Subtitles/CC turned off solution?

[parth31533]

Hey man, I was working on this, and had a similar approach like yours using the YT APIs, but in my case the videos are from oldser years and there are 7hrs long, added to that most of the older videos from 2019 have their closed captions turned off from the creator. i have thought of using AssemblyAI approach to downthe audio file and then run that to assemblyAI but that approach takes a lot of time in my case.

do you have these issue if the subtitles are turned off ?

my rep: https://github.com/parth31533/YT-Project/blob/main/Josh.ipynb?short_path=69eb1b4

Br,
Parth

[therohitdas]

Hi @parth31533

Transcript being turned off is a huge issue. Why don't you extract audio using ffmpeg and then use whispher/other model to extract text? Speech-to-text models have come a long way.

If I had to speed things up, I'd cut the audio into X chunks and then run the API request to convert speech to text parallelly.

Sorry for the late reply; I have notifications turned off.