Isolation Script - Final step at creating the dataset.
This will crop and trim the targeted speaker down to 30 second clips, then trim silence from them, so they're at most 30.00 seconds or less.
This will also convert them for use in audio training, like mono audio, 16-bit, etc.
Pre-Processing Script
This script will assist you by ripping the ENG audio track from a folder, and save the audio-only WAV files to a different directory, so you can start the diarization process.
Rename identify-single-speaker2.py to identify-the-speaker.py
Main Script
After you process and diarize the audio into the two folders, If it's a movie and one speaker, then you're done and this script was mostly useless. If it's a TV series and you're identifying the same speaker in 3000+ episodes, then this script is your lord and savior ;)
It processes the isolation speakers for you by playing a small section it collects from the JSON files, and then asks you if what it played was the speaker you want to identify.
Once you identify the speaker as the one you want, it isolates and exports it to a CSV file, where you can then process it using the rest of the package scripts.
