search

linto-ai / whisper-timestamped

Multilingual Automatic Speech Recognition with word-level timestamps and confidence
GNU Affero General Public License v3.0
1.87k stars 150 forks source link

simple split audio file example using whisper-timestamped #111

Closed silvacarl2 closed 10 months ago

silvacarl2 commented 1 year ago

do you have a simple split audio file example using whisper-timestamped floating around?

8-)

Jeronymous commented 1 year ago

What do you mean (by "simple split audio file")?

silvacarl2 commented 1 year ago

i apologzie i was not more clear. given the output from whisper-timetamped, are you aware of a simple tool that can read its json output once it has processed an audio file in MP3 and then produce MP3 output that splits the original MP3 by word into individual MP3 clips that contain each word?

this could/would be extremely helpful for batch processing a large amount of audio data into indivial clips.

Jeronymous commented 10 months ago

Interesting.

But that's another project, that should be more generic, just not linked to the notion of words. Something that takes an audio, timestamp intervals (in usual format) and split the audio.