We expect new-fave to be primarily used as a command-line tool. This page outlines that usage. Using new-fave this way does not require you to do any python programming, but if you would like to import new-fave into a python project of your own, see the page on Python Usage.
To use new-fave, you will need to have python installed on your computer. Currently, new-fave supports python versions 3.10, 3.11, or 3.12. If you are not sure whether python is installed, or what version is installed, here is a good tutorial for figuring that out.
In this documentation, when code is meant to be run at the command line, the code snippet will begin with # command-line
Once you have python successfully installed, you can install new-fave at the command-line like so:
After installing new-fave, the fave-extract executable will be made available. You can access a minimal help message by just running fave-extract with no arguments.
Usage: fave-extract [OPTIONS] COMMAND [ARGS]...
Run new fave-extract
Options:
--help Show this message and exit.
Commands:
audio-textgrid (audio-textgrid)
Run fave-extract on a single audio+textgrid pair.
corpus (corpus) Run fave-extract on a directory of audio+textgrid...
subcorpora (subcorpora) Run fave-extract on multiple subdirectories.fave-extract has three sub-commands. Which one you use will depend on how your data is organized.
In the simplest case of a single audio/textgrid pair, your best option is the audio-textgrid subcommand. For example, if you had the following files in a data/ directory:
data
├── speaker1.TextGrid
└── speaker1.wavTo use all default settings, you would run the following:
To customize the way fave-extract audio-textgrid works, including how to incorporate speaker demographics into the output, see the customization documentation.
If you have all of your audio file/textgrid pairs in a single directory, then the corpus subcommand is your best option. An example file organization would look like this:
my_corpus
├── speaker1.TextGrid
├── speaker1.wav
├── speaker2.TextGrid
└── speaker2.wavThe corpus subcommand will only work if the file names are the the same for the audio/textgrid pairs. That is, if your audio files are named something like speaker1.wav, and your textgrids are named something like speaker1_aligned.TextGrid, the corpus subcommand won’t process them.
To use all default settings, you would run the following:
To customize the way fave-extract corpus works, including how to incorporate speaker demographics into the output, see the customization documentation.
If each audio file/textgrid pair is in its own directory inside of a larger project directory, then the subcorpora subcommand is the best to use. An example file organization would look like this:
big_project
├── speaker1
│ ├── notes.txt
│ ├── speaker1.TextGrid
│ └── speaker1.wav
└── speaker2
├── notes.txt
├── speaker2.TextGrid
└── speaker2.wavThe corpus subcommand will only work if the file names are the the same for the audio/textgrid pairs. That is, if your audio files are named something like speaker1.wav, and your textgrids are named something like speaker1_aligned.TextGrid, the corpus subcommand won’t process them.
To use all default settings, you would run the following:
To customize the way fave-extract subcorpora works, including how to incorporate speaker demographics into the output, see the customization documentation.