 4x. If polygons
convention pays off at this stage. Each object set – matching within the UV channels are too close, a higher Edge Padding number will cause overlaps. A value
low- and high-poly – must be activated separately for rendering of 3 pixels or even less is reasonable, but here we’ve opted for 4. Last, the Output Resolution
the maps. This can be done quite simply by activating and/or needs to be set, as well as the format, which we can set to TGA. If the vehicle is for an in-game
deactivating their checkboxes. cinematic, we can use a larger resolution such as 2,048 or more.

14 15

16 Use an optimal projection cage xNormal gives

users the ability to take advantage of an implemented
ray distance calculator that puts a projection cage around the
high-poly to measure the ray distance between the high- and
low-poly meshes. How accurate the calculation is depends
entirely on the time that the calculator is originally given, but so
long as the calculator measures as more detailed, the maps will
look better. Any calculation that lasts longer than one minute is
fine in principle, because the cage has had enough time for the
calculations to expand correctly. After this stage all that’s left is
to copy the results and generate the texture maps. The final
steps in this process involve combining of all the parts to
complete both maps in Photoshop. If the Normal map detail
isn’t quite strong enough, you can try moving to CrazyBump,
which is perfectly capable of increasing the intensity needed for
the final result.

16

3D Art & Design 147

 Vehicle

Move to post-production
Build texture maps in Photoshop and finalise your asset

17

17 Build two maps Now Photoshop can be used to paint

two new maps; one to keep the diffuse and the other for
specular information. For this vehicle, the diffuse will contain
18

just a few variations and mostly dark colours with only a few
details. For both maps we can use some rusty, dirty and
bare-metal textures from portals like www.cgtextures.com and
overlay them where it makes sense. Reduce the visibility of the
respective layers and make intensive use of the Eraser Tool to
add some variation to the specularity. With a combination of
brushes and tones try to add a Mad Max-like look to the vehicle.

18 Ready to roll out The complete low-poly object can

be exported out of 3ds Max, including the collision
geometry, as an FBX file. The asset and textures are now ready
for import into the UDK game engine. If you want to achieve
more details on your mesh, just use detailed Normal maps like
developer Epic did on their assets. The principle behind this
idea is having a tileable Normal map added to the base Normal
map of the vehicle in the Material Editor. For rendering inside
3ds Max, we need a material that contains a Normal/Bump as
the Displacement map source.

Mapping alternatives
If you’d rather not use xNormal for this process, then no
worries. You can easily use the 3ds Max built-in map
generation tool named Render to Texture. This needs a
low- and high-poly too, but there are several options that
need to be set, such as using the existing UV channel
instead of an automatic unwrap once selecting a low-poly.
When a high-poly is selected, the low-poly will be equipped
with a projection cage. That cage needs to be reset to work
properly and must be close to the high-poly.

148 3D Art & Design

 Artist info
Gallery

Luigi Memola
Born in Mexico, Luigi was
adopted by Italian parents, who
gave him a passion for art
Personal portfolio site
www.luigi-memola.com
Country Italy
Software used Rhinoceros 4,
KeyShot, Photoshop

Work in progress…

The design of the bike piloted by a droid belongs

to my wider personal project in which I created a
series of vehicles. The chapter CERN05 describes an
extreme competition between motorcycles and
droids that are piloted remotely
Luigi Memola AEG27Cern 05, 2012
3D Art & Design 149
 Vehicle

Easy-to-follow guides Create

Bunkspeed
take you from concept
to the final render
Artist info

Peter Blight
Personal portfolio site www.
peterblightconceptdesign.com
renders
Adaptive Field Tank 2013
Bunkspeed
Pro 2014 60-day
trial with
the disk
Country Australia
Software used
Bunkspeed Pro 2014 Here we’ll look at how Bunkspeed can be used to showcase
Expertise Peter Blight is a
freelance Vehicle Concept a sci-fi vehicle concept in the shortest possible timeframe
Designer specialising in sci-fi
Peter Blight Sci-fi and fantasy concept designer

B
unkspeed’s suite of products is famous for Putting myself inside the mind of an evil computer-world
achieving beautifully rendered images at breakneck overlord perfecting his secret weapon, my process is to
speed. Here I have re-imagined an original concept I achieve a final render as fast as I can while still making it
created several years ago, the Adaptive Field Tank. interesting. You can see examples on this vehicle of some
The AFT is half jet, half tank, built to take advantage of very quick and dirty modelling. For model completionists,
the unusual physics of cyberspace. In reality, you can’t just there is a link on my site to the full-scene digital model kit
stick a jet engine on a tank without causing problems – with the original Prototype-A Adaptive Field Tanks
treads don’t move that fast and the aerodynamic drag would (available for a modest donation to my digi-supplies fund).
prevent it from reaching a decent speed. In cyberspace, the For this tutorial, I present the AFT Prototype-B.
treads only materialise where they contact the ground and I’ll take you through the process of importing the model,
are attached by code alone. This keeps the ship in ground using pre-made and generated custom materials, decal
effect for pursuing hyper-fast wheeled vehicles in areas application, environment setup and rendering. The tutorial
inaccessible to air attacks, with the additional advantage of version of the tank has treads which materialise as a solid.
being tethered to very grippy tread segments to absorb the To mimic the main illustration with pure energy elements,
recoil of its powerful turret and weapon pods. Make no experiment with the emissive material settings and apply to
mistake, you don’t want this beast on your tail. wherever you feel the tank needs some glow action.

150 3D Art & Design

 Create Bunkspeed renders

ěũũ4++ũ2!#-#ũăũ+#ũ$.1ũ3'#ũ
343.1(+Ĕũ$4++ũē. )ũ,."#+ũ
$.1ũ(,/.13(-%Ĕũ-"ũ
22.!(3#"ũ,3#1(+2Ĕũ
 ũ-"ũ"#!+ē

Concept
I conceived the Adaptive Field Tank as
a vehicle that bends the rules of
physics in cyberspace to generate
field-linked treads only where it
contacts the ground. This frees up the
body to be as aerodynamic as possible
– for something covered in weapons,
at least! It’s a jet/tank hybrid, with a
dash of Japanese magnetic levitation
thrown in for good measure.

3D Art & Design 151

 Vehicle

Prepare and render your model

Quickly set up and complete a Fast mode render of your work

01 Getting started Open up Bunkspeed and select the

Hybrid renderer (if this is not already selected) then hit
New Project. Under the Project Menu, select Import Model.
Select the file AdaptiveFieldTankTutorial.obj from the cover CD
provided with the magazine. In the Import Settings menu,
select Materials tab, check Ignore Texture References and then
press OK. Ensure that you have the Fast button pressed in
order to take advantage of the realtime preview. Some quick
navigation tips include Alt+ hold left mouse button to rotate,
Alt+ hold right mouse button to zoom and Alt + spin your
mouse wheel away from you to exaggerate the perspective for
01
a more dynamic pose.

02

02 Create custom materials In the right tool panel,

select the Material tab. Here we can make the whole
thing prettier with just a couple of tweaks. Select the default
material’s Material Type drop-down box and change it to Metal
with roughness set to 9.82. For the headlight/eye sensors, right
click in the black space next to the default material and select
New Material. From the Material Type drop-down box, select
Emissive. Set the intensity to ten and the colour to pure red.
Click on the material and drag it over the headlights and eye
sensors of the tank to assign.

03 Browse pre-made materials In order to give you

more choice of materials you can join the Bunkspeed
site and then use that login to access their asset/material
library directly from within the software. Click the top-right
hyperlink within the application Login to Bunkspeed and the File
Library tab, and you’ll see two icons appear beneath the tab to
toggle between the local and web library. Go into the Metallic
Paint sub-folder and grab Plasma Red and Metallic Night. Using
the screenshots in this tutorial as a guide, apply each material
to the various body parts to achieve a nicely balanced paint job.
Alternatively, I’ve provided these specific materials with the
03
cover disc tutorial files.

152 3D Art & Design

 Create Bunkspeed renders

Decal use
Sometimes decals will
appear to be floating over
the surface of an object
rather than stuck to it.
This is caused by the
scale of the objects in the
scene being too small,
and is easily rectified. In
the quick menu panel in
the top of the screen,
ensure the Selection Tool
is set to Model and that
the sub-selection is also
Model. Select the Object
Manipulation Tool and
click the Scale icon. Hit
Ctrl+A in order to select
all, and then scale the
scene up several times
using the manipulator
inside the viewport before
04 re-applying the decal.

04 Assign a decal In order to efficiently assign a decal it’s helpful to do so in an orthographic

view and set render mode to Preview while positioning the image. You can either rotate the
existing camera to look directly down at the tank and check the Orthographic checkbox in the Camera
05

Properties tab or use the top view from the multi-viewports option in the View menu. Right click in the
black space next to your materials and select New Decal. Select DecalByJamienListon and drag it onto
the model as shown. Click Project from Current Ortho Camera button with the decal selected then
scale/move to fit. Uncheck Detatched and ensure Multiple Part Decal is checked.

05 Environment setup In the File Library tab, go into the Environment sub-folder and
grab the Studio 008 HDRI. Drag it into the viewport to assign it to the scene. Click on
the little wireframe planet icon for the Environment settings and set Gamma to 4 and brightness
to 0.77. The Bunkspeed site has a huge number of HDRIs to choose from, though I found this one
to be the most appropriate for a cyber tank. If an HDRI has a weird ground but a decent sky, you
can create simple geo in Bunkspeed to block it out and assign a ground texture.

06
06 Rendering The
most satisfying part
is render time, especially
when using Bunkspeed PRO
2014’s new Fast mode.
Accurate is almost as quick,
depending on whether you
want to see more accurate
glows/reflections etc. I
recommend using Fast
regardless just to bang out a
bunch of stills to decide on an
angle before setting up an
Accurate render at full
resolution. After you’ve
picked the optimal angle,
hover the mouse over the
floating menu towards the
top of the screen and click the
camera-shutter icon Render.
Set it to render to 3,000
pixels wide, the Render Mode
to Quality, and the Number of
passes to 1,500. Output as a
.PNG in order to avoid
compression artefacts, and
you’re good to go!

3D Art & Design 153

 Vehicle

Easy-to-follow guides

Model a
take you from concept
to the final render
Artist info

Ben Simonds
fighter jet
Personal portfolio site
www.bensimonds.com
Location London
Software used
Aerial Pursuit 2013
Blender, GIMP, Photoshop Learn how to build a dramatic scene featuring a
Expertise Modeller, 3D
generalist and VFX artist using realistic fighter jet being hotly pursued by the enemy
a range of software
Ben Simonds is a 3D artist at Gecko Animation

H
ere you’ll learn how to model a fighter jet composite. We’ll cover blocking out the model mixing procedural techniques with hand-painted
from scratch, then construct a dramatic from references, before creating the details with a maps while using Blender’s node materials.
scene featuring a dogfight high above the mix of retopology and modelling methods. After finishing up the jet itself we’ll move on to
clouds. Throughout we’ll be using Blender for Next we’ll move to making materials for the creating the final image, adding two pursuing jets
modelling and rendering the 3D elements, GIMP fighter jet, with a view to rendering a final image into the background along with other exciting
for painting textures and Photoshop for the final with Blender’s Cycles renderer. We’ll focus on features such as explosions, contrails and clouds.

Begin modelling Snapping

surfaces
Block out the fighter jet with simple placeholder geometry
Throughout the modelling
01 02 process I made use of
Blender’s surface-
snapping tools to create
new topology over the
surface of existing
objects. For example,
when modelling the wing
as shown in Step 3, you
can turn on Snap with the
magnet icon in the header
of the 3D viewport, then
set the Snap Target to
Face and turn on Snap

01 Start simple Begin with basic primitives, such as

cylinders, spheres and cubes, to input modifications
(extruding and scaling to capture the overall shape of each
Projection. These options
make it much easier to
model surfaces with holes
and other details in them.
major component of the model). This gives an early idea of You can also keep the
where the challenges will lie when building the model and also 03 overall surface smoother
acts as a starting point for creating more-complex meshes. by using a simpler shape
for the guide.

02 Build the fuselage With the model blocked out,

you can start from the front and begin shaping the
fuselage (the main body of the jet). Use the blocked-in model 01 The jet blocked out with
as a guide alongside your reference. Keep the poly count
relatively low at this stage, as we’ll be applying a Subdivision
Surface modifier later. It’s best to use a Mirror modifier to
03 Shape the wings You’ll need to find some reference
for the cross-section of a fighter jet’s wing to get the
aerofoil shape right. Model these with curves, convert them to
simple geometry
02 Using a Mirror modifier to
model the fuselage. Edge
loops are added around hard
model symmetrically. After finishing the fuselage, move on to meshes, then space them out and loft them together with the edges to retain sharpness
the air intakes. These are relatively box-like in shape, so start Bridge Two Edge Loops operator in Edit mode. This provides us
03 To construct the wings, first
out with a cube, then gradually manipulate it into shape. Now with a wing shape that we can then use as a guide to construct create curves for the
insert edge loops and extrusions to produce geometry that will the wings. Apply the same process to produce the horizontal cross-sections, then loft these
to make a wing shape and
maintain its sharp edges when subdivided. stabilisers (the rear wings) and the rudders. then model the actual wing

154 3D
O 3DArtist
Art & Design
 Model a fighter jet

ěũ(%'3#1
#3Ĵ!#-#ē +#-"
ěũ(%'3#1
#3ē +#-"
ěũ#7341#2

Concept
A modern fighter jet
bursts out of the
page, pursued by
enemy jets, as a
missile streaks by in
the foreground and
explosions punctuate
the distant sky.

3D Art & Design155

 Vehicle

Detail and texture

Now the main forms are established, we can start to layer on details

04 Model the missiles Adding details to the jet will

mostly consist of examining reference material and
modelling elements, such as individual panels on the body of
the jet, the interior of the cockpit and the exterior areas of the
jet engine at the rear. Of course, no fighter jet is complete
without an arsenal of explosive weaponry strapped to it’s belly,
so try building three different kinds of missile based on
real-world references, using the same techniques as those to
build the fighter jet itself. These are mainly cylinders with
some simple geometry added on for the fins, with a few extra
04
details here and there.

05 UV unwrapping
It’s now time to UV
unwrap your jet’s elements.
Thankfully, Blender has a
great set of tools for UV
unwrapping, so start by
adding seams around the
main sections of the fuselage.
These include the nose, as
well as the upper and lower
surfaces of the wings. You
also need to unwrap the
different parts of the
underside of the jet. You can
use Blender’s default Unwrap
operator to unwrap these
pieces and pack them into a
single UV grid. You’ll find this
grid will be very helpful later
on when you begin to bake
your textures. 05

06 Multiple UV maps Because we’ll want to add

decals to the model later, at this stage it’s helpful to
create a couple of extra UV sets. In addition to a standard
large camouflage pattern on the wings and the shark’s mouth
design on the nose cone. Try projecting specific sections of the
model from the most relevant Orthographic perspective – the
04 The missiles for the fighter jet
are modelled with simple
cylinders, extrusions and
extra parts for the fins
unwrap, which gives each piece of the model it’s own unique top for the upper surface of the wings, the bottom for the belly 05 Mark seams for UV
bit of UV space, we can create two other UV co-ordinate sets. of the jet and the left or right for the sides of the rudders. This unwrapping the fighter jet
Project large chunks of the model from key angles for the first provides a UV map that we can use for painting decals and 06 Creating various UV sets is a
(mainly from the top or the side). This provides a very simple other details, such as panelling on the model, without worrying bit of extra work at this stage,
but it makes texturing tasks
UV set that we can use to paint broad textures, such as the too much about them becoming distorted. much easier later on

06

156 3D Art & Design

 Model a fighter jet

07 Bake textures Once we have the fighter jet

unwrapped, we can add an Ambient Occlusion map
for the whole model using the baking tools in Blender’s Render
tab (note: you have to set the renderer to Blender Internal for
this, as Cycles doesn’t currently support texture-baking). Now
bake your AO map to the first UV set you created.

08 Paint a camouflage pattern Here we’ve gathered

various textures for the fighter jet as separate images,
which can be combined in Blender when making materials. Begin
by defining an overall base colour for each part of the jet with a
camouflage pattern and, for extra character, also paint in the
shark’s mouth design for the nose cone of the jet on this texture. To
act as a guide, export the UV layout created as an image, use the
UVs>Export UV Layout operator in Blender’s UV Image Editor and
open this up as a layer in GIMP.

09 Attach decals to the jet Now paint the decals for

the jet on a separate texture with an Alpha channel,
then build a couple of military-style designs as well as some
random pieces of text and numbers. Place these appropriately
07
on the wings, rudders and the fuselage of the jet.

Drawing 08

textures in
Blender
To create the impression
of complex panels and
rivets over the surface of
the jet, I partly used a
Displacement map.
Rather than laboriously
painting this in GIMP, I
opted to model the
outlines of the panels with
curves in Blender. I used
the exported UV
co-ordinates I created
earlier as a guide and laid 09
out the scene from a
top-down perspective. I
used an Orthographic
camera to render a
texture that perfectly
07 Baking an AO map. Make sure 09 Some of decals on the jet
10 Make grunge and scratch textures Now build up
a few different seamless textures for dirt and scratches
to be used in various materials covering the jet. To make a
you have the renderer set to
matched up with the jet’s Blender Internal and turn up 10 Take a square selection of texture seamless, first offset it in GIMP and place the
UV co-ordinates. This the Samples settings for your your base texture, offset the
Ambient Occlusion image to position the seams boundaries of the image in the middle of the canvas. Select the
could later be combined in the middle, then apply the
with my other textures. 08 Paint the camouflage pattern Resynthesise filter on the
seams and use the Resynthesise filter to fill them with a
for the body in GIMP model’s boundaries continuous texture. This should produce a smooth result.

10

3D Art & Design 157

 Vehicle

Realistic surfaces
Apply Blender’s node materials ready for rendering

11 Construct a Diffuse shader Initially make a Diffuse

shader for the body of the fighter jet, then take the
Diffuse BSDF shader and begin incorporating your textures.
Combine these with Color Mix nodes before plugging them
into the Diffuse Shader node. Join the camouflage texture with
the decals texture, then add in some further details using the
seamless grunge textures. Because the different textures use
various UV co-ordinate sets, you can include some Attribute
nodes (into which you can enter the name of the UV set you
wish to use). Plug the Vector Output of these into the Vector
Input of the image textures to let them know the correct UV
co-ordinates to use. For the seamless textures, use the Image
Texture node’s blended box-mapping feature to apply the
textures without the need for UV co-ordinates.

12 Apply glossy reflections Now you can combine the

Diffuse shader with a couple of Glossy BSDF shader
nodes using Mix Shader nodes. You can also reuse some of
your textures to affect the Color and Roughness inputs for the
Glossy shaders, to provide a bit of variation in the glossy
reflections. Apply two Glossy shaders: one for broad soft
highlights over the shape of the jet and the other to give some
sharp reflections on top. This isn’t physically correct, but it
gives extra control over the look of your material. Use a couple
11
of Layer Weight nodes to control the mixing of the shaders.

13

13 Displace elements
The Cycles renderer
supports a couple of methods
11 By creating a Diffuse shader
for the jet, you can use node
groups to keep your material
nodes organised
12
for defining surface texture 12 The two Glossy shaders
for your materials. Here we shown on their own, plus the
overall node setup
opted to plug the panels incorporating both diffuse
The Cycles render preview texture into the Displacement and glossy reflections

Input of the Material Output 13 Some displacement brings

The new Cycles render engine has a fantastic live preview for getting instant feedback on node. This provides some detail to the surface of the
your materials and lighting. You can enable this in the 3D viewport if you have the renderer wings and fuselage
set to Cycles and it will constantly update with a rendered preview. I used this when creating fake displacement for the
14 Linking the jet into the final
my materials to get an idea of how they looked under some simple lighting, then again later material and highlights the BLEND file as a linked library.
when making my final scene to tweak lights, World settings and render options. It’s best to panels that were drawn in. This keeps the scene live if
split off a smaller 3D viewport in your window layout and use this as your preview render you need to go back
We also used a Multiply
while you work in another 3D viewport. This will give you a rendered preview that updates 15 Rendering the volumetric
quickly, all while keeping it simple to select and edit an object in your main 3D viewport. Math node to tone down the smoke and trails in the scene
intensity of this displacement. as separate scenes

158 3D Art & Design

 Model a fighter jet

Showcase
Develop the scene Artist
Build a dramatic environment
Ben Simonds
Ben Simonds is a 3D artist and a director of Gecko
Animation Ltd, a small VFX and animation studio
based in London. He’s been using Blender for about
seven years. He also produces tutorials and articles for
his website, www.bensimonds.com, and is the author
of Blender Master Class, a book all about creating 3D art
using Blender.

Pirate Captain Blender,

GIMP (2010)
This scurvy sea dog was
created a few years ago for a
portrait competition. At the
time, subsurface scattering in
Blender was relatively new
and I wanted to create an
interesting but lifelike
character in the software.

14

14 Link the jet into a new BLEND File To create the final image, start a new BLEND file
to work on lighting and to render the jet in. To link the jet into this new file, assign all of the
objects making up the jet into a single group, then link this group into the new BLEND file. This
enables us to go back and modify the jet in it’s original file and have the final scene update
automatically when we reload it. It also enables us to create a couple of duplicates of the jet to
act as the other fighters in the scene.
Abandoned Warehouse Blender, GIMP (2011)
An abandoned warehouse full of rubble. Modelling was completed
in Blender, textures in GIMP and rendering in V-Ray.

Fourarmed II
ZBrush, Blender
(2010)
Ben has created a vast
amount of Blender
work over the years
that straddles
15
everything from
believable human
likenesses to
15 Build up smoke trails To add some extra intensity to the image, you can include some
smoke trails created with Blender’s smoke-simulator tools. These have to be rendered
with Blender’s older render engine – Blender Internal – which supports volumetric materials.
spaceships and
captivating soft-body
sculpts such as this
Render these effects in separate scenes, then link in objects like the jet to act as masks (using the detailed creature!
Mask Layers options when rendering).

3D Art & Design 159

 Vehicle

Post-production Simulating explosions

Refine the lighting and render settings For my final scene I wanted the jet to be pursued by other
fighters, among explosions and missiles. To create some
explosions and missile trails I used Blender’s smoke-
simulation tools. I made a couple of separate files with
different smoke simulations in – one for a fast-moving
missile with a trail of smoke behind it, another for some
smoke erupting from the wing of the jet and another for full-
on explosions. The smoke-simulation tools can be pretty
fiddly, so creating these in separate files will enable you to
concentrate on these elements in isolation. After you’re
happy with the result you can combine them with the other
pieces of the final scene.

16

16 Position the lights The lighting in this scene is pretty simple. We used a HDR sky map
to provide some global illumination lighting and a sun lamp for the main directional
lighting. We also added a large flat cube below the jet (out of shot on final image) and assigned a
light-blue Emission material to it. This provides some lighting from below to mimic reflected light
from the clouds. You can also try adding a bright-orange point light on the damaged wing, just
where the smoke is supposed to be pouring out, to create an orange glow on that side of the jet.

17 Rendering and
final composition
This final composite was
completed in Photoshop.
GIMP is preferred for painting
textures, but Photoshop has
flexible layer-management
for making multilayered
images. We rendered out
each of the elements for the
image as separate passes,
while completing a small
amount of compositing in
Blender. We then opened all
the passes as layers in
Photoshop. At this point you
can add aerial perspective
(fading distant objects), as
well as some extra motion
blur. You can also tweak the
colours of the various cloud
and smoke layers.

16 The lighting setup, showing

the Emitter mesh below the
jet, the sun lamp above and
the small orange point lamp
on the wing
17 Compositing the final image 17
in Photoshop

160 3D Art & Design

 Gallery
Artist info

Moran Tennenbaum
A 2D turned 3D artist, Moran
specialises in environments,
props and characters
Personal portfolio site
www.morantenn.com
Country USA
Software used Maya, ZBrush,
Photoshop, Mudbox, xNormal,
V-Ray, NUKE

Work in progress…

This project was created as part of the

Demo Reel class at Gnomon. The design and
intricate details presented a welcome challenge
in terms of modelling and textures
Moran Tennenbaum Waldo, 2013
3D Art & Design 161
 Vehicle

Easy-to-follow guides
take you from concept
to the final render
Artist info

Jonathan
Williamson
Personal portfolio site
'33/Ėĵĵ!%!..*(#ē!.,ĵ +#-"#1
Countryũ-(3#"ũ33#2
Software used +#-"#1
Expertiseũ
ũ,."#++#1Ĕũ2!4+/3.1ũ ěũũ'#ũē +#-"ũ2!#-#ũăũ+#ēũ'(2ũăũ+#ũ
-"ũ3./.+.%8ũ2/#!(+(23ũ6'.ũ+(*#2ũ (2ũ1#"8ũ3.ũ1#-"#1ũ-"ũ
3.ũ" +#ũ(-ũ#5#183'(-%ũ 2'.6!2#2ũ'.6ũ3.ũ42#ũ3'#ũ
ı1#+3#"Ĕũ-"ũ3#!'#2ũ+#-"#1ũ ,4+3(/+#ı,3#1(+ũ31(!*ũ.-ũũ
.-+(-#ũ3ũ..*(# !.,/+#7ũ,."#+

Create professional
wireframe renders
Make professional-looking wireframe renders in
Blender using its Cycles feature

R
endering wireframes in Cycles can be the Blender Internal, is my preferred way to
done using a few different methods. create decent wireframe renders. By
One option is to use the Wireframe compositing two renders you can create
Input node within a Cycles shader node tree. excellent results and keep a fine level of
Alternatively you can create actual control over the final effect. It’s also quite easy
wireframes via the Wireframe Mesh Operator to set up.
in Edit mode. Additionally, you can combine The Blender Internal method just takes a
your Cycles render with a Blender Internal few steps, but to make the process
render through the compositor. manageable we will utilise Blender’s scenes,
The first method, using the Wireframe duplicates of our model(s) and a series of 01
Input node, is the simplest – it also renders materials. We will need a material for the
everything in triangles at the highest base material, the wireframe material and a
subdivision level. However, this tends to result
in a very messy-looking render and it’s not
very flexible. With the second option, you can
white matte material. Last, we’ll need to set
up a simple compositing node tree to
combine the two renders together from each
01 Duplicate objects to a new scene The first thing
to do is create a new scene and duplicate or link all of
your model(s) into it. Duplicating objects is the most flexible
have a great level of control, but it relies on of the render engines. method, but linking them can be more convenient as it’ll use
actual geometry, which can become The true benefit to making wireframes in the same mesh data for each instance of the model in each of
expensive very quickly and also clutters up this way is that you are able to take full the scenes. In this case, I’ve chosen to make a new scene with
the scene. In light of these two, the third advantage of Cycles’ powerful node-based linked objects, as this way if I ever change something in the
option, combining your render with one from shader system. mesh, the changes will propagate across both scenes.

162 3D Art & Design

 Wireframe renders

02 Set the render engines for each scene Once both scenes are created you need to
specify which render engine to use for each one. The first scene, which will be used for
actually rendering the model, should be set to Cycles. The second scene, which will create the
wireframe render, should be set to Blender Render (also known as the Internal Engine).

02

03 Create wireframe material for second scene Next it’s time to create the
wireframe material in the second scene, (which should be the one set to Blender
Render). The first step (and most important part if you’ve chosen to link your objects) is to
change what the active material is linked to. Materials can either link to the data of an object, or
to the object itself. When the link is set to Object, you can use a separate material for each
instance of that object. In this case, I’m setting the link to Object. Now you need to make a new
material, since changing the link to Object has emptied the active Material slot. On this material
set the type to Wire and the Diffuse colour to black. You should also enable Transparency and 03
increase Z Offset to something around 0.01.

04 Produce white matte material for the frame

Now repeat the previous steps for the white matte
material, but this time leave the type set to Surface, set the
05
05 Render both
scenes Once the
materials are created and
Diffuse colour to white and then enable Shadeless. Also, go into added to each object for
Edit mode, select the whole mesh, then press Assign in the rendering, it’s time to actually
material. This material will be used to tell the compositor what render both scenes. Since
to ignore, as we’ll be multiplying the black wireframe over the we’re using multiple scenes,
Cycles render. While creating this material, it’s also a good idea this is most easily done from
to set the World Background Color to white as well, or you may the compositor. Switch over
find it slightly conflicting later on in the process. to the compositor, then go to
Input>Render Layer Node for
04 each scene, then specify the
scene from the dropdown
menu on the node. To render
a specific scene, just press
the Camera icon on the
Render Layer node.

06
06 Composite the
two renders
together Finally, to put it all
together we need to multiply
the wireframe render over the
Beauty render. This can be
done by adding a Color> Mix
Node and then setting the
Blend Type to Multiply. This
method is not as fast to set
up as some alternatives, but it
tends to yield the best results
with the most flexibility.

3D Art & Design 163

 Animation
Allow your 3D art to truly come to life by
learning how to animate
166 Animate a dialogue shot
Create seamless and believable Maya
dialogue shots in 12 easy steps

171 Gallery
CTN Animation Expo Kirk & Lucy by
Angel Navarro

172 Shatter moving characters

Animate a fractured character, as if
they are made of something brittle

178 Master blend cycles

Achieve Assassin’s Creed’s style of
animation in your artwork

185 Gallery
General Shellshock by
Koen Koopman

186 Animate vehicles

Take your car out for a spin and create
some cool stunts 171

166

164 3D Art & Design

178

185 186

172

3D Art & Design 165

 Animation

ěũ(+.%4#Ĵ'.3Ĵ313ē,
ěũ(+.%4#Ĵ'.3Ĵ(-+ē,
ěũ(+.%4#Ĵ
-(,3(.-ē,/Č
ěũ(+.%4#ē65

The audience’s attention is

typically centred on the
character’s face, especially
during a shot with dialogue

Animate a
dialogue shot
Digital-Tutors reveals how to create seamless and
Easy-to-follow guides believable Maya dialogue shots in 12 easy steps
take you from concept

T
to the final render ackling a shot that involves facial animation can be a Since you are trying to convey real-world movements into a
rather challenging process that takes up a great deal 3D virtual world, knowing just how human faces move is an
Artist info

of time. In fact, so much work is involved that it’s best extremely important step. With this in mind, before you even
to approach the facial animation as if it is a completely start attempting to create believable, well-crafted facial
separate shot in and of itself. There is as much planning and animation, you need to carefully study how your own face
care involved in animating the face as there is in the rest of deforms and moves around. Have a go at looking in the mirror
the animation. and reciting the dialogue yourself. What do you see? The
In this tutorial, we will condense the process into 12 amount of planning that is involved can seem like overkill, but
important steps, ensuring the facial animation of your shot is having this type of preparation in place will certainly speed up
Digital-Tutors
rock-solid and can hold up on its own. the process and help push the animation to the next level. As
Personal portfolio site When animating the face, head and mouth, the same ever when working in 3D, taking the time to get things right is
666ē"(%(3+343.12ē!.,
techniques that you use to animate the entire body can be absolutely key.
Expertiseũ(%(3+ı43.12ũ'#+/2ũ
13(232ũ,*#ũ,.5(#2Ĕũ%,#2Ĕũ applied. Timing, lead and follow, drag and exaggeration are all The character that we are going to animate will be saying:
5(24+ũ#Ăũ#!32Ĕũ"#2(%-2ũ-"ũ extremely vital to accomplishing a great animation. Any time “Ahhhhh… Have you seen the rest of my body?” The shot is 77
#,#1%(-%ũ,#"(ũ 8ũ/1.5("(-%ũũ
Ĕũũ-"ũ"(%(3+ũ the audience notices something strange in the animation, frames long. Our example character is an ogre, but any
13ı31(-(-%ũ1#2.41!# that’s the moment they will stop buying into your character. character rig with good facial controls should suffice. You can
That’s the last thing that you want to happen. find the tutorial project files free with the disc.

166 3D Art & Design

 Animate dialogue

Get to grips with facial animation

Important tips for creating your dialogue shots

01 Create a dialogue chart As we mentioned earlier,

planning is incredibly important when working on a
shot with dialogue. Write down each word of dialogue and
then break it up into syllables. Next, track the timing for each
word and study where the emphasis hits, or the points where
the character speaks most loudly in the audio. Most words
should have one of these emphasised sounds. Take notes on
what syllable they occur on, and then use the information as a
guide as to where to push the lip sync in your animation.

01

02

02 Act out the scene Animators are actors. The best way to get ideas for the scene is
to get in front of a camera and act out the shot. Don’t be shy or feel awkward – all
animators do it. Try to become the character during the acting process. Play the audio in the
background and play out the scene over and over again. When you think that you have got the
shot, play back the video that you’ve captured to double-check. Only stop when you are entirely
happy with your performance. A good rule of thumb for this is to force yourself not to settle for
anything less than 15 to 20 minutes in front of the camera. This technique will help you to really
get into the acting of the shot, and give you plenty of different options to pick between.

03 Block in six key

poses Use the
video that you shot in order
to find the proper storytelling
poses that truly describe the
facial animation. For each
storytelling pose you should
make sure to pose each area
of the face, rather than just
the entire head. This
technique will help you to
work out the type of facial
pose that you want to hit for
each specific word in the
dialogue. Try to keep the
poses to an absolute
minimum; you really only
want to block in the most
important ones in at this
stage. For example, our
03 blocking pass consisted of
just six main poses.

3D Art & Design 167

 Animation

04 Create
breakdowns The
breakdowns are the poses
04

in-between the storytelling

poses that describe how the
character is getting from one
pose to the other. Study your
reference and use that as a
guide to help you create your
breakdowns. Look at the
spacing between each pose
that’s in your reference and
incorporate that into your
animation. It’s very important
for you to remember that
breakdowns are not just
in-betweens. You should be
thinking about drag,
exaggeration and poses, all of
which can add a more
appealing look to the action.

05

05 Exaggerate the movements Once you have all

the proper breakdowns in place, the brunt work of the
animation is almost done! Make sure the timing is exactly how
you planned it and your holds are in the right spot. Once you’re 06
happy with the timing, go in and start pushing some of the
poses to really exaggerate the facial movement. For example,
when the character brings his head down, drag the eyebrows
and have them offset a few frames. You could even play with
the scale attributes for the head to add some squash and
06 Animate the eyebrows It’s important to have strong eyebrow poses to sell the
facial animation. Don’t overdo it by having the eyebrows moving during every single
word. Instead, find a few important eyebrow shapes that you want to hit during the dialogue. In
stretch. Exaggerating movement is just as important with the ogre animation there are really only two main eyebrow poses: frowned and surprised. It’s
facial animation as it is in the rest of the body. how you transition between the two that helps sell the animation.

168 3D Art & Design

 Animate dialogue

07 Fine-tune the eyelids After the overall

movements of the head are looking how you want
them, it’s time to go in and start fine-tuning some of the
smaller details. The eyelids may seem like the least important
area on your character, but they can really help sell the
emotion of your character if used right. For example, in our
animation the eyelids have been opened very wide in the spots
where the eyebrows are raised. This makes our character’s
face feel like it’s all one cohesive unit, working in sync.

07

08

08 Add in blinks When adding blinks to your facial animation, make sure that you
aren’t adding them just because you think the character’s eyes must be getting dry. If
you want to add in blinks with the correct timing you need to think about the emotional state of
the character. Is he angry? Sad? In our animation, the dialogue has a concerned tone to it, so the
character’s blinks can be minimal. In this shot there are a total of three blinks; the first is during
the head turn at the beginning, the next is when the character looks up at camera, and the final
one is when he bobs his head. Remember that you can also use the controls under the eyelids
to be able to really make each blink feel fleshy and tactile.

09 Time for lip sync

Once the head
movement and face are
09

working properly, it’s then

time to work on the lip sync.
This should be relatively
simple because you created
the dialogue chart in step
one. The first thing we need
to do is block in the jaw
opening and closing. Typically
this is done at the start and
end of a word, however, it
can be very easy to fall into
the trap of opening and
closing the jaw far more than
is actually needed. Keep
referring back to your
reference video to see how
much the jaw moves while
the dialogue is being
delivered and don’t overdo it.
Once this is completed you’re
already halfway done with
the lip sync.

3D Art & Design 169

 Animation

Animation tutorials
Find hundreds of full-length tutorials to help you become a
better animator and learn today’s leading creative
applications at www.digitaltutors.com. You can join
thousands of artists by becoming a member and learn
using the same training used by top schools, studios and
artists all around the world.

11

10

10 Work on mouth shapes When you have blocked

in the jaw movement, you are then able to go in and
start fine-tuning the shape of the mouth. Add some
asymmetry to the mouth shapes in order to help add
believability. Animate each lip control to get a sort of peeling
effect when the mouth opens. Think about the arcs that the
jaw is taking during this stage. You can really add a very fleshy
look to the lip sync by animating the entire rotation axes as
well as the translate axes. Have fun with this and don’t be
afraid to exaggerate!

11 Animate the cheeks A great place to make the

character’s face feel fleshier is in the cheeks. When
animating the face you do not want any area to feel dead, so
utilise some of the facial controls in order to make the face feel
like it’s all one unit working together to create the shapes. For
example, when the ‘B’ sound is being expressed in our shot, try
using the facial controls to puff out the cheeks. This adds a
sense of realism to the shot, while also making the lip sync feel
more natural.

12 Polish the animation Once you have animated the

lip sync, it’s then time to go in and add that final 10%,
which – as you will see – will really push your facial animation
to the next level. Simply track the arcs of the corner of the
mouth, nose and eyebrows. Tweak the eyelids in order to get a
nice fleshy look during blinks. Once the polish pass is finally 12
completed, then your facial animation is finished!

170 3D Art & Design

 Gallery
Artist info

Angel Navarro
A Spanish character artist living
in Sweden, he has a passion for
animation and characters
Personal portfolio site
www.angelnavarroart.com
Country Sweden
Software used MODO 701,
ZBrush, Maya, Photoshop

Work in progress…

This is a promotional poster for CTN Expo, featuring two of

the characters, with a film-noir style. As a team we put this
together for the webinars at VirtualAnimators.com. I was
responsible for modelling, texturing, posing and rendering
Angel Navarro CTN Animation Expo Kirk & Lucy, 2013
3D Art & Design 171
 Animation

ěũũ1(.42ũ2!#-#ũăũ+#2ũ
(-!+4"(-%ũ
ũ/1.)#!3ũ
ăũ+#2
ěũũ43.1(+ũ2!1##-2'.32

Concept
In this tutorial we are
going to shatter a
pre-fragmented
character by geometry
using mParticles and a
little ADM (Advanced
Data Manipulation) in
3ds Max 2014.

172 3D Art & Design

 Shatter moving characters

Easy-to-follow guides
take you from concept Shatter moving
characters
to the final render
Artist info

Ice Man Shatter 2013

Anselm von
Seherr-Thoss Here we’re going to animate a fractured character, as if they are made
of something brittle, using hand-animated control objects
Username: 3delicious
Anselm von Seherr-Thoss is an award-winning VFX TD and VES (Visual Effects Society) member
Personal portfolio site
www.incendii.com

H
Country USA ere you will be learning how to pre-fragment a Gazprom commercial where a fast running group of horses
Software used 3ds Max 2014, character using 3ds Max and RayFire. We will shatters: tinyurl.com/3DAGazprom.
RayFire (optional)
then pipe the fragments into mParticles and lock You will also learn basic ADM (Advanced Data
Expertise Anselm has worked
on VFX for the likes of Avatar them to the original moving character. Based on hand- Manipulation), which is the former PFlow Toolbox#3.
and Star Trek: Into Darkness placed geometry objects we will then release these Furthermore you will learn how to optimise your mParticles
fragments over time creating the effect of a crumble while flows and prevent a major PhysXplosion, which is caused by
the character is walking. This effect was used in a intersecting geometry.

Prep work
We need to fragment and prepare our character before we pipe it into PFlow

01 02

01 Ensure the scene is set This initial step is small but

important. Make sure your scene is in Film/24fps and
100-130 frames long. Also make sure Realtime is unchecked,
02 Check the provided character The initial scene
provided has a character with a PointCache modifier
in it. Re-link the point cache to the file provided, then make a
03 Fragment the
snapshot with
RayFire Select the snapshot
as mParticles playback will be affected by that. Finally, make snapshot since we want to keep the original walking character and start RayFire. If you don’t
sure your units are Generic and Centimeters. Units will have an as the driver for our fragment. Overall it’s always a good idea to have RayFire open the second
effect on the overall speed and look of things. make a copy of your original object and put it on another layer. step as a MAX file, you will
find the character already
broken up in there. In RayFire
Conform meshes for particles load the character into the
Impact Object group and
Once you have the fragments there is one last important
open the Fragmentation tab.
step before we can make those particles! The fragment
pivots need to be central and the Xform should be reset. Choose Voronoi Uniform as
Once that’s done you can convert them all to meshes. The the Fragmentation Type and
pivot centre of each fragment will be the particle pivot. around 1,500 fragments, then
EditMeshes carry less data then EditPolys and are therefore hit the Fragment button. Your
faster, so they should be your preferred mesh choice when
working with particles. result should look like the
03
thumbnail in the screenshot.

3D Art & Design 173

 Animation

Head into PFlow

Pipe the fragments into PFlow and set up mParticles

04 Create the PFlow

It’s now time to set
up the PFlow for the
character and fragmentation
effects. Hit 6 on your
keyboard to open the Particle
View. From the depot drag an
Empty Flow into the view.
Make sure you see 100 per
cent of your particles in the
viewport and that the
Sub-Sampling is set to
Frames. Our character is
moving fairly slow in the
provided example, so in this
case we don’t need require a
04
huge amount of precision.

05 Bring in the fragments Add a Birth Group from the depot and connect it to the root
event. Add all your fragments into the Birth group and hit Update Particles from
Objects. This will make every fragment a particle and will also inherit shaders and mapping.
When you set your Display Node to Geometry you should see the fragments as particles.
05

06
07

06 Lock the particles to the moving character When you scrub the timeline you
will notice that the particles are just standing there. In order to make it move like the
original character we need to unhide it again and in the PFlow create a Lock/Bond test. Pick the Avoid PhysXplosion
character as Lock On Object. Set it to Lock To Surface and Animated Surface. Now scrub and the
shapes should somewhat follow the character’s walking motion. When you scrub the timeline you will see that the particles
just fall down and some might bounce out of the shape.
This is because there is potential shape interpenetration.
07 Get physical In order for mParticles to work we need an mpWorld and an mpShape.
Drag an mpWorld operator into the flow and hit Create New Driver. This will add a
helper in the world centre called mParticle World. Make sure this has Ground Collision and
The engine tries to separate those shapes causing an
explosive motion, so to counteract this it’s important to play
with the Weld Threshold and Interpenetration Tolerance in
Gravity active. Next make an mpShape, which will make every particle a PhysX shape that can the mpShape and the Sub Samples in the mpWorld. Lower
Restitution and Friction values can cure the spasms. It’s up
collide with other shapes. Adjust this to Convex Hull and set the Display to Wireframe. You can
to you to find the best working path depending on what
see that some shapes will overlap due to the Convex nature of the shape. We can counteract this look you are after.
by adjusting the Weld Threshold value and Interpenetration tolerance (see Avoid PhysXplosion).

174 3D Art & Design

 Shatter moving characters

Create trigger geometry

Make mpShapes obey the Lock/Bond and trigger fragmentation

08 09

08 Turn off PhysX features Drag an mpSwitch under mpShape and check Speed and
Rotation to match and follow legacy operators. This will keep them in place despite the
Lock/Bond operator telling them otherwise. Or, you can check Turn Off Simulation and PhysX will
be turned off and will kick in when told to. This can result in explosive behaviours, but it is faster.

09 Create the trigger geometry We want the fragments to fall eventually, so an easy
way is to include geometry that surrounds the particles at some point and triggers an
event change when they are inside the trigger object. For this I just hand-animated two spheres that
trigger the arms and a plane that triggers the remainder top down. You can alter the timing as you
like using your own geometry; just make sure it surrounds the particles you want to trigger.

11

10 11 The second event of free-fall Now that you have a

Test operator to send particles into a new event with new
rules you can design the free-fall part of the effect. Select the

10 ADM inside object trigger Data Test This the only time we are going to utilise ADM.
Add a Data Test under the mpWorld, then open the Data view and drag a Select Object
sub-operator into the view. You need this OP every time you want to select something from your
mpWorld, hold down the Shift key, then drag it down under the
PFlow event. This will prompt you if you want to copy or
instance the operator – pick either one. The mpWorld helper is
scene. Add all your trigger geometry, then drag a geometry sub-operator into the view and connect the box object in the viewport and only references this
it. Your trigger objects are geometry, so this is the OP of choice. Set it to Inside Object and this will operator. Now you should have a second Display operator as
test if a particle is inside the objects or not. Create an Output Test into the view and connect it. well, so pick a distinguishable colour from the first event.

3D Art & Design 175

 Animation

Optimise the flow

Apply new rules to the particle simulation

12 PhysX with a spin As the particles from the first event are loose, the Lock/Bond
should not just fall down and collide with the ground, they should have some spin to
them. Drag a Spin operator above the mpWorld in the second event and play with the Spin
amount you like to achieve the desired effect. In order to make mParticles obey legacy
operators we need an mpSwitch, so drag a second mpSwitch right under the Spin and set it to
obey the legacy spin.

13 Tame the overall effect Some particles might spin a lot and some might bounce
based on your Restitution and Friction settings in the mpShape and mpWorld helper. To
tame this behaviour it’s always good to have a little Drag introduced. mParticles has its own
Drag operator for this, so drag an mpDrag under the mpSwitch. Check that you want to apply
Drag to Angle and Rotation and play with the Amount value until it eventually suits the effect
that you’re looking for.

12

13

14

14 Reduce unwanted jumping or PhysXplosion There are a few factors that can
cause undesired behaviour in your mParticles – usually Restitution/Friction and
Subframe Factor/sampling. A lower sub-stepping might cause less chaotic bounce but less
accuracy along the way. The Sleep Threshold is also worth playing with, as it takes particles out
of the simulation until they are hit by another particle, so raising the thresholds for Energy and
Bounce will tame particles once landed. There is no magic number here to fix everything, just
find a good balance of all these different values.

176 3D Art & Design

 Shatter moving characters

15 Cache or bake the final result If you plan to render

Artist
Showcase
on a render farm it is a good idea to cache the system
either with Cache Disk, which is new to PFlow with ADM, or
with the mParticles World. Set it to Viewport/Render and
check Cache Test Result, then hit Update. If you bake it with the
mpWorld helper click Cache/Bake Particles and let it run
through the timeline, then check Use Baked Cache. This will Anselm von Seherr-Thoss
ensure that every machine sees the same thing at render time. I‘ve created visual effects for movies, commercials
and music videos for about nine years. My special
fields are particle and smoke/fire/fluid simulations. I

1hour
creation time
live and work out of New Orleans, where I run Incendii
LLC Visual Effects. I have worked at studios like BLUR,
Pixomondo, Atomic Fiction, Frantic Films and Psyop.

Resolution:
1,280 x 720

Snow horses 3ds Max 2010, PFlow Toolbox#2/3, V-Ray (2010)

This is a collection of RnD I did for a Gazprom commercial. I want
to show the still I really like – it’s a snow horse. The image is, in its
essence, this tutorial.

15

Bonus round
You could also use the PFlow Baker script (www.scriptspot.com/3ds-max/scripts/
PFlow-baker) and bake the particles into meshes. Then you can pipe those into a new
particle system where you spawn smaller debris from the falling pieces. Adding more detail
by emitting from the baked particle surfaces like this can really enhance the overall effect.
You will find a bonus max file with baked out particles and additional debris with the disc.
Constructor - A Particle System 3ds Max 2010, PFlow
Toolbox#2/3, V-Ray (2010)
You can watch this video in motion at: www.vimeo.
com/14597952. The high-res model was pre-fractured with
deconstructor by Marc Lorenz then passed on to Particle Flow’s
BirthGroup and triggered with a deflector.

Venus 2.0 3ds Max 8, V-Ray (2006)

This was an art piece I did while learning Particle Flow. I used the
PFlow Toolbox#1 and a Max script that leaves particle trajectories
as splines. The statue is the Venus de Milo that stands in the
Louvre in Paris. It was rendered with V-Ray.

3D Art & Design 177

 Animation

Easy-to-follow guides
take you from concept
to the final render
Artist info

Jonathan Cooper
Personal portfolio site
www.gameanim.com
Location Canada
Software used
MotionBuilder
Expertise Making interactive
characters and worlds ever
more believable

ěũũ
2222(-Ĵ4-ēũ
– compatible with
MotionBuilder 2010
onwards

178 3D Art & Design Supporting image is from marketing material and does not necessarily reflect the final look of the tutorial
 Blend cycles

Master
blend cycles
Assassin Sprint 2012
Here we will learn how to create the Sprint Impulsion acceleration
cycle, an animation that captures the Assassins’ free spirit
Jonathan Cooper is an animation art director at Ubisoft Montreal

T
o achieve Assassin’s Creed’s style of animation,
we require a vast amount of animation cycles
and transitions, bringing the assassin to life for
Concept
even the simplest ground movement before we layer Starting with a longer
on cutting-edge technology such as reactive physics shot of a mocap actor
and IK. running, we’re going
Here you’ll learn the process for making a flawless run to be extracting a
section in order to
cycle using motion capture, as well as a relatively quick and create a perfectly
simple looping technique that can be applied to any kind of looping cycle. This
motion where seamless cycling is required. For this example can then be used for
we’re using Autodesk MotionBuilder but the same animating a game
character, such as
techniques can be applied in any software that has the
Connor Kenway.
ability to blend sections of animation with one another.

3D Art & Design 179

 Animation

Direct your mocap

Great mocap starts with good acting, so get the most out of your actors

01 Inspire your actor’s imagination Because the run

cycle is perhaps the single most important animation in
your game, ideally you should have several to choose from,
01

covering a variety of attitudes, speeds and intensities. An

essential tip here is not to ask an actor to simply run fast, but
instead to give him or her visual direction in the form of: ‘you
are chasing after a stolen purse’, or ‘you are escaping a burning
building’ and so on. Employing visual references is more likely
to provide a range of performances rather than simply
adjusting the speed of the capture. You can even try tying
weights to legs, if your actor is willing, or modifying the terrain
for wildly varied motion styles.

02
02 Be kind to the talent Don’t leave high-energy
actions like running to the end of the day’s mocap
recording, especially if your actor is more thespian than
stuntman, or is not in the peak of physical condition. This will
avoid fatigue creeping into your recordings. Additionally,
starting with a quick burst of action is often a fun and
energising way to start off a session. Remember, never criticise
a performance, because actors are emotional beasts that
require encouragement and inspiration to bring something
amazing to their performances. The same can be said for
animators in general.

Acting is for actors

Often an animator’s first approach when directing actors is
to show the action themselves, but this is severely limiting.
We pay actors a lot of money to bring their talents to our
03 Use all of the volume Unlike stationary actions, running requires you to maximise
the use of the mocap volume – the space the cameras can record – as even large studios
typically have only enough room for a dozen or so footsteps. Ensure your actors run diagonally
session, so why not use them? Performances are much
better when an actor feels as though they are more
involved in the process. The direction you provide should
across the volume for maximum recording distance, having them run out of the space before be broad enough to enable them to come up with their own
interpretation of what you originally wanted. Nine times
decelerating if possible. This minimises wastage of mocap data when on a budget and gives you
out of ten you will find that this will be much better than
as wide a range of footsteps to choose from as possible. However, It’s easy to reposition and what you originally envisioned.
realign performances, as we shall see later.

03

180 3D Art & Design

 Blend cycles

Set up your scene

Before editing, it’s important to ensure you have a cleanly set-up scene

04 Convert to the control rig Once your mocap is delivered, the real fun begins! Before
we can edit anything in MotionBuilder, we must move the motion from the skeleton to the
control rig. Do this by selecting Bake(plot)>Bake(plot) to Rig from the large blue character button. You
may also want to deselect Constant Key Reducer in the Options box to retain maximum fidelity. Hit
Ctrl+A to display the control rig in the viewport if it isn’t visible already. Get into the habit of plotting
between the rig and skeleton often, for an easy before/after comparison between changes.

05

05 Enable the Story mode In order to broadly manipulate the motion, we’ll use
MotionBuilder’s Story mode. While mostly used for editing cinematic cutscenes, Story
mode is also a powerful blending tool. Open the Story tab in the Navigator and turn Story mode
on if it isn’t already. Right-click the space to the left of the timeline and choose Insert>Character
Animation Track. Choose your character from the Character dropdown menu, then right-click in
04
the empty timeline next to your newly created character track and select Insert Current Take.

06
06 Select a segment From the variety of examples,
you’re looking for one that best suits the personality
of your character, keeping the rest as backups for other
characters. You’ll want to select a range of at least three steps
that have a constant momentum. Ignore the acceleration and
deceleration likely captured at the extents and avoid noticeable
actions that would stand out in a repetitive loop. Using the
Assassin_Run.FBX file we’ll drag the edges of the clip to frames
305-329. You’ll want to start and end cycles on a passing pose,
with one leg up, to make similar poses most recognisable.

07 Scrub your animation Moving the clip back to

zero, you can make scrubbing through keys more
readable by zooming to the extents in the timeline. Do this by
manually typing in the start/end times in the S: and E: boxes at
either side of the timeline, as well as in the story by clicking on
your clip and hitting F. Note that the Story timeline differs from
the viewer’s – the one that you’re editing and one that you’ll see
when you press Play. You can jump back and forth one frame
at a time by hitting Ctrl, then the left/right arrow keys.

07

3D Art & Design 181

 Animation

Refine and repair

Clean up your clips and locate the ideal join

08 Clean up your scene With the clip at zero, notice

that the numbers above and below it are out of sync.
The top numbers are scene-based and the bottom are internal
08
to the clip, which can cause issues on earlier versions of
MotionBuilder when copying and pasting clips. Plot to the rig by
right-clicking it and selecting Plot Whole Scene To Current 09
Take. Right-click and delete your original Story clip and plot
back from the rig to the Story tab once more. This resets the
clip’s local timeline and is a good habit to get into to better
understand the real length of clips when you move them
around. The Global In/Out indicator is the clip’s position in the
scene, whereas the Local In/Out setting is local to the clip.

09 Align to an axis Now that you have isolated what

you want and because we captured the motion at an
angle, we must align to an axis to make the looping workable.
You can edit the position and rotation by clicking on the Show/
Hide Ghost (eye) button in your character track. With the clip
highlighted, you should now see a green line that represents
your character’s travelling motion. On the first frame of your
range, select the Translation tool and enter ‘0.00’ into the X,Y
and Z fields. Now select Rotation tool and manually rotate the
orientation to have the green line best match the Z axis. The
Top view, via Ctrl+T, is the best angle to rotate to the axis. Hit
Ctrl+E to return to the Perspective view.

Rapid iteration is key

Working in this manner to quickly create cycles enables us
to see them where it counts – in the game engine. An
animation is never done until it’s fully playing and blending
in real-time with all the other cycles and transitions. Only
then can we tell whether the action is good enough and fits
the character’s personality. Videogame animators often
have many animations being worked on at once and make
adjustments before exporting, review while playing, then
rinse and repeat in this manner until the game not only
looks good but feels good with the controller in our hands.

10
10 Find the best join
This is the key
element of what we are trying
to achieve, making the join
seamless and unnoticeable.
It’s important to identify two
complete footstep actions –
not just single key poses
– that are most similar. In our
example we chose only three
steps, so notice the first and
third steps are the two similar
actions. We will be using the
latter two as the basis for our
cycle. This means that frame
8 will then be our starting
pose. Here we’ve identified
the second passing pose as
our best join frame.

182 3D Art & Design

 Blend cycles

Blend the motion

The crux of this process is choosing to match motions rather than poses

11

12

11 Loop the footstep Now we are going to blend back

across one full step to ensure the looping motion is
seamless, with the momentum of every body part retaining
13

constant velocity throughout. We need to avoid the noticeable

hitch that would occur if we simply pose-matched the first and
last frames. What we’re essentially doing here is, over the
duration of the final footstep, blending back into the keys before
the starting footstep, so we end our cycle at exactly the same
pose as the start, with all the same momentum as the first
frame. In our three steps, there are two similar left-to-right
actions we can blend across.

12 Blend the motion back onto itself With passing-

pose 2 (frame 8) as our desired start frame, shrink the
clip start to reflect this. Now right-click the clip, select Copy,
move the timeline slider to exactly after the clip, right-click the
track and select Paste. We now have two identical clips
immediately following one another. Ignoring forward
movement for a second, jump into the Front view via Ctrl+F.
Select the second clip and drag its start half-way across the
original. The X shape in the Story tab shows the original clip
blending back across the keys prior to its first frame.

13 Adjust the blend duration Now set the timeline to

be the length of the first clip only, (8-24). Ensure looping
is enabled via the Loop button to the right of Play and hit Play.
You now have a seamless loop. Play around with how long you
blend back across the cycle to itself. Longer durations provide
the most seamless results but will increase foot-sliding,
whereas shorter durations are most stable but will increase the
visibility of the join. Only use the Front view at this stage in
order to see the seamless looping.

Cycling on the spot

Never remove the forward (Z axis) movement keys from
your character’s root/pelvis bone to see your character
cycling on the spot, unless you want them to plane forward
in an unnatural manner when you key them forward again.
We don’t move linearly in real life and instead push a little
with each step in a rhythmic fashion. To see a stationary
cycle, add a second layer and key the character backwards
linearly, removing the layer again later if required for
exporting into your game engine.

3D Art & Design 183

 Animation

Close the loop

All that remains is to clean up the action to make it game-ready
14 15 0?

14 Match the forward position Moving back to the

Perspective view, we still have the issue of the clip
returning to its starting forward position. Rectify this by
selecting the Translation tool once more, entering the Side view
via Ctrl+R and repositioning the second clip forwards or
backwards until you reduce foot-sliding, which is bound to
occur during the blend anyway. MotionBuilder has an Auto
Match button in Story mode, but because we only want to
match the forward position and not the up/down and left/right
we can avoid it here. You can use the handy Ghost skeletons to
line up your clips with minimal guesswork.

16

15 Clean up the foot Plot from the Story to the rig and back while deleting old clips to
produce one clean clip, which you should reposition back at the origin. Plot the Story edit
onto the rig one last time and turn off Story mode. You’ll likely have some lateral foot-sliding
remaining that occurs across the blend, so eliminate this by modifying the offending keys,
ensuring you don’t change the initial or final pose. Cleaning up the minor foot-sliding is easy with
MotionBuilder’s FCurves. In the Front view, select the left foot controller and open the FCurves
window in Navigator. Select the Translation X curve and modify the foot inwards before it lands.

Not just for running

This technique can be used for any cycling action you wish as long as you blend across two
similarly large motions rather than two subtle ones. This masks the blend more effectively
and reduces the foot-sliding and other cleanup work that follows. Assassin’s Creed required
many cycles to blend between in real-time – sometimes only for a few frames at a time. With
this handy technique, your own projects can also have many quickly created cycles, but how
creatively you use them is up to you!

16 Improve the silhouette From this point onwards it’s

important to retain the cycling motion while editing. In
Assassin’s Creed everything we avoided in this tutorial, such as
17 Final adjustments Let’s quickly raise the shoulders
for a more heroic posture. Turn on AnimLayer1 and the
AutoKey (key) buttons, as well as selecting Full Body in the
17

asymmetry, acceleration and deceleration, was re-created in Body Parts dropdown menu, all under the Key Controls tab.
the game engine to give personality and a sense of effort. How Next set Reach R to 100% for the head and shoulder
much work you have remaining depends on how close your controllers under the Character Controls in the upper right.
original mocap was to the desired result, but now that we have This enables us to blanket keys across the entire animation
a flawless cycle an animator is free to begin the real creative unhindered. Finish off by pulling the shoulders up and back and
work of adding appeal and personality. Remember that while tilting the head down slightly on your cycle’s first frame for that
videogame animations need to look good from 360 degrees, classic Assassin’s Creed look. Now we’re ready for a first pass in
runs are seen mostly from the rear. the game engine – many more remain!

184 3D Art & Design

 Gallery
Artist info

Koen Koopman

Username: TBKoen
Personal portfolio site
www.tubuh.nl
Country The Netherlands
Software used 3ds Max,
Mudbox, After Effects,
Photoshop, V-Ray

Work in progress…

This was an image I created for fun and

practise, revised from an older version with my
current knowledge of sculpting, lighting, colour
palettes and composition
Koen Koopman General Shellshock, 2013
3D Art & Design 185
 Animation

186 3D Art & Design

 Animate vehicles

Animate
vehicles
One giant leap
for car-kind 2013
Jahirul Amin is a freelance rigger, animator and an
associate lecturer at the NCCA, Bournemouth

Easy-to-follow guides
take you through the
creation process
Artist info

Jahirul Amin
The control rig
Personal portfolio site
www.warpeaceandpixels.com
Country UK
Software used Maya
Expertise Jahirul is an expert
animator and rigging genius

ěũ 8ũ2!#-#ũăũ+#2
ěũ("#.ũ-(,3(.-2
ěũ#+/$4+ũ2!1(/32ũ$.1ũ 8
ěũ43.1(+ũ2!1##-2'.32

3D Art & Design 187

 Animation

Animate
Hey, who took
the bridge?

vehicles in Maya
H
aving modelled and rigged our car, we can add texture to the piece. However, there’s no enable you to see the path your object (here, the
now finally take it out for a spin. We’ll be substitute for first-hand resources, so if/when you car) is following. They also enable you to edit the
hurtling down a stretch of rough road, then drive, really feel the way the car moves in response motion live in the viewport.
taking a jump over a broken wooden bridge. This to changes in surface, shifting gear and so on.  Before we hit the steps, I’ll run through the
little stunt will definitely test out our rig and we’ll In terms of approach and technique, we’ll come process briefly. First we’ll set our animation
need to play with the suspension and tip the car at this task as if we were completing a character preferences, create a custom shelf, frame our work
using the pivot controls. We’ll also have to make animation. We’ll lay out the broad strokes to begin and make a camera to work to. Then we’ll block out
sure the car doesn’t penetrate the environment as it with and try to get the timing right before adding the motion of the car, first moving it from A to B,
travels over the bumpy surface, most likely by any finer points. When the timing is just right, we then working and reworking the timing to get it
adding extra keyframes, then finally adding weight can add detail layer by layer, tweaking that timing as looking believable. Next the layering begins, so
as the car hits the ground after the jump. So grab we go. The problem with attempting any detailed ensure the wheels don’t penetrate the surface of the
the supplied tutorial files and let’s take a look. work before the timing is sorted, is that if you need road and add weight by playing with the suspension.
Thanks to the millions of petrol heads out there, to significantly rework it, you’ve got to undo the Finally we’ll add details such as the damage to the
you’ll find more reference than you could ever need detail added on top in order to fix it, which I find car on landing, the movement of the wooden bridge
online. This is especially handy to help you work out pretty soul-destroying. Technique-wise, we’ll be struts as the car hits them, and animate the
the timing, as well as to observe small details that using the very useful Editable Motion Trails, which character in the driving seat.

Start the engine Modifications to the rig

Check the settings & create a camera I’ve made a few tweaks to an original rig I created to make it
fit for purpose. First I added an attribute to allow for
pivoting independently from the driver or the passenger
side. I also created a control for the steering wheel and
hooked it up to the Front Wheels Turn attribute, so the
steering wheel will turn with the tyres. Finally I popped a
character (Box Boy) into the car, parented his IK hand
controls to the steering wheel and constrained his root_ctrl
to the car’s body_ctrl. Check out the scene file ‘rapier_rig.
ma’ to see these changes and have a play around before
beginning the tutorial.
01

01 Tweak the animation preferences

Before we drive, we need to check the animation
settings. Go to Window>Settings/Preferences>Preferences
and scroll down to the Settings category. Under the Working
Units, set the Time to PAL (25fps) and then go down to the
Time Slider tab. Under Playback, set the Playback Speed to
Real-time [25fps]. If you’re used to working at 24 or 30 frames
per second, feel free to set it to this. As the car will travel
02
between keyframes, set the Default In and Out Tangents to
either Auto or Spline under the Tangents in the Animation tab.
03 Custom shelves & shortcuts Create a new
custom shelf and call it ‘car_anim’. In this shelf we’ll

02 Place the camera Go to Create>Cameras>Camera

and rename this ‘render_cam’. Open the Render
Settings under Window>Rendering Editors and set the Image
make a shortcut to return to our camera view, as well as a
shortcut to select the main controls that we’ll be keying in this
shot. Looking through the render_cam viewport, go to
Size to either HD 720 or HD 1,080. Select the newly created View>Bookmarks>Edit Bookmarks. Select New Bookmark and
03
camera in the Outliner, then in the viewport panel go to then Add to Shelf. If you aren’t quite decided on a final camera
Panels>Look Through Selected. Position your camera where angle, create a few bookmarks to switch between. Next select
01 Setting Time preferences you think is best and turn on the Resolution Gate to help you the following controls: main_ctrl, body_ctrl, fpTyre_ctrl,
before we start frame your shot. I ended up tweaking the camera slightly fdTyre_ctrl, rpTyre_ctrl and rdTyre_ctrl. Open up the Script
02 Framing the shot throughout the shot, so as to frame the animation better as I Editor and highlight the commands from the History and
03 Creating camera bookmarks
progressed. However, working to one main camera means you middle-mouse-drag them onto the custom toolbar. When it
and selection tools can cheat a few things should you need to. prompts you to Save Script to Shelf as Type, select MEL.

188 3D Art & Design

 Animate vehicles

Move from A to B 04

It’s all in the timing; details can wait

04 Apply guide curves Before I begin animating, I like to create a curve to help me figure
out the path the object will follow. Go to Create>CV Curve Tool or Pencil Curve Tool and
make sure the Curve degree for either is set to 3 Cubic. Navigate to the top view and create a nice path
from the top-right corner of the environment and over the bridge. If you’re using the CV Curve Tool,
make sure you make several clicks from start to finish, enabling better editing of the curve. Now tweak
the vertices, as they’ll currently be flat on the surface. Once you’re happy with this, use the car’s
globalSRT_ctrl to position and rotate it in the direction of the curve .

05 Block out the controls To translate the main mass

of the car we’ll block out the main_ctrl. Only focus on
translating the car in the X and Z axes initially and don’t worry
05

about the up-and-down translation too much, other than on the

bridge. When the timing is refined later on, we’ll come back
and clean the up-and-down translation, so the car will sit on the
surface better. As we have to deal with both Translate X and Z
in the Graph Editor, things can get pretty tricky when trying to
get them to work together. Luckily we can use the Create
Editable Motion Trail tool, which you’ll find in the Animate
menu. With the main_ctrl selected, run the tool to create the
curve of the path the car is following. You can now – live in the
viewport – use the handles to edit the trajectory. This is a very
useful tool when animating a shot like this.

06
06 Re-time the animation & make an impact
Once the main blocking is complete, we can really
start concentrating on the timing and try to get it to a state
where only minimal changes may be needed later. Try to think
hard about the weight of the car and how the front (where the
majority of the mass is) will behave. Use the Front Wheel Pivot
and the Rear Wheel Pivot attributes to lift the car up from the
broken bridge, and then also to add weight as it comes crashing
down. As the car comes down, we can consider the principles
found in a bouncing ball animation: the spacing increasing
gradually and then the sharp contact, to help convey a sense of
weight and impact.

07 Replicate the drag At this point we’ve got the car

running at some speed and the first turn is pretty
sharp. To help strengthen the idea of a speeding car, use the
Front Drag attribute on the main_ctrl to delay the rear of the
car. This will also help enforce the idea that the main weight of
the car is at the front, causing the rear to fall behind. As the car
Where is the weight? begins to straighten up before the jump, you could also add
During this process I was constantly some drag in the opposite direction. Doing this will bring a little
considering where the weight was in the car. texture to the animation to prevent it becoming too clean-
For this model the engine would be at the looking and boring.
front, so as the car swings the rear will drag.
As the car comes down from the jump, it’s the
07
front that will lead and hit first. Stunts like
those performed in The Dukes of Hazzard had
the boot stuffed with weights to soften the
landing and stop the car from nose-diving.
Think about what would be physically
plausible first and if you need to exaggerate
slightly, do so.

04 Using curves as guides helps 06 Bring impact to the landing

to project the flight path using the Graph Editor
05 Use the Dope Sheet to help 07 Add texture by dragging the
edit the timing rear of the car

3D Art & Design 189

 Animation

Adjust the tyres & suspension

Apply weight & keep the tyres on track to maintain the illusion

Motion trails
Whether you’re animating
a character or a
hard-surface object, you’ll
want to make sure the
arcs created are appealing
to the audience’s eye. I’ve
supplied a script created
by a colleague,
Constantinos Glynos,
called ‘CG_ 08
VertexMotionTrail_v2.py’.
This enables you to take a
selected vertex and create
a motion trail from it. Here
I’ve checked that the trails
08 Stay on the surface By now we should have the
main beats of the animation, so we can begin cleaning
up and layering in the details. Go through the tyres one control
09 Animate the suspension Once the wheels are
sitting well on the surface, use the body_ctrl to bring
the feeling of suspension as well as further appeal to the
created by the tips of the at a time, using Translate Y to keep them on the surface rather animation. As the car makes the first big turn, rotate the body_
car’s wings flowed.
than penetrating it. Keep the geometry of the wheels and the ctrl away from the turn and then hit it in reverse. Delay the
environment in smoothed mode, as viewing unsmoothed may body_ctrl so it’s always trying to catch itself. As the car comes
give slightly odd results. As the surface is so bumpy, especially crashing down, you can really help sell the impact by rocking it
as it travels up and down the bridge, you’ll probably find that back and forth a few times before settling. With the
you have no choice but to set a key on almost every other combination of the movement in the wheels and the
frame – if not, a frame here and there. suspension, this will help add some life to the animation.

09

10
10 Tweak the tyres
For the wheels,
disable the autoSpin attribute
11 Place the driver To have this car with either blacked-
out windows or no driver would look odd, so pop BoxBoy
in to let him take it for a spin. We can use this to our advantage,
on the main_ctrl and use the as it enables us to replicate the bumpy ride. By adding some
Spin attribute on each tyre delay and overlap to the upper body and the head, we can sell
control. This enables us to the idea of this car careening around and slamming down. You
edit the speed of the tyres don’t have to spend too much time here as the motion of the
spinning at different stages. car is fast, but even these small details help add some reality to
For example, when the car is the animation. You’ll also have to use the Elbow attributes in the
travelling through the air, the IK hand controls to stop the elbows from popping about.
front wheels need to slow
down but the back wheels 11
should spin at a faster rate.
You can apply something
similar for when the car
makes its first major turn. You
may find that due to the style
of the hubcaps it’s pretty hard
to see the wheels spinning,
but adding motion blur can
help sell this effect. To turn 08 Make sure the wheels aren’t 10 Turn the wheels and add
penetrating the surface some spin to them
the front wheels, use the
09 Add more weight by rocking 11 Create further drag and
Front Wheels Turn attribute and tilting the main body of overlap by animating the
on the main_ctrl. the car BoxBoy driver

190 3D Art & Design

 Animate vehicles

Add more details & texture

Small points will make all the difference, so let’s add some

12 Include bumps & breaks In order for the front of the

car to crunch the floor on landing, the licence plate will
have to penetrate the floor for a couple of frames. Although
12

there’s no control for the plate, we can animate it to move

slightly backwards as it contacts the floor. You may have to edit
the pivot point to be able to rotate it from a better position. To
do this, simply hit the Insert key on the keyboard before you set
your keyframes.

13 Apply further jitter For greater interest, we can

include more detail for the car’s vibrations and jitters.
Use the autoJitter_ctrl with the following settings: Master Ctrl
Speed: 1, Master Ctrl Size: 1, Body Rock Speed: 6, Body Rock
Size: 0.2, Body Up Down Speed: 5, Body Up Down Size: 0.2,
Bonnet/Boot Speed: 7, Bonnet/Boot Size: 0.1, D Door/P, Door
Speed: 7, and D Door/P Door Size: 0.05. Try experimenting in a
clean scene with just the car, as it’s hard to see the effects in a
fast-moving vehicle – but they do add to the overall animation.

12 Apply some minor damage 14 Use a bend deformer to add

some vibrations to the struts
13 By adding subtle jitter on top
of our main animation we can 15 Clean up the animation’s 13
increase the believability curves before signing off

14 Interaction with the environment The last

additional details can affect the struts as the car lands.
To avoid animating each piece of wood individually, select the
majority of the horizontal struts and navigate to Create
Deformers>Nonlinear>Bend. Rotate the bend deformer 90
degrees in the Z axis and animate the Curvature attribute,
found under Inputs, with very small numbers. I’ve hand-
animated some of the vertical struts using the Rotate Z channel
to reinforce the illusion of weight slamming down.

15 Clean up the animation curves Finally, check the

curves and clean them up or delete any keys from
attributes that have no animation on them. You’ll know which
attributes have no animation as you’ll see a straight line in the
14
Graph Editor. And now, you can take that brisk Sunday drive!

Light up your animation

Lights are awesome and I have a tendency to use them
often when checking my animation. I like to create a
directional light with Use Depth Map Shadows turned on
and an ambient light with a very low Intensity of around
0.25. Adding Viewport 2.0 and turning on the Screen-space
Ambient Occlusion really helps to show you if your models
are penetrating the environment. Minor penetration isn’t a
bad thing, as it helps to sell the idea that one object is sitting
on top of another – but go too far and it will look wrong!

15

3D Art & Design 191

 of al
r
al ci
fe
tri Spe

Enjoyed
this book?
Exclusive offer for new

Try
3 issues
for just
£5 *

* This offer entitles new UK direct debit subscribers to receive their first three issues for £5. After these issues, subscribers will then pay £21.60 every six
issues. Subscribers can cancel this subscription at any time. New subscriptions will start from the next available issue. Offer code ZGGZIN must be
quoted to receive this special subscriptions price. Direct debit guarantee available on request.
** This is a US subscription offer. The USA issue rate is based on an annual subscription price of £65 for 13 issues, which is equivalent to $102 at the time
of writing compared with the newsstand price of $14.99 for 13 issues, being $194.87. Your subscription will start from the next available issue.
 The UK's best 3D magazine
dedicated to the 3D artist and
the 3D community. You've
never seen a magazine so
stuffed full of gorgeous 3D
artwork before…
Published 13 times a year, 3D Artist delivers the finest

About
tutorials, features and interviews with people working
in the industry, in addition to the reviews and what

the the latest happenings are in the 3D community. Each

issue also comes with a free CD that is packed with

mag high value 3D content such as software and models.

subscribers to…

Try three issues for £5 in the UK*

or just $7.85 per issue in the USA**
(saving 48% off the newsstand price)
For amazing offers please visit
www.imaginesubs.co.uk/3da
Quote code ZGGZIN
Or telephone UK 0844 249 0472 overseas +44 (0) 1795 592 951
iPad iPhone Android phone Android tablet Apple Mac Windows PC
 /P
%JTD
/P
1SPCMFN
Many of the files you’re looking for can be
found on the magazine’s website
*NBHJOF
EJHJUBM
FEJUJPOT
To get the most out of your digital
editions, be sure to enjoy all of our
BSF
B
OFX
BOE
FYDJUJOH
fantastic features, including:

XBZ
UP
FYQFSJFODF
PVS
r
;PPNBCMF
UFYU
BOE
QJDUVSFT
r
*OBQQ
CSPXTJOH
XPSMEMFBEJOH
NBHB[JOFT
r
4FBSDIBCMF
UFYU
BOE
CPPLB[JOFT r
5BLF
ZPVS
DPMMFDUJPO
XJUI
ZPV
r
3FBE
PGá
JOF

5P
CVZ
NPSF
*NBHJOF
EJHJUBM
FEJUJPOT
BOE
GPS
UIF

MBUFTU
JTTVFT
BOE
CFTU
PGGFST
QMFBTF
HP
UP XXX(SFBU%JHJUBM.BHTDPN
 The
3D Art
& Design
The ultimate guide to getting
the most out of your
3D projects

Includes nd
sa
advice, tips from
techniquestry
indu
experts

VOLUME 03

www.imaginebookshop.co.uk