codekingpro
/

audio-reborn

stem-separation

audio-transformation

Model card Files Files and versions

🔱 Audio Reborn

A professional GPU-accelerated audio transformation pipeline designed to bypass Content ID fingerprints using advanced AI synthesis.

🏗️ Project Architecture (GPU-Only)

Stage 1: Deep Audio Analysis (`1_Analysis`)

Tools: Librosa (GPU-backed via CuPy/Torch), Crepe.
Goal: Extract BPM, Key, Pitch, and Spectral Centroid.

Stage 2: AI Stem Separation (`2_Stem_Separation`)

Tools: Demucs v4 (HTDemucs) - CUDA Enabled.
Goal: Separate Vocals, Drums, Bass, and Other.

Stage 3: Multi-Layer Transformation (`3_Transformation`)

Tools: Pedalboard (VST3/GPU), Audiomentations (Torch-based).
Goal: Phase shifting, Micro-latency, Random EQ, Reverb.

Stage 4: AI Identity Regeneration (`4_Regeneration`)

Tools: RVC v2 (CUDA), Meta AudioCraft (MusicGen/AudioGen).
Goal: Complete Timbre replacement and Melody-conditioned synthesis.

Stage 5: AI Mastering & Quality Check (`5_Mastering`)

Tools: Matchering, Pyloudnorm, VISQOL.
Goal: Professional loudness and artifact verification.

🛠️ Infrastructure Requirements

OS: Windows (PowerShell)
GPU: NVIDIA (CUDA Toolkit 11.8/12.1)
Python: 3.10+ (Inside venv)

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support