search

mohdfareed / msd-p18538

Interactive Training Robot for Fire Safety
0 stars 0 forks source link

Add live transcription #13

Closed mohdfareed closed 10 months ago

mohdfareed commented 10 months ago

Added live transcripting through a microhpone to the terminal as a demo.

I used a free version of Google's API for testing, but if all other options are of comparable speed then this might become a problem.

I added a simple transcription algorithm for now that is too slow for what we want, but it has a lot of room for improvement.

Resolves #11

mohdfareed commented 10 months ago

There are alternatives that we can try with built-in live transcription like: https://www.assemblyai.com

mohdfareed commented 10 months ago

We can create a demo for our review of live transcription, what do you think? @BrianMonclus @umayo

BrianMonclus commented 10 months ago

@mohdfareed Sorry, for not replying sooner... Just a busy week on my end. I think that it would be great to have this. Although I think it will likely be for the phase 4 demonstration, as we have to get some other stuff done before then, ie. summary doc and presentation. If there are no other concerns I will likely merge this request and close the current milestone later today.

mohdfareed commented 10 months ago

I actually wouldn't merge it just yet, I found a couple of bugs I wanna fix first. Also let me know what you think of its current performance whenever you get the chance.

mohdfareed commented 10 months ago

I've fixed most the bugs in transcription and cleaned up the audio module. I think this is good to be merged at this point.