search

usefulsensors / openai-whisper

Robust Speech Recognition via Large-Scale Weak Supervision
MIT License
62 stars 24 forks source link

30 seconds limit #43

Closed narakai closed 1 year ago

narakai commented 1 year ago

Thanks for the great work. It seems that the record time limit is 30 seconds. Is it possible to set a longer time?

nyadla-sys commented 1 year ago

Segmenting the audio into 30-second blocks and performing inference on each block is both feasible and it needs to be worked on ...

narakai commented 1 year ago

Thanks for the reply, just want to process a .wav file which > 30 seconds instead of recoding a voice, will try to use ffmpeg to segment the file first.