Page 1 of 1

Transcripts?

Posted: Sun Feb 05, 2017 12:50 pm
by JohnMayo
A few years ago a company offered to do a free transcript for the podcast. I jumped at the chance and that resulted in this page: http://comicbookpage.com/PodcastContent ... script.php

At a cost of $1/minute, I opted not to pay for any transcripts even though the company did an excellent job. I simply don't have the budget for $1/min transcripts

I've been playing around a little lately with computer generated transcripts and figured I'd ask the question of if you all saw any value in them or not.

Obviously, computer generated transcripts would be far from perfect but they'd be better than nothing. (Or would they?)

As a proof of concept, I've put together a really rough page with a computer generated transcript of Weekly Comics Spotlight #492: http://comicbookpage.com/PodcastContent ... script.php

and Weekly Comics Spotlight #493: http://comicbookpage.com/PodcastContent ... script.php

I'd appreciate feedback on if this is worth continuing with or not.

Re: Transcripts?

Posted: Sun Feb 05, 2017 7:13 pm
by drew
its close - like 80 percent - obviously would need a manual edit but it aint bad - what program is it-dragon naturally?

Re: Transcripts?

Posted: Sun Feb 05, 2017 9:26 pm
by JohnMayo
drew wrote:its close - like 80 percent - obviously would need a manual edit but it aint bad - what program is it-dragon naturally?
No, it isn't Dragon Naturally Speaking. It is a different approach which I've only slightly automated. The workflow involves a number of steps to go from an MP3 file to the transcript which I am experimenting with automating to the point of usability.

Ideally, I'd like to avoid the need for a manual edit as that takes time I don't exactly have to spend on this.

Re: Transcripts?

Posted: Mon Feb 06, 2017 7:49 am
by JohnMayo
I've added a computer generated transcript of Weekly Comics Spotlight #494: http://comicbookpage.com/PodcastContent ... script.php

That is probably as far as I'll go with the Weekly Comics Spotlight until I get more feedback indicating it is worth the time it takes to produce the webpages.

Re: Transcripts?

Posted: Mon Feb 06, 2017 1:31 pm
by fudd71
Interesting, I guess the real questions is what is the purpose? Is there an audience of deaf fans you hope to reach? Do you want the transcripts for historical archival purposes? To produce a book? How would that be different that the already existing archive of your articles? I don't think a large number of regular listeners would switch to reading transcripts, but you never know. Personally I don't think I would use the transcripts.

Re: Transcripts?

Posted: Mon Feb 06, 2017 1:45 pm
by JohnMayo
Personally, I see a few uses for the transcripts (in no particular order):
1) easier to search/find things in the backlist of episodes
2) improved searchability of the podcast/website
3) allows for others to pull quotes easier
4) content could be leveraged for books/ebooks

Re: Transcripts?

Posted: Sun Feb 26, 2017 5:49 pm
by JohnMayo
I've been playing around with the IBM Watson Speech to Text service. I've pumped almost all of my episodes through the system to mixed results.

Based on the ones I've checked, human editing is needed before the transcripts could be considered generally usable.

I've made a copy of the Episode Archive page with links to the auto-generated transcripts:
http://comicbookpage.com/PodcastContent ... cripts.php

Also, another benefit of transcripts is that it would open up the contents of the podcast to the deaf.