Google I/O 2011: The YouTube Caption API, Speech Recognition, and WebVTT captions for HTML5

Naomi Black, Cynthia Boedihardjo, Jeffrey Posnick

Captions make video accessible and more importantly: discoverable to search. Learn how we're live captioning Google I/O with a new open source gadget, discover caption features in the YouTube Data API based on speech recognition, and find out more about WebVTT, a new format to help broadcast TV subtitling make the leap to HTML5 and the Web.

Category:

Science & Technology

Tags:

• YouTube
• accessibility
• translation
• speech recognition
• App Engine
• captions
• Google APIs
• webVTT
• HTML5
• naomi black
• cynthia boedihardjo
• jeffrey posnick
• broadcast captions
• naomi bilodeau

[GreenyLiveshow]

I have been hearing alot about HTML 5.

[farukba]

funny I have done this as part of my masters thesis and used the SAPI engine from microsoft to do exactly the same thing as google I/O for automated timings

it was work done for a company that needed to transcribe timings for large amounts of texts for language learning software

I achieved almost 95% accuracy if the sound quality was appropriate with english and french texts;

I have done this early march and also supported spelling correction errors to improve on bad recognition

[Abdullu]

@kindofquirky Thanks kindofquirky, didn't know that. It works with my desktop. ... Though when accessing Youtube via mobile web app or a smartphone native app, I still don't know how to make subtitles more readable.

[kindofquirky]

@Abdullu

you can add a background to the captions by pressing 'b' when captions are on. Also pressing '+' and '-' will increade/decrease the text size. The problem is that these key presses need making obvious to the user. There doesn't seem to be a way to find out all the keypresses for the player other than searching the web. Really there should be a 'help' link near the player that opens a page/popup describing how to use the player in full.

[mlpc47]

I installed new software, Omnipage 18 and it now stopped my Voice recognition from working. I uninstall the software and reinstalled Omnipage 17 and the Voice recognition will work again. In the past, there was not a problem. What has started this, as in the past there were no problem?

[Abdullu]

White captions on white background. Google, it's a long way to real accessibility. Quite ironic to see this gotcha throughout the Caption API presentation.

[MsPwain]

Is there an editor for webvvt, cause I would like to use it on my videos. Preferebly for Linux.

[MsPwain]

That is very nice to see, that Google are constantly trying to make the web more accessible. I like how Google is never sleeping (like some other company starting with the letter M). This is exactly *the* reason I like Google.

On a side note, .srt-format has been a standard for a long time for DVDs. This looks like an extention to that standard?

[siglmgga]

VERY good video - captioning at google - join the ccac - ccacaptioning at gmail dot com