Skip to content

fishaudio/audio-preprocess

Repository files navigation

Fish Audio Preprocessor

PyPI Version

中文文档

This repo contains some scripts for audio processing. Main features include:

  • Video/audio to wav
  • Audio vocal separation
  • Automatic audio slicing
  • Audio loudness matching
  • Audio data statistics (supports determining audio length)
  • Audio resampling
  • Audio transcribe (.lab)
  • Audio transcribe via FunASR (use --model-type funasr to enable, detailed usage can be found at code)
  • Audio transcribe via WhisperX
  • Merge .lab files (example: fap merge-lab ./dataset list.txt "{PATH}|spkname|JP|{TEXT}")

([ ] indicates not completed, [x] indicates completed)

This code has been tested on Ubuntu 22.04 / 20.04 + Python 3.10. If you encounter problems on other versions, feedback is welcome.

Getting Started:

pip install -e .
fap --help

Reference