elan-ev / vosk-cli

Apache License 2.0
2 stars 9 forks source link

Display avg confidence of transcription #15

Closed owi92 closed 1 year ago

owi92 commented 1 year ago

This extends the write_webvtt_captions function with a calculation of the average confidence coefficient of the transcription. Confidence will be displayed after a successful transcription.

Arnei commented 1 year ago

Code looks good to me and works, but leaves me wondering why you would make an effort to calculate this just to display it on the command line output. Can Opencast consume it this way?

lkiesow commented 1 year ago

Can Opencast consume it this way?

Opencast could consume stdout. An alternative option would be to introduce something like a --meta metadata.json command line argument and then write that JSON file. That would probably be more robust to parse. But then, I don't know if there are any more meta information we want to pass to opencast. So, not sure what's the better option.