Can we auto-generate word-level timed transcripts from caption/subtitle SRT/WebVTT?

[nfreear]

A quick and dirty solution may involve dividing the time taken per phrase, by the number of words in the phrase...

(A more sophisticated solution may involve counting the number of syllables...)

[nfreear]

• Stackoverflow: questions/405161/detecting-syllables-in-a-word
• http://www.nltk.org/
• https://code.google.com/p/hyphenator/wiki/en_PublicAPI
• http://www.liedtke.it/phpHyphenator/example.php

Example:

• http://iet-embed-acct.open.ac.uk/dev/mediaelement-demo/media/ou-student-experiences-openings-captions.srt