Data Knot is a set of machine learning tools that are optimized for low latency and real-time performance. The tools can be used with any type of audio input, including optimizations for Sensory Percussion sensors and ordinary drum triggers.

Data Knot includes low latency onset detection, onset-based and real-time descriptor analysis, classification and clustering, regression and prediction, corpus analysis and querying, synthesis and modeling, audio, data, and CV processing, and a slew of other abstractions that are optimized for low latency applications.

Data Knot is built around the FluCoMa Toolkit and requires v1.0.9 or higher to be installed for this package to work.

You can join the Data Knot Discord server to keep up on future updates/news as well as make suggestions, share corpora and musical examples etc...
https://discord.gg/Vy2zuKQkqN

Max 8.6.5 or higher (Mac/Windows).
FluCoMa v1.0.9 or higher - installed from Package manager or here
All abstractions work in Windows, 64-bit, and M1/Universal Binary 2.

Intro to Data Knot
What's New in Data Knot v1.0
Performance and Musical Examples

Onset Detection Tutorial
Descriptors Tutorial Creating a Corpus Classification Tutorial Regression Tutorial Sensory Percussion Setup

• BREAKING CHANGE - @enable renamed to @active in dk.datagranular, keeping it in line with other parameter names elsewhere
• BREAKING CHANGE - @roundrobin now only has two options, 0 or 1, removing the option to manually set how many values are used for the @roundrobin attribute
• added dk.metro~, a metro alternative with some quirky features
• added dk.onsetdisplay~, a UI utility for fine tuning dk.onset~
• added new buffer/list utilities (dk.buf2list and dk.list2buf) for changing formats from buffers/lists while remaining in the high-priority thread
• added a new @extractmode (sinusoidal decomposition) to dk.transientextract~ via the @extractmode sines attribute
• added @inputfilter to all dk.descriptors~ variants (d.descriptorframe, dk.descriptors~, dk.descriptorsrt~) that allows you to use the same filter messages as you would with dk.filter
• added new @matchcount attribute to dk.controllermatch, dk.corpusmatch, dk.corpussampler~, and dk.gridmatch that sets the number of matches that are returned as iterated items playing them back as a "chord"
• added new bending modes to dk.databending (@jitter for timing jitter, and @repeat for data repeating)
• added a new outlet to dk.corpusmatch and all other corpus-based objects (dk.controllermatch, dk.gridmatch, etc...)
• added new dump and dumpdict messages to all corpus-based objects that outputs the current entries (filtered or otherwise)
• added a new outlet to dk.filter which reports whether the criteria was met or not
• dk.filter now allows you to filter messages from dk.spectralshape~ and it's variants (dk.spectralframe, dk.spectralframert~)
• dk.filter now accepts a filter reset message to clear the current filter
• dk.regressorcreate~ now accepts messages for fluid.mlpregressor~, allowing you to fine tune the regressor settings as desired
• added new @maxbuffertime to dk.onsetframe~ and dk.realtimeframe~ for setting the recording time in the circular buffer
• added a new sample folder of CD-skipping samples (for dk.metro~ helpfile)
• added 32 sample window analysis to dk.setupcreate~ for future additions (you should reanalyze your setups)
• added new flags to dk.classcreate and dk.clustercreate for future additions and improved metadata accuracy (you should recreate your classifications)
• all files that load .json files will now check to see if the right type of file is being loaded and throw an error if not (e.g. loading a concat analysis .json into dk.corpusmatch)
• all descriptor objects (dk.descriptors~, dk.melbandframe, dk.classspectralshape~, etc...) now return buffer names that reflect the contains of the buffer (e.g. 22233descriptorbuffer)
• all synthesis algorithms (dk.avdog~, dk.rungler~, etc...) can how accept all the @outputmode options (iterated, lacedlist, rawlist) from dk.regressor
• all synthesis algorithms (dk.avdog~, dk.rungler~, etc...) can now accept the output from the dump outlet of the same algorithm directly for chaining synths or for manually storing the preset dict for recalling
• you can now stop and restart training in dk.classcreate, allowing you to build up the classes slowly rather than having to do them all at once
• dk.filter will now accept arguments in addition to @filter attributes, simplifying using dk.filter
• dk.plotter can how highlight multiple items at once
• dk.corpusplayer~ / dk.sampler~ should respond more quickly to killall and stop messages
• added more clarification for creating a dict for your own synthesis algorithm in the dk.regressorcreate~ helpfile
• fixed a bug where many corpus loading objects were not properly filtering based on index

• BREAKING CHANGE - Project name changed from SP-Tools to Data Knot. All abstraction names changed from sp.* to dk.*.
• BREAKING CHANGE - dk.corpusconvolver~ now contains internal onset and descriptor analysis, no longer requiring input from dk.descriptors~
• BREAKING CHANGE - @xy renamed to @xyout in dk.gridscale allowing for more scaling options
• BREAKING CHANGE - @ramp renamed to @ramptime in dk.scatter~, dk.scramble~, and dk.waveguidemesh~ to disambiguate it from a synthesis parameter
• BREAKING CHANGE - @absolutelength renamed to @lengthabsolute in dk.triggerlength~ keeping it consistent with the order of words elsewhere
• BREAKING CHANGE - @train renamed to @trainmlpin dk.setupcreate~ keeping it consistent with dk.classcreate and dk.clustercreate
• BREAKING CHANGE - @gain in dk.sampler~ changed from dB to %, keeping it consistent with all other uses of @gain across the package
• BREAKING CHANGE - @pitch, @loudness, and @spectral renamed to @pitchcomp, @loudnesscomp, and @spectralcomp across all concat/corpus players (dk.concatplayer~, dk.corpusplayer~, etc...) to avoid conflicting naming overlap when nested in abstractions
• BREAKING CHANGE - load and loadfolder deprecated for all objects that load files (.json or audio files). Instead you now send a read or readfolder message no additional arguments and it will open the file dialog menu
• BREAKING CHANGE - read and load functionality renamed to readfolder and loadfolder in dk.corpuscreate to accomadate for the ability to load a single sample
• BREAKING CHANGE - reset message deprecated for all objects that accepted both reset and clear messages
• BREAKING CHANGE - slew in dk.melbandfilter~ renamed to ramptime, keeping it in line with other parameter names elsewhere
• added dk.avdog~ as an AVDog-inspired synth voice
• added dk.corpusclustersampler~ as an all-in-one version of corpus clustering and playback (in line with dk.corpussampler~ and dk.concatsynth~)
• added dk.crosstalk~ as a way to remove crosstalk from multiple simultaneous sources
• added dk.cvcontrollers~ as a way to extract meta-signals from control voltage input
• added dk.cvcross~ as a way to seperate control voltage into low and high frequency components
• added dk.cvgate~ as a way to send incoming control voltage to mutilple outputs
• added dk.cvonset~as CV-based onset detection
• added dk.cvscatter~ as a way to send incoming control voltage to mutilple outputs
• added dk.cvsmear~ as a way to send incoming control voltage to mutilple outputs
• added dk.fourses~ as a Fourses-inspired oscillator
• added dk.gongue~ as a Gongue-inspired drum voice
• added dk.justfriends~ as a Just Friends-inspired oscillator/modulator
• added dk.mangrove~ as a Mangrove-inspired oscillator
• added dk.melbandfilter~ to allow for melband analysis-based filtering
• added dk.onsetfreeze~ as an onset-based freeze
• added dk.onsettap~ as onset-based delay taps
• added dk.quantussy~ as a Quantussy-inspired oscillator
• added dk.regressor as a way to output parameter regression based on new input
• added dk.regressorcreate~ as a way to create a regressor based on training data
• added dk.rungler~ as a Rungler-inspired oscillator
• added dk.snare~ as a Plumbutter snare-inpsired snare voice
• added new descriptor type (spectral shape) and corresponding objects (dk.specltralshapeframe, dk.spectralshape~, dk.spectralshapert~, dk.classspectralshape~)
• added dk.spectralshapefilter~ as a way to create filters from the output of any of the spectral shape objects
• added dk.transientextract~ as a way to extract transients from incoming audio
• added dk.transientreplace~ as a way to replace transients in incoming audio
• added dk.triggerdelay~ to create additional timed triggers/gates from incoming triggers/gates
• added dk.ultrasound~ as a Plumbutter-inspired Ultrasound filter
• added dk.waveguidesnare~ as a physical model of a snare
• added weights messages to all corpus/concat matching abstractions (dk.corpusmatch, dk.corpusconvolver~, dk.concatmatch, etc...) for varying the relative weight of each descriptor
• added dump message and corresponding outlets to all synthesis objects (dk.karplus~, dk.waveguidemesh~, etc...), dumping out parameter names, ranges, and values
• added the ability to load a folder of files into dk.concatcreate via the readfolder and loadfolder allowing all concat synthesis abstractions to load the folder of samples for resynthesis
• changed analysis hop size in dk.concatcreate (from an overlap of 8 to an overlap of 4) to make analysis faster and shrink analysis file - you should renalayze your concat corpora
• added the ability to load a single file into dk.corpuscreate via the read and load messages which will then be sliced for corpus-based playback elsewhere in the library (dk.corpusmatch, dk.corpussampler~, etc...) - you should reanalyze all your corpora
• added onset detection messages/attributes (sensitivity, lockout, floor, etc...) to dk.corpuscreate to slice a single file when loaded via the read and load messages
• added a new outlet to corpus matching abstractions (dk.controllermatch, dk.corpusmatch, etc...) showing the buffer~ slicing
• added @invertmatching to dk.corpusmatch and dk.corpusconvolver~ that allows you to return the furthest matches in the corpus (rather than nearest)
• added @active attribute to all matching-based abstractions (dk.classmatch, dk.corpusmatch, dk.concatmatch, and all their variants, etc...) allowing for quickly enabling/disabling the matching process
• added the refer message to all corpus and concat match objects (dk.corpusmatch, dk.corpussampler~, dk.concatmatch, dk.concatsynth~, etc...) allowing you to point the abstractions to an existing buffer~ causing it to be immediately analyzed and made available for playback/resynthesis without having to save anything to disk first
• added outputs to trigger-based objects (dk.triggerbounce~, dk.triggercloud~) for indicating when process is complete (for chaining processes)
• dk.filter can now parse messages from the class-based descriptor objects (dk.classdescriptors~, dk.classkitchensink~)
• added @select attribute to dk.descriptorframe, dk.descriptors~, and dk.descriptorsrt~ allowing for selecting only a subset of the descriptors to come out of the list output (buffer output is unaffected)
• added @select attribute to dk.controllers, and dk.speed allowing for selecting only a subset of the meta-descriptors to come out
• added gamepad-specific features to dk.gridscale making it a more general purpose 2d scaling and transformation abstraction, including @deadzone, @curve, @zoom, @transformation, etc...
• added @input modes to dk.spectralonset~ allowing for optional filters optimized for different input types
• added @every attribute to dk.probability~ allowing for passing on only every nth event (similar to dk.sift~)
• added @smearcurve attribute to dk.smear~ allowing for further shaping of the smearing envelope
• added @envelope attribute to all onset-based objects (dk.onset~, dk.descriptors~, dk.corpusconvolver~, etc...) to allow for the use of logarithmic (default) or linear differential thresholding in amplitude-based onset detection @mode (primarily useful for material with large/loud sustain like cymbals and gongs)
• added @offsetmute attribute to dk.lpg~ allowing for the use of the offset parameter to manually control the lpg filter
• added overdub message to dk.datalooper~ allowing the layering of data onto existing loops
• added train message to dk.clustercreate and dk.setupcreate~ to have parity to dk.classcreate (dk.setupcreate~ no longer has a done message)
• added midinote to dk.sampler~ to change the default MIDI note when transposing samples
• added @size, @color, and @algorithm attributes to dk.plotter and made a 4d projection the default for corpus matching
• added @threshold to dk.descriptorstrt~ allowing loudness thresholding below which no descriptors will be output
• added @range to dk.triggerdelay~ allowing for random selection of delay times from a given range
• all ..create objects (dk.classcreate, dk.clustercreate, dk.setupcreate~) now show entries counting from 1 instead of 0
• dk.classcreate now shows the amount of entries per class, rather than a running total, as you create a classifier
• added clear message to dk.shattercreate~
• added clear and reset messages to dk.scramble~ allowing you to directly control active components
• added a new corpus (corpus_banjo.json)
• the included Voice corpus has had all samples increased by 6dB so it more closely matches the level of other corpora
• created dk.demosound to use as the main musical example UI across all the help files and musical examples
• all help files revamped and consolidated in many places to make them as light-weight as possible
• dk.plotter can now take messages prepended with dictionary allowing you to connect the output of dict and fluid.dataset~'s dump outlet directly to it
• dk.concatcreate/dk.corpuscreate now have their metadata prompt window happen when sending a write message, rather than when loading files to analyze
• dk.boppad had a number of bugs that were fixed with regards to quadrant recognition
• made metadata pop-up window more consistent with a pop-up window (no toolbars along the sides etc...)
• added progress bar to dk.corpuscreate
• added @scale attribute to dk.melbandframe and all melband objects to output units in dB rather than amplitude
• all melband-based objects (dk.melbandfilter~, dk.descriptordisplay, etc...) will all auto-adapt their functionality based on the incoming melband range
• dk.triggerlength~ only updates the length of the triggers at the start of the trigger
• dk.sampler~ can now take messages directly from dk.plotter for corpus playback
• exposed control parameters for UMAP for the objects that use dimensionality reduction (dk.controllermatch, dk.gridmatch, dk.plotter)
• the read message now allows you to load a file directly from live.drop to dk.corpuscreate and dk.concatcreate
• dk.corpussampler~ now takes input from dk.plotter directly for navigating corpora
• improved and fixed bugs in how values of 0. to 1. are parsed by sample/class selecting objects (improving interoperability)
• changed how sample rate is reported within the objects to fix inconsistency issues when working inside Ableton Live
• changed contact mic EQ filter used in @input modes across all onset detection objects (dk.onset~, dk.spectralonset~, etc...) so it also works with contact mics on cymbals
• dk.speed can now take signal input as well as bangs/lists
• improved gate duration accuracy in dk.triggerlength~
• improved how attack and hold are computed in dk.triggershape~
• improved how the rate parameter in dk.sequenceclock~ works, allowing for faster changes when at slower tempos
• renamed num_files to num_samples in dk.corpuscreate to make more sense with segment-based corpus analysis
• added mic correction example (if you have HIRT installed) to the helpfile of dk.convolver~
• fixed bug in dk.clustercreate where clear-ing was not working properly
• fixed bug in dk.mfccrt~ that was causing it to output all of the time
• fixed a bug in dk.sampler~ which meant that refer'd buffer~s weren't triggering properly
• fixed bugs in the zoom function of dk.gridscale
• fixed bug when analyzing a single long file in dk.corpuscreate where the amount of slices were capped to 256. if you have used this with long files, try again
• fixed bug where dk.shattercreate~ was not working properly
• fixed bug(s) in dk.triggeralign~ that stopped working under certain circumstances (heavy CPU load primarily) and improved timing accuracy overall
• fixed bugs in pitch analysis when working with sample rates that are multiples of 48k
• slightly improved the audio quality of some of the built-in musical examples (declicking and denoising a few)
• removed Snippets from package
• added "collections" in the sidebar inspector, for UI objects, and all the patches that were once snippets

• BREAKING CHANGE - objects with "train" in the name (sp.classtrain, sp.clustertrain, sp.setuptrain~) have been renamed to be more consistent with other objects (sp.classcreate, sp.clustercreate, sp.setupcreate~)
• BREAKING CHANGE - sensitivity parameter in all onset-based objects has been fine-tuned to work across a wider range of instruments and materials so your previous settings may likely need adjusting
• BREAKING CHANGE - threshold parameter renamed to floor in all onset-based objects (sp.onset~, sp.onsetframe, sp.descriptors~, sp.melbands~, sp.mfcc~, sp.sines~, etc...)
• BREAKING CHANGE - threshold parameter renamed to sensitivity in sp.novelty~ keeping it inline with other onset detection objects macro controls (sp.onset~, sp.spectralonset~)
• BREAKING CHANGE - mode parameter renamed to timestretchmode in sample playback objects keeping to remove clash with the new mode attribute
• BREAKING CHANGE - input mode order and modes changed on all onset-based object (sp.onsetframe, sp.descriptors~, sp.melbands~, sp.mfcc~, sp.sines~, etc...)
• BREAKING CHANGE - clusters trained with sp.clustertrain now count from 1 instead of 0 and are returned as ints instead of a symbols (so a 1 instead of a "1")
• BREAKING CHANGE - sp.classifierdisplay @drum attribute reworked to @defaultdisplay
• BREAKING CHANGE - fade in/out and curve parameters renamed in sp.corpusplayer~ to be consistent with the new sp.sampler~ (attack=in, hold=out(and inverted), attackcurve=incurve, holdcurve=outcurve)
• overview patch (SP-Tools Overview.maxpat) completely revamped with better discoverability and package navigation
• added snippets for common use cases to SP-Tools package (accessible via the snippets side panel)
• restructured Max Package folders to make it easier to navigate and add additional corpora
• restructured github repo to separate versions (Max, pd), Max for Live devices, and additional corpora
• added class-combined versions of all descriptor analysis objects making it easier to get classification combined with descriptor analysis (sp.classdescriptors~, sp.classmfccs~, etc...)
• added sp.controllermatch to browse corpora using a single parameter
• added sp.controllerptich to create pitch-quantized melodic phrases from a single parameter
• added zero-latency convolution (sp.convolver~) which allows for reverb, amp sim, mic correction, and any other convolution-based audio processes
• added convolution-based corpus matching via sp.corpusconvolver~ which combines corpus matching and realtime convolution
• added combined abstractions for common processes (sp.corpusanalysis~, sp.corpussampler~)
• substantially improved core concatenation algorithm
• added radius and neighbor selection in concat matching
• added pitch and loudness compensation to concat playback
• added transpose and pretranspose to concat playback
• added sp.corpusclustermatch for creating clusters from corpora, allowing for grouping of similar sounds within a corpus (e.g. kick, snare, hat, etc...)
• added objects for shaping and processing descriptor lists (sp.descriptorcurves, sp.descriptorreplace)
• added an envelope follower with some quirky features (sp.envelope~)
• added sp.gridscale for scaling and shifting XY-based controller inputs (combines well with sp.gridmatch and the new sp.boppad)
• added pitch and time quantization with sp.quantizepitch and sp.quantizetime~
• added data and clock sequencers with sp.sequenceclock~ and sp.sequencedata~
• added Schmitt filter for descriptor filtering for when you need hysterisis in the filtering (sp.schmitt)
• added various audio processes for splitting and decomposing signals (sp.scatter~, sp.scramble~, sp.shatter~, sp.shattercreate~, sp.sift~, sp.smear)
• added spectral-difference onset detection (sp.spectralonset~)
• added new kitchensink descriptor type, as a combination of descriptor/melband/MFCC buffers which comes in onset, real-time, and class-based variants (sp.kitchensink~, sp.kitchensinkrt~, sp.classkitchensink~)
• added neural network classifier type to sp.classtrain, sp.clustertrain, and sp.classmatch for greatly improved speed and accuracy
• added support for the BopPad controller (sp.boppad)
• added input modes to all the physical modeling objects (sp.karplus~, sp.resonators~, sp.shaker~, sp.waveguidemesh~) allowing for realtime actuation of the models, in addition to onset/trigger-based actuation
• added ability to reverse and loop samples, and jump directly to sample position in sp.corpusplayer~
• added @mode attribute to sp.onsetframe~ to switch between amplitude and spectral onset detection (also impacts all onset-based descriptors objects (sp.descriptors~, sp.melbands~, sp.mfcc~, sp.sines~, etc...))
• added ability to seed clusters in sp.clustertrain
• added utility for subdividing a spaces of arbitrary size and steps (sp.slicecurve)
• added utility for time-alining trigger/gates with descriptor output (sp.triggeralign~)
• added utility for increasing the length of triggers and gates (sp.triggerlength~)
• added AHR-style envelope with sp.triggershape~
• improved time-alignment for trigger/gate outputs of descriptors objects (sp.descriptors~, sp.melbands~, sp.mfcc~, sp.sines~, etc...)
• reworked sp.classifierdisplay to automatically load class names from all class matching objects
• reworked sp.descriptordisplay to automatically display the correct type of descriptor
• added dump output message to all class-based objects (sp.classmatch, sp.classdescriptors, etc...)
• added mic correction (via a new input mode attribute @input 5) to all onset-based descriptor analysis objects for better descriptor results when using only an SP sensor
• added @buffer attribute to sp.corpusplayer~ to simplify non corpus-based sample playback
• added @history attributes to all realtime descriptor objects (sp.descriptorsrt~, sp.melbandsrt~, sp.mfccrt~, sp.sinesrt~) to smooth descriptors
• added @input modes to sp.onset~ allowing for optional filters optimized for different input types
• added onthreshold and offthreshold to sp.novelty~ given you finer grain control of thresholding
• improved (and fixed) melband transposition in sp.datatranspose
• added optional arguments to all corpus matching abstractions to load corpus and setups automatically (e.g. sp.corpusmatch corpus_china setup_snare will automatically load the China Cymbal corpus and load/enable the snare setup)
• added ability to load default files to all .json loading objects (e.g. sp.resonators~ resonators_bell)
• added header to every .json file to differentiate them (you should reanalyze all your corpora, classes, and setups)
• when loading a setup into any corpus matching abstraction (sp.corpusmatch, sp.corpusconvolver~, sp.corpussampler~) it will automatically enable that setup
• improved default starting values in sp.controllers so it gives useful results while it autocalibrates
• added ability to trigger sp.triggerbounce~ and sp.triggercloud~ with bangs as well as trigger/gates
• added ability to trigger synthesis objects (sp.lpg~, sp.karplus~, sp.resonators~, sp.waveguidemesh~) with bangs as well as triggers/gates
• added 8 more abstractions to the puredata version
• changed FFT settings for medium and all sizes for corpus and setup analysis to have better pitch accuracy in the larger time scales. You should re-analyze all your own corpora and setups.
• you can now load mp3 files into sp.corpuscreate
• fixed M4L device loading bug(s)
• abstractions that only function as part of an existing abstraction (sp.lpgcore~, sp.folderloop, etc...) have had their sp. prefix replaced with widget and have had their help/reference pages removed

• added synthesis and physical modeling objects (sp.karplus~, sp.resonators~, sp.resonaroscreate~, sp.shaker~, sp.sinusoidplayer~, sp.sinusoids~, sp.waveguidemesh~, sp.lpg~, sp.lpgcore~)
• added Max for Live devices for some of the new processes (Resonators, Sinusoids, Waveguide Mesh)
• added new descriptor type (sines) and corresponding objects (sp.sineframe, sp.sines~, sp.sinesrt~) (you should reanalyze all your corpora)
• added utilities for filtering and creating clusters of triggers and gates (sp.probability~, sp.triggerbounce~, sp.triggercloud~, sp.triggerframe~)
• added absolute start and length parameters to sp.corpusplayer~ (via startabsolute and lengthabsolute messages)
• added scramble transformation to sp.databending
• added slew parameter to sp.speed
• added ability to loop floats and ints (as well as conventional descriptors) to sp.datalooper~
• added proper Max documentation via reference files, autocomplete, etc...
• added sample rate adaptation to all realtime and offline analyses. previously things were optimized and assumed for 44.1k/48k but now everything works at every sample rate (up to 192k)
• added some puredata abstractions and help files to the package (in the puredata folder)

• BREAKING CHANGES - all objects that had a separate control inlet, now take those messages in the left-most inlet
• added new "ramp" objects for structural and gestural changes (sp.ramp, sp.ramp~)
• added new "data" objects for transforming, looping, and delaying descriptors (sp.databending, sp.datadelay, sp.datagranular, sp.datalooper~, sp.datatranspose)
• added novelty-based segmentation for determining changes in material type (sp.novelty~)
• added timestretching functionality to sp.corpusplayer~ and the Corpus Match M4L device

• added Max for Live devices (16 total) which cover (nearly) all the functionality of SP-Tools
• Max codebase further commented and tidied

• added Max for Live devices for some of the main/flagship functionality (Concat Match, Controllers, Corpus Match, Descriptors, Speed)
• added sp.gridmatch abstraction for generic controller-based navigation of corpora
• added support for the Erae Touch controller (sp.eraetouch)
• improved path stability when loading example corpora

• added "concat" objects for real-time mosaicking and concatenative synthesis (sp.concatanalysis~, sp.concatcreate, sp.concatmatch, sp.concatplayer~, sp.concatsynth~)
• added ability to apply filtering to any descriptor list (via sp.filter)
• improved filtering to allow for multiple chained criteria (using and and or joiners)
• updated/improved pitch and loudness analysis algorithms slightly (you should reanalyze corpora/setups/etc...)

• added ability to filter corpora by descriptors (baked into sp.corpusmatch via filter messages)
• added improved/unified corpus playback with sp.corpusplayer~
• add realtime analysis abstractions (sp.realtimeframe~, sp.descriptorsrt~, sp.melbandsrt~, sp.mfccrt~)
• added new stereo corpus (corpus_plumbutter.json)
• improved corpus analysis to work with stereo files and files shorter than 100ms as well as adding more comprehensive metadata
• added sp.corpuslist abstraction for visualizing and playing samples in a corpus in list form
• removed old playback abstractions (sp.corpussimpleplayer~, sp.corpusloudnessplayer~, sp.corpusmelbandplayer~)

• added "setups" (corpus scaling and neural network prediction/regression)
• added "controllers" (meta-parameters extracted from onset timings and descriptor analysis)
• added four new abstractions (sp.controllers, sp.speed, sp.setupanalysis, sp.setuptrain~)
• added new corpus (corpus_voice.json)
• added @roundrobin mode to sp.corpusmatch