2

Is it possible to get an approximate duration of each word in an audio file? The closest thing (for audio files from youtube videos) is to download the captions file as an srt. The srt will then have the duration for each sentence in the video.

I was wondering if it is possible to somehow get the duration for each word in a sentence. Maybe not accurate but something around that ?

Karup
  • 2,024
  • 3
  • 22
  • 48
  • What's the connection to `python`? As for the question: words are delimited by periods of quiet. Search for those intervals with only the background/ambient noise level. Anything between those are words. – boardrider Jan 27 '16 at 09:37
  • Unless you have some extra information, it's a non-trivial task. There are some libraries which can provide you with detection or regions of speech vs. non-speech, aka speech activity detection, but that's only a starting point since it will not give you boundaries between words.\ – Itamar Katz Jan 27 '16 at 10:24
  • @boardrider: In a natural speech, there is no pause between most words. – Itamar Katz Jan 27 '16 at 10:25
  • @ItamarKatz can you please point out such libraries (that detect regions of speech). That might be a starting point as you said. – Karup Jan 27 '16 at 15:37
  • It depends on what language/platform you use. Look for "voice activity detection" on google as a starting point – Itamar Katz Jan 27 '16 at 15:51
  • Difficult to answer your question without knowing what format your audio file uses. – boardrider Jan 27 '16 at 19:48

0 Answers0