Searching content INSIDE audio and video
April 27, 2008 by Mel AclaroAs a developer of online media, one of the drawbacks I've always hated about placing media up on the web, such as videos, podcasts, webcasts, audio streams, and so forth is the hit I take in discoverability.
That is, the major search engines can't find any of the words that are spoken inside a video or podcast. Let's say, rather then typing this weblog entry, I instead decide to make it a v-log (video blog) entry. You get the picture? Same thing, by the way, for those nifty videos of your property listings, as well as any video testimonials you've placed on your web site, let alone if you're one of the growing number of agents who publish their own version of online talkshows. Nope. Can't find what you say.
But I'm hopeful. Everyzing is a company that I've been keeping an eye on. I actually posted about them in an article I wrote last month on another blog. Everyzing has been perfecting their speech-to-text technology that essentially ingests audio and video content and converts the spoken word to text-based metadata. That, in turn, becomes discoverable. But the feature that I think is really slick is that, rather than the video simply appearing in the results page of a keyword search, it's possible for the keywords themselves to point to the precise timecode in the video/audio file at which the keywords were uttered.
You can see how this works by going to a site at which Everyzing's technology has been implemented. At Boston Globe's online website, for example (click the image below), conducting a multimedia search on the words "short sale" (in quotes so you get the exact term), you can click on the timecode link to go directly to the relevant point that's one minute and forty-four seconds into the video rather than wait to scrub through the video from the beginning.
As a developer of online training media for real estate professionals, I find this technology as making for some compelling applications in elearning. Now, rather than scrubbing through an entire video in a learning module to find the exact place where the instructor was recorded as having talked about the "mortgage forgiveness debt relief act", for example, participants could be given the ability to search for the phrase and be directed to the exact point in the video segment.
Now, all I have to do is talk my company into making the investment... ;-)
| Attachment | Size |
|---|---|
| bostoncom.jpg | 76.76 KB |















Matt Jones
President /CEO
FavoriteAgent.com
Post new comment