An Open-Source Project to Unify Audio Processing and Generation

🩺🎧 Fix all your podcast, video or live stream audio! 🎧🩺

Mp3 to wav super resolution model for audio restoration & enhancement. U-Net + Discrete Wavelet Transform (DWT) Architecture

bin2bin, a Time-Frequency Generative Adversarial based method for Audio Packet Loss Concealment

A research based project which uses steganography and ML/deep learning algorithm to reconstruct the lost audio signals from a corrupted file.

🔬 AI-powered audio restoration platform implementing breakthrough research in blind enhancement. Transform any corrupted audio - speech, music, recordings - into professional quality through Self-Organized Neural Networks and advanced source separation. The future of audio recovery.

Short-Time Spectral Attenuator, plus attempt to reverse the process in order to have a noise-based music effect

A high-performance C++ port of HRAudioWizard - an audio enhancement tool that uses high-frequency compensation (HFC) to improve the quality of low-fidelity recordings by synthesizing plausible high-frequency content.

McGill University - Digital Audio Signal Processing

🎧 Enhance MP3 audio quality by removing compression artifacts with a wavelet-based deep learning model for superior sound restoration.

Code and demo site for complex-bin2bin speech signals inpainting

A studio-grade, fully automated VHS audio restoration suite. Uses a hybrid AI engine to separate and clean vocals while preserving background music and singing voices. Includes RTX 5090 fallback support.

A high-performance local AI pipeline for restoring VHS audio, transcribing with Whisper, and translating subtitles using NLLB-200.

It is a professional-grade digital audio workstation designed for precision waveform editing, complex multitrack mixing, and industry-leading audio restoration for film, broadcast, and podcasting