Tongue uses speech recognition to get the location of each individual word inside an audio or video file, then compiles these together in any order requested. This uses google speech recognition for the bulk processing, and then pocketsphinx to check individual words.
This also includes "autosub", a Python script used to handle the google speech recognition. The license is included and the original is available here.
To view the help for tongue use any of the following commands:
tongue help
tongue --help
tongue -hFirst have a folder containing mp3 or mp4 files, ideally only containing the voice you want to mimic.
Then run:
tongue add <filename>on each of them to generate a .srt file for each, containing the location of each word said. Note that this will take a while to run, often around 10 minutes per hour of audio.
To view a list of the words found, use:
tongue listTo make your subject say a phrase, use any of the following:
tongue say "The phrase you want them to say"
tongue say The phrase you want them to sayTo quickly see if all the words for a phrase are available without doing the full speech generation, use any of:
tongue check "The phrase you want to check"
tongue check The phrase you want to check