YouTube just released a new feature for video content generators: annotations. As you can see in this example video of someone jumping out of an aircraft the annotations show up whereever you want to put them on the screen and are time-based. You can even add links and hotspots to other videos and search results which means you can do interactive games using several videos.

Now this is all cute and nice, but what I’d want is API access to these annotations. This would allow us to provide not only captioning of the video for the hard of hearing but also information for blind visitors. I’ve written about this before, you can easily create an interface to have timed captioning on youtube but playing the captions back is trickier as you have no means of syncing the video (if the video buffers in between the captions and the video get out of sync).

Now, if YouTube came up with an API access to these captions that fires an event every time a new caption starts with the type of caption and its text value, it would be dead easy to update a hidden form field with that text (or an ARIA live region) to provide a poor man’s captioning and information for the hard of hearing.

YouTube could become both a larger consumer faced product by enabling more disabled visitors to gain access and a means of captioning video that is intuitive and easy to use.

I’d be happy to help out!

• check out youtube annotations

One thing that has been annoying me for ages is that no video player on the web allows you to write comments for a specific time in the video that get displayed as plain text. Viddler allows you to comment at a certain time and it appears in the video, but the benefits of time based captioning both in terms of accessibility and SEO didn’t quite transpire to any video site maintainers yet. Edit: Darn, I hadn’t looked at Viddler for a long time, it actually does this now, well done!

Google just released a JavaScript API for YouTube which makes it dead easy to control a video with JavaScript. You can start, stop and jump to a certain time of the video but more importantly - you have events firing when something happens to the player. This made it easy for me to whip up a proof of concept how time-based captioning might work as an interface. Click the screenshot to see it in action.

Start the video and hit the pause button to add a new caption. You can delete captions by hitting the x links and you can jump back to the section of the video by clicking the time stamp.

Check the source for how it is done. In order to make this a service, all you need to do is have a backend script that gets all the form fields and store it in a DB.