tts-recorder
Records sentences/utterances for TTS training.
Requirements
- Python 3.8.10
- Python packages
- Librosa
- pysoundfile
- pynput
- Conda installable (using e.g.
conda install -c conda-forge portaudio)- ffmpeg
- portaudio
- pyaudio
Supported commands when it asks for your next sentence
devices- Lists all the devices for recording. These can be passed in as the first arg. E.g.
python -m main <device idx>.
- Lists all the devices for recording. These can be passed in as the first arg. E.g.
undo- Undoes progress and allows you to re-record the previous sentence.
assemble- Outputs metadata.csv in the output directory.