I record conferences and produce them in an HTML player. I have a need to transcribe some content and integrate it into the player.

I found this question, but the answer seems to say that the transcription is mixed into the video file. I do not want that.

What I need, or could work with, is a transcript that is in a separate text file with the timings that the sections begin. XML, CSV, or some other standard is perfect.

I have thousands of hours per year that I would potentially need to do this on. There needs to be very little effort involved to get the transcripts for this to be viable.

What would be perfect is that I tell the program what videos I want a transcript for, click go, then it produces the files. Any OS is fine, though I prefer Windows.

Update: I learned that Google employs a very good transcript engine, but you cannot seem to take advantage of it without actually uploading files to YouTube. I asked if there was a way around this on Superuser, but got nothing. To further complicate things, YouTube decides if the video with be transcripted without notice to the user. There is no option to select it or generate it on command. So in expansion of this question, what options do I have available to me regarding YouTube's engine? Is there a way around this? I can upload to YouTube as long as I can keep it private and the transcript will be generated at time of encoding, if that's the only option.

1 Answer 1


Adobe premiere has a speech analysis system built in, after cs4 I believe. Machine transcription systems tend to be very poor though - in my experience it is only useful for a vague sense of where you are in the video. See below:


Human transcription services are getting cheaper as many of them use Amazon's "Mechanical Turk" to provide the human labour. The quality is much better, but will still require editing for readability. See the link for more info: http://waxy.org/2008/09/audio_transcription_with_mechanical_turk/

  • Thanks for the link. I have had premiere on my PC for years and still haven't used it. I will give this a go and report back.
    – user3643
    Jul 27, 2013 at 15:10
  • With the Premiere transcription thing I always found that the time spent hunting through the transcript and checking for errors was greater than the time it would take to just write it yourself. Especially as the errors are sometimes not apparent unless you're reading it while listening to the voice, and even then a few always slip through. We have to transcribe everything we produce and we use an online transcription service. The price beats paying one of the office staff for the time it takes for them to do the transcription (and then for me to check the timings) by an order of magnitude.
    – stib
    Jun 4, 2015 at 3:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.