Welcome to WebmasterWorld Guest from

Message Too Old, No Replies

Indexing the Audio Part of Video Files - in text!



4:36 am on Jul 16, 2008 (gmt 0)

WebmasterWorld Senior Member tedster is a WebmasterWorld Top Contributor of All Time 10+ Year Member

You may not have noticed this yet, but Google has thrown a new technology gadget into the US political arena. To quote CNet [news.cnet.com] "Videos from YouTube's Politicians channels are automatically transcribed from speech to text and indexed. Using the gadget, you can search not only the titles and descriptions of the videos, but also their spoken content."

This is a cutting edge achievement, and only time will tell how well it works. Google acknowledges that the entire process still needs refinement, and they've been working on it for quite a while. But when it is all worked out, just imagine the audio tracks of all your videos being indexed and searchable.

The user friendly side is already in place: "Additionally, since speech recognition tells us exactly when words are spoken in the video, you can jump right to the most relevant parts of the videos you find."

Learn more about this latest gadget on Google's blog post, In Their Own Words [googleblog.blogspot.com]


12:40 am on Jul 18, 2008 (gmt 0)

WebmasterWorld Senior Member tedster is a WebmasterWorld Top Contributor of All Time 10+ Year Member

I can't find the reference right now, but I'm pretty sure that Google was working on audio-to-text quite a few years ago. And in a parallel effort, using OCR to pull text out of images, too. Does anyone else remember this?

Robert Charlton

6:12 am on Jul 18, 2008 (gmt 0)

WebmasterWorld Administrator robert_charlton is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

Seems to me I remember at several of the tech expos in San Francisco during the dot-com boom that there was at least one company, maybe more, pushing speech recognition systems for video that would provide both searchability and closed captioning.

I don't have a recollection of who they were... there were so-o-o many companies then with new technologies... but Google wasn't involved with the software I'm thinking of.


11:26 am on Jul 19, 2008 (gmt 0)

5+ Year Member

If it works, it is a great technology indeed.

On the other hand... why political speeches...? ( I know why but )

I guess the only reason is because it's not perfect yet?
"Oh, [the_politician] didn't say this and that, it's just that the words weren't spelled clearly enough for the speech recognition system..." ...*hah*... viral bombs away. (j/k)

Sorry for the PR department but who would believe that this tool was developed for political videos only, especially with the candidates censoring which video can get into the widget. I mean gadget.

ok got it, it's probably much cooler than saying
Coming up October in YouAdWords:
... keyword-based contextual ad targeting following the audio transcript.
... precisely *timed* *on-topic* *clickable* ads in the video overlays.
... or not.
We don't know yet.

either way, while it may be last in YouTube's priorities... I'm not sure what excuse they'll come up for those who've been saying all along, that this could be effectively used for detection to battle off copyright infringement.


If they do it right, and not try to evade responsibilities, this may very well bring YouTube that 'holy grail' [news.cnet.com] Schmidt was looking for ( and save the money in the cookie jar from being fed into lawyers' pockets in the same time )


5:02 pm on Jul 19, 2008 (gmt 0)

WebmasterWorld Senior Member jimbeetle is a WebmasterWorld Top Contributor of All Time 10+ Year Member

While some of the transcript snippets you see may not be 100% accurate

So, is it really smart of Google to start with the political channel?

Featured Threads

Hot Threads This Week

Hot Threads This Month