The Speech-to-Speech translation works accurately and in real-time.
However, it has one limitation.
I am not sure how BigBlueButton is converting Speech-to-text but it misses punctuation marks. The speech-to-text converted by BBB has no punctuation marks. For example, it won't have question marks or escalation marks.
So the translated speech will lack the emotions of the speaker.
Next Step
The solution is to not rely on BBB's speech-to-text translation but to extract speaker's audio and convert it into speech using Google's speech-to-text that supports punctuation marks.
Hence, I am looking for a way to extract speaker's audio in real-time during a BigBlueButton session.
Any help would be greatly appreciated!
- Manish