barelyMusician active

Engine Sound Generator is a procedural audio synthesis tool that generates physically-modeled engine sounds using the Web Audio API. It implements waveguide-based synthesis techniques derived from research on physically informed car engine sound synthesis, simulating intake, exhaust, and engine block vibrations with configurable parameters like cylinder count, waveguide lengths, and reflection factors. The tool offers multiple implementations including a JavaScript AudioWorklet version and a WebAssembly-compiled version for improved performance. It integrates with Three.js for spatial audio positioning and includes Doppler effect simulation using DelayNodes. Parameters like RPM, muffler configurations, and individual component volumes can be adjusted in real-time. The WebAssembly version eliminates audio glitches present in pure JavaScript implementations while maintaining the same API surface. All synthesis is done procedurally without sample playback, making it suitable for dynamic vehicle simulations where engine characteristics need to change based on gameplay state.

Procedural Music Generator

Unity · Standalone

Generates melodies procedurally using Perlin noise mapped to musical scales. The noise curve is controlled by seed, octaves, lacunarity, and persistence parameters to produce deterministic, infinitely variable melodies suited for adaptive game music. Same seed always produces the same melody, allowing for reproducible results. The tool quantizes smooth Perlin noise curves to chosen musical scales (major, minor, pentatonic, blues, chromatic) with configurable BPM, note range, and length. Supports rhythm and rests via a separate Perlin noise track. Includes a web-based demo with real-time playback (MIDI export is listed on the roadmap, not yet implemented). Note: This project has evolved into SeedSong, a more complete system with multi-instrument support and genre presets. The original repository was archived in 2026 and is maintained as reference for the Perlin noise melody generation approach.

Pink Trombone

Standalone · Engine-agnostic

A programmable version of Neil Thapen's Pink Trombone that models the human vocal tract using physical simulation. The tool exposes audio parameters for glottal intensity, frequency, tenseness, and vocal tract shape, allowing direct control over tongue position and diameter. Implemented as a Web Audio API worklet processor with an optional interactive visualization. Developers can create, manipulate, and remove vocal tract constrictions in real-time to produce specific phonemes or arbitrary vocal sounds. The system handles both voiced and voiceless sounds through tenseness and loudness parameters. Includes presets for common phonemes (fricatives, stops, nasals, vowels) with precise index and diameter values. Built for web-based applications requiring real-time procedural speech synthesis without relying on recorded samples or TTS engines.

Raveler

Wwise

Raveler is a Wwise plugin that runs RAVE (Realtime Audio Variational autoEncoder) models for real-time timbre transfer via neural audio synthesis in game audio contexts. The plugin provides direct integration of trained RAVE models into Wwise effect chains, enabling neural processing of game audio with adjustable latent space manipulation. The plugin exposes controls for model performance parameters including latent noise injection, prior sampling, and dry/wet mixing. It offers direct manipulation of up to 8 latent dimensions with bias and scaling controls, all of which can be bound to RTPCs for dynamic runtime control. Buffer settings allow balancing between audio quality and latency based on project requirements. Based on the RAVE VST project, Raveler brings research-grade neural audio synthesis techniques into production game audio workflows through Wwise's standard plugin architecture. Note: the core is released under CC BY-NC 4.0 (non-commercial), which restricts use in commercial products.