Major Update Incoming — Tiny AutoFUS v2
We are moving beyond static audio enhancement.
The next version of Audio Optimizer evolves into a context-aware edge AI signal adaptation system powered by Tiny AutoFUS v2.
✨ What’s new
🧠 Environment Profiling The system now captures a short acoustic fingerprint and adapts its DSP strategy depending on environmental characteristics (room reflections, noise conditions, open vs enclosed spaces).
⏱ Temporal Adaptation Layer AI decisions are stabilized over time using smoothing and adaptive-rate control, turning the model into a real-time signal control loop rather than a simple inference block.
📊 Signal Health Monitor New internal metrics track clipping probability, noise floor drift, and dynamic range behavior — enabling AI-driven signal integrity management.
🔁 Model Hot-Swap Architecture The neural controller is now modular, allowing future TinyML models to plug into the DSP pipeline without architectural changes.
🎯 Direction
This update shifts the project from an “AI audio app” to a reference implementation of ultra-light edge AI for adaptive signal processing.
Still:
- Fully offline
- ~25 KB neural model
- Real-time operation
- Runs on standard Android hardware
More details coming soon.
Audio Optimizer is a lightweight Android app that applies real-time adaptive equalization to your device’s audio — all processed locally, with no internet or cloud required.
Unlike static EQs, it dynamically adjusts based on live audio content, enhancing clarity and balance whether you're listening to music, watching videos, or on a call.
📥 Download audio_optimizer_v1.0.apk
(Works on Android 8.0+)
🔒 100% offline — no data collection, no telemetry.
Audio Optimizer captures system audio directly from your device — not through the microphone.
✅ Compatible with:
- YouTube
- Spotify
- Netflix
- Games
- Phone calls
- Any app producing sound
How?
→ Uses Android’s built-in MediaProjection API (same as screen recording apps)
→ Requires one-time user approval (no root, no special permissions)
→ Supported on most Android 10+ devices
- Real-time adaptive equalization
- On-device neural processing (25 KB PyTorch Mobile model)
- Low-latency native DSP: Biquad filters, FFT, noise-aware gain
- Built-in WAV export for analysis
- Lightweight (~1.2 MB APK)
- Fully offline, privacy-first design
- Android Studio (with NDK & CMake)
- JDK 11+
- Android SDK ≥ API 24
- Clone the repository:
git clone [https://github.com/Kretski/audio-optimizer-android.git](https://github.com/Kretski/audio-optimizer-android.git) ## Screenshots
