Noise/minimum mic level?

[ankitson]

I often want to dictate over long periods of time with my headphones on, and there's sometimes TV or music or other transient speech in the background. The mic picks all of those up and FluidVoice transcribes it, so transcription only reliably works in silent settings.

I think if we could specify a minimum volume floor/silence threshold for FluidVoice manually, we could make it work in noisier settings too. I have done some work with local transcription / diarization pipelines on my homeserver, my guess is we can expose the VAD threshold to the user without too much difficulty? What do you think - i can take a crack at a PR.

[altic-dev]

I think it has to be auto adjusted to make it seamless but should be doable. Not sure if I will be happy with something that 'works for you' type PR tho. You can def take a crack but I've warned you ahah. Because this comes adds extra config for general users which on paper should be handled by me to dynamically handle it - perhaps?

[ankitson]

lol I understand it will have to be good code, well thought out feature to merge here, not just a hack :) I didn't understand this part "should be handled by me to dynamically handle it - perhaps?"