Skip to content

Faster whisper

Requirements#

Requires the audio-dataset-converter-faster-whisper library.

Plugins#

Transcribing audio#

The following commands loads raw audio files (i.e., ones without a transcript) and applies the fw-transcribe filter to generate a transcript using faster-whisper, with the result then being stored in Festvox format:

adc-convert \
  -l INFO \
    from-data \
      -l INFO \
      -i "./raw/*.wav" \
      -t sp \
    fw-transcribe \
      -l INFO \
    to-festvox-sp \
      -l INFO \
      -o ./festvox

Tools#

Generate SRT subtitles#

The adc-srt allows generating subtitles in SRT format from audio and video files.

The following example generates subtitle files for all .mp4 file, alongside the video files:

adc-srt \
  -l INFO \
  -i ./input/*.mp4

In this example, the .srt files generated from .wav files get placed in a separate output directory:

adc-srt \
  -l INFO \
  -i ./input/*.wav \
  -o ./output