christian-byrne/audio-separation-nodes-comfyui
Separate stems (vocals, bass, drums, other) from audio. Recombine, tempo match, slice/crop audio
What it solves
This project provides a set of custom nodes for ComfyUI that allow users to manipulate and separate audio tracks. It solves the problem of isolating specific sound elements (like vocals or drums) from a mixed audio file or video, as well as performing basic audio editing tasks like tempo matching and cropping within a node-based workflow.
How it works
The toolkit integrates Hybrid Demucs to split audio into four distinct stems: bass, drums, vocals, and "other" (which includes instruments like guitars and keyboards). It also includes utility nodes for combining waveforms, shifting speed, matching the tempo of two tracks via onset detection, and replacing audio tracks within video files.
Who it’s for
It is designed for ComfyUI users who want to incorporate audio processing, remixing, and vocal isolation into their generative AI workflows, such as replacing background music in videos with AI-generated audio.
Highlights
- AI-powered separation: Uses Hybrid Demucs to isolate vocals, drums, bass, and other instruments.
- Tempo synchronization: Includes tools to detect BPM and time-stretch audio to match tempos between tracks.
- Video integration: Ability to replace the audio track of a video file with a new audio source.
- Waveform manipulation: Nodes for cropping audio and combining tracks using mathematical operations (add, mean, subtract, multiply, divide).
Related
- Project
- Project
- Project
- Project