Made to run commonvoice data through codec2 via ffmpeg. Requries the following:
- ffmpeg
- libcodec2
- sqlite3
- libmp3lame
-
Generate entries in the database to track the progress of converting each file:
mix prepare "~/Projects/commonvoice_17_0_en_codec2/audio/en" -
Run the conversion, specifying the relative path:
mix garble "~/Projects/commonvoice_17_0_en_codec2"
Altogether:
mix prepare "~/Projects/commonvoice_17_0_en_codec2/audio/en"
mix garble "~/Projects/commonvoice_17_0_en_codec2"Unpacking tar files recursively:
find . -type f -name '*.tar' -exec sh -c 'tar -xvf "$1" -C "$(dirname "$1")"' _ {} \;