Divyansh Kulshrestha lineageon

AUDIO::DSP::SYSTEMS

REAL-TIME AUDIO ENGINEERING LOW-LATENCY DSP AI AUDIO MODELING VST3 PIPELINES

╔══════════════════════════════════════════════════════════════╗
║                 SYNTHESIS CONTROL TERMINAL                  ║
╠══════════════════════════════════════════════════════════════╣
║ ROLE        :: Audio Developer / DSP Engineer               ║
║ DOMAIN      :: Real-Time Systems + AI Audio Processing      ║
║ ARCHITECT   :: Low-Latency Runtime Design                   ║
║ SIGNAL PATH :: Hardware ↔ DSP ↔ Neural Inference            ║
║ STATUS      :: ACTIVE                                       ║
╚══════════════════════════════════════════════════════════════╝

SYSTEM::OVERVIEW

Building real-time audio systems where
traditional DSP meets neural inference.

Focused on:

Ultra low-latency audio architectures
Frequency-domain DSP engines
JUCE-powered VST pipelines
Hardware audio routing systems
Neural denoising integration inside C++ runtimes
AI-assisted spectral enhancement workflows

I engineer systems that behave more like hardware racks than software apps.

MASTER_PROJECT::Adaptive Audio Denoiser Engine

PHASE_01 :: CLASSICAL DSP
---------------------------------------------
Input Signal
    ↓
STFT / Frequency Domain Transform
    ↓
Spectral Subtraction Engine
    ↓
Dynamic Vocal Phase Isolation
    ↓
Noise Profile Attenuation
    ↓
Real-Time JUCE Buffer Processing
    ↓
Output Stream

PHASE_02 :: NEURAL DSP MIGRATION
------------------------------------------------
Current Direction:
DeepFilterNet-inspired lightweight inference
RNN-style temporal filtering structures
Real-time streaming neural enhancement
Optimized CPU-safe runtime deployment
C++ inference bridge experimentation

DSP::CONTROL_PANEL

`GAIN::C++ (JUCE Framework)`

[█████████████████░░░] 88%

`THRESHOLD::DSP Engine`

[██████████████████░] 90%

`DRY_WET::Python + ML`

[███████████████░░░░] 75%

ACTIVE::MODULES

C++ JUCE VST3 SDK

DSP FFT Spectral Math

PyTorch NumPy Python

Git VS Code FL Studio Figma

LOW_LEVEL::STACK

Audio Runtime Layer
├── C++17
├── JUCE Core
├── VST3 SDK
└── Real-Time Thread Management

Signal Processing Layer
├── Spectral Analysis
├── Frequency-Domain Processing
├── Dynamic Filtering
└── Buffer Synchronization

Neural Audio Layer
├── Python Prototyping
├── NumPy
├── PyTorch
└── Lightweight Inference Exploration

Creative Environment
├── FL Studio
├── VS Code
├── Git
└── Figma

CURRENT::FOCUS

> optimizing neural denoising for real-time playback
> reducing inference latency inside JUCE runtimes
> spectral enhancement + phase recovery experiments
> hybrid DSP + ML processing architectures
> hardware-aware audio routing systems

REALTIME::PIPELINE

MIC_INPUT
   │
   ▼
[ AUDIO DRIVER ]
   │
   ▼
[ JUCE CALLBACK ]
   │
   ▼
[ FFT WINDOW ]
   │
   ▼
[ SPECTRAL PROCESSING ]
   │
   ├── Noise Reduction
   ├── Phase Recovery
   ├── Harmonic Analysis
   │
   ▼
[ OPTIONAL ML INFERENCE ]
   │
   ▼
[ LOW LATENCY OUTPUT ]

ROUTING::CHANNELS

X

https://x.com/lineagebeats

MAIL

divyanshkulshrestha10524@gmail.com

Navigation Menu