How to embed generated subtitles with timing? Like each spoken word is highlighted - any open source for this?

[FurkanGozukara]

What i mean is like in this video each word is highlighted and also embedded to the video

Do you know how can we do such thing?

https://twitter.com/GozukaraFurkan/status/1745466705592164835

Lets say I have generated subtitles with accurate time via Whisper

How can we embed them to the video?

[ryanheise]

For inspiration, take a look at #435 (reply in thread) which makes a video out of audio, but you may be able to adapt it to suit your purposes. That snippet was using a .ass file as input, but Whisper can now generate an SRT file with word highlighting like this:

whisper --output_format srt --word_timestamps True --highlight_words True YOURINPUTFILE

Once you have that SRT file, you can use a command like the one in the linked comment to try to overlay that subtitles file over a video, but use your SRT file instead of a .ass file.

[FurkanGozukara]

awesome looking right now

[FurkanGozukara]

i made .ass file like this but it is not working any ideas? i get no error.

I added .txt to upload here.

a.ass.txt

[ryanheise]

I'm saying you should use the SRT file instead of .ass because Whisper doesn't generate .ass files, it only generates SRT files. Here:

but use your SRT file instead of a .ass file.

And to clarify, by that, I mean use the SRT file generated using the command I gave you.

[FurkanGozukara]

ok i made it work :)

[FurkanGozukara]

For inspiration, take a look at #435 (reply in thread) which makes a video out of audio, but you may be able to adapt it to suit your purposes. That snippet was using a .ass file as input, but Whisper can now generate an SRT file with word highlighting like this:

whisper --output_format srt --word_timestamps True --highlight_words True YOURINPUTFILE

Once you have that SRT file, you can use a command like the one in the linked comment to try to overlay that subtitles file over a video, but use your SRT file instead of a .ass file.