Manual diarization help
For generating ground truth for speech diarization jobs ... or maybe also for subbing idk
-
Press run. Use the 3 seconds of lead time given to sync the audio file closely as possible.
-
Use number keys (both the ones above the keyboard and numpad are fine) to indicate when which speaker talks.
-
Export as a txt file
-
Use a text editor to modify the file for use. e.g. replacing the default "Speaker 1" "Speaker 2", etc, replacing the default time span syntax, adding subtitle text, etc.
License: MIT
Future of the work: I will not be developing the tool further. Feel free to fork it if you really really want to do something with it.