# prophesier/diff-svc

Singing Voice Conversion via diffusion model

Repository: https://github.com/prophesier/diff-svc
Canonical: https://ross.abutalabs.com/products/diff-svc
Language: Jupyter Notebook
License: AGPL-3.0
License Family: copyleft
Last push: 2026-06-06T09:49:46+00:00

## Health v2 (maintenance only)
Score: 62/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 86, release rhythm 8, longevity 100
- inputs: {"age_days": 1411, "days_push": 88, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2717, forks 812 (observed 2026-08-28T04:07:12.175242+00:00)

## What it is
Diff-SVC is a deep learning project that performs singing voice conversion using diffusion models, transforming input singing audio into a target singer's timbre with basic pitch correction support. It includes preprocessing, training, and inference pipelines for training custom voice conversion models on user datasets.

## Use cases
- convert a singing voice into another singer's timbre
- train a custom singing voice conversion model on my own dataset
- do real-time voice changing during singing
- apply pitch correction to converted vocals
- run voice conversion inference on a consumer GPU
- create AI cover songs with a target voice

## When to choose
- you need diffusion-based singing voice conversion with trainable custom timbres
- you want GPU inference on modest hardware like a GTX 1060
- you need both offline conversion and real-time voice-changing support

## When to avoid
- you need a production-ready, commercially licensed system (AGPL-3.0, academic-use disclaimer)
- you only need speech-to-speech conversion rather than singing
- you want a plug-and-play app with no training or Python setup

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, audio-processing, speech-recognition, llm-training
- domain: machine-learning, artificial-intelligence, deep-learning
- platform: python, cross-platform, windows
- tags: singing-voice-conversion, diffusion-model, voice-conversion, timbre-transfer, pitch-correction, vocoder, hubert, audio-synthesis, audio, gpu, linux

## Member repositories
- prophesier/diff-svc (main) score 62

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:12.175242+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:15:20.390752+00:00, confidence not recorded.
  - readme: https://github.com/prophesier/diff-svc (fetched 2026-08-28T04:07:12.175242+00:00, sha 7c1641131bdd)
- Data as of 2026-08-30T08:39:29.467469+00:00.