# maum-ai/voicefilter

Unofficial PyTorch implementation of Google AI's VoiceFilter system

Repository: https://github.com/maum-ai/voicefilter
Canonical: https://ross.abutalabs.com/products/voicefilter
Homepage: http://swpark.me/voicefilter
Language: Python
License Family: other
Topics: source-separation, audio-separation, speech-separation, pytorch, voicefilter
Last push: 2024-07-25T10:15:31+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2721, "days_push": 769, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1216, forks 240 (observed 2026-08-28T04:04:01.255392+00:00)

## What it is
An unofficial PyTorch implementation of Google AI's VoiceFilter system, which isolates a target speaker's voice from noisy mixed audio given a reference audio enrollment of that speaker. It was an early open-source implementation of the paper but is no longer maintained.

## Use cases
- separate a target speaker's voice from noisy audio
- implement VoiceFilter target voice separation in PyTorch
- reduce word error rate of speech recognition by isolating a speaker
- study speech separation model architectures
- extract one speaker's audio given a d-vector reference

## When to choose
- you want a simple reference implementation of the original VoiceFilter paper in PyTorch
- you are studying speech separation and want a minimal codebase to learn from

## When to avoid
- you need a maintained, reliable production implementation - the author explicitly warns the code has known mistakes
- you want VoiceFilter-Lite or newer, better-maintained implementations
- you need support or reviewed issues/PRs - the project is unmaintained

## Facets
- artifact type: library
- maturity: abandoned
- function: audio-processing, machine-learning, deep-learning
- domain: speech-processing, machine-learning
- platform: python
- tags: voicefilter, speech-separation, source-separation, pytorch, unofficial-implementation, target-speaker-extraction, audio

## Member repositories
- maum-ai/voicefilter (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:01.255392+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 2026-08-30T06:17:11.826475+00:00, confidence not recorded.
  - readme: https://github.com/maum-ai/voicefilter (fetched 2026-08-28T04:04:01.255392+00:00, sha 8b200594033b)
  - homepage: http://swpark.me/voicefilter (fetched 2026-08-29T12:25:10.997398+00:00, sha 902a7b43693d)
- Data as of 2026-08-30T08:39:29.467469+00:00.