ruvnet

ruvnet / musica

Public

Structure-first audio separation via dynamic mincut — spectral graph methods in Rust

10
3
100% credibility
Found Apr 11, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

Musica is a zero-dependency Rust library for ultra-low-latency audio source separation using graph partitioning, ideal for hearing aids, embedded devices, and real-time music stem splitting.

How It Works

1
🎧 Discover Musica

You hear about a clever tool that magically pulls apart mixed sounds like voices from noise or instruments from songs, perfect for better hearing or fun music experiments.

2
📁 Pick your audio

Grab a recording from your phone or computer, like a noisy conversation or favorite track, and load it into the app.

3
🔧 Choose sounds to separate

Tell it what to find, like speech from background chatter or drums from music, using simple buttons.

4
✨ It separates instantly

Watch as the tool swiftly untangles the sounds into clear separate streams, feeling the magic of clean audio emerge in seconds.

5
🎵 Listen and tweak

Play back the separated parts, adjust volumes if needed, and hear how crisp everything sounds now.

✅ Enjoy your clear audio

Save or use your enhanced tracks for hearing aids, music remixing, or sharing noise-free recordings with loved ones.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is musica?

Musica separates mixed audio into clean sources—like speech from noise or music stems like vocals, bass, and drums—in real-time with sub-millisecond latency. Built in Rust, it runs on embedded devices, MCUs, hearing aids, or browsers via WASM, handling streaming binaural audio or 6-stem music separation without any ML models or training data. Developers get WAV I/O, benchmarks, and APIs for quick integration into audio apps.

Why is it gaining traction?

It crushes neural separators on speed (0.2ms vs 10ms+) and size (0 bytes model), making it viable for edge where others fail, plus full interpretability via graph cuts for debugging. No dependencies beyond one mincut crate, 276 passing tests, and scripts for WASM checks or test audio downloads draw devs seeking reliable DSP without black boxes. For music apk github or reproductor de musica github projects, the multitrack mode delivers stems instantly.

Who should use this?

Embedded engineers building hearing aids or smart mics need its streaming enhancer with audiogram support. Browser audio devs want WASM for live separation in telecon apps or youtube musica github tools. Arduino hackers or MCU firmware writers targeting audio—like musica arduino github or musical berlin event recorders—get low-latency crowd tracking for thousands of speakers.

Verdict

Promising for real-time audio separation on constrained hardware, but at 10 stars and 1.0% credibility, it's early—docs shine with benchmarks, but real-audio evals lag. Try for prototypes if latency trumps SOTA quality (1-5dB SDR).

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.