« Searching For Last Minute Gifts? | Main | AV/ATW and Video Search »

December 15, 2004

Yahoo! Video Search Beta

There were some rumors a few weeks back about Video Search products coming in 2005. Well, we're ready to show you what we've got today--and to ask for your feedback. An early Yahoo! Video Search Beta is now up on Yahoo! Next, our preview site for new technology and applications. I've spent a few hours with it in the last few weeks. Go try it out and let us know what you think. Remember that it's a beta product.

Why Video?

The costs of producing video content have been steadily decreasing in recent years. Between the adoption of broadband Internet connections, and easier to use video editing software, it's no surprise that we're seeing a lot more video content make its way on to the Internet. And what's out there today is just the tip of the iceberg.

The Backstory

But there's more to the story here than the blossoming world of on-line video and building a video search system: it's often not easy for a web crawler to find downloadable and streaming video content. Unlike web images and most audio files, videos aren't always easy to discover. In many cases, they're hidden behind complex JavaScript, Flash-based players, and other non-crawler friendly obstacles. That's exactly why we've talking to a lot of our existing media partners, many of whom have sizeable video assets which have yet to be indexed.

Enter RSS

When we started thinking about how to make it easier for anyone to expose video and other rich media content, one of the first things we thought of was podcasting and RSS. Podcasting uses RSS Enclosures to provide an audio file along with a news item or blog posting in an RSS feed.

So rather than build a completely new way to do this, we decided to see what it takes to make RSS Enclosures work for video content as well: video enclosures. It's not a new idea but we think it's one whose time has come.

At the most basic level, this is just a matter of pointing to a video instead of an MP3 file.

Instead of this:

<enclosure url="http://www.example.com/baby_walks.mp3"
 length="64358" type="audio/mpeg"/>

You could use this:

<enclosure url="http://www.example.com/baby_walks.mov"
 length="2144275" type="video/quicktime"/>

For many publishers, that's all it takes. The beauty of this is that there's existing infrastructure for handling simple enclosures. Many RSS readers already consume enclosures just fine.

In the very near future the Yahoo! Video Search crawler will support indexing video enclosures in RSS feeds.

Metadata Extensions

As Marc Canter has noticed, we could all benefit from a bit more metadata to go with this growing pool of media. Who published this video? What formats are available? How is it licensed?

From our point of view, it means we can build a much better video search. You might want to filter results based on some of that metadata (title, actor, file format, etc). But it also opens up so many more doors. For example, your news aggregator might use your preferences to figure out which videos to download: Windows Media or Quicktime? High bandwidth or low? Heck, we can see entirely new rich media aggregators and tools being built--something like the popular iPodder currently used for podcasting. And when they are, this metadata becomes all the more important.

To get this started, we're suggesting an optional set of metadata extensions that we've been calling "Media RSS" (yes, we're so creative with names). They're aimed at publishers who'd like to provide a rich set of metadata about the media being published. Our video search system will also support these Media RSS extensions in addition to video enclosures (see the FAQ and the draft spec).

In addition, we're working with several other companies and organizations to help refine these ideas. They include: AtomFilms, Creative Commons, Buzznet, Ourmedia, and Broadband Mechanics.

Evolution

This is all about helping RSS evolve to handle all the media types we might want to attach to RSS feeds. Early last month when InternetNews.com asked me about enclosures, I said two things that are worth repeating in this context:

Enclosures are more evolutionary that revolutionary.

That's exactly why we want to start by building on the foundation of enclosures rather than introducing a brand new format.

RSS is a reasonably flexible format for distributing content. Enclosures are to RSS what attachments are to e-mail.

In other words, enclosures aren't just for podcasting anymore. If you want to be a video blogger, enclosures should work for that too.

If you're a user, aggregator author, or content provider who'd like to get involved with the development of Media RSS please join the rss-media group. Of course, you can also leave a comment here or trackback this post to let us know what you think.

Jeremy Zawodny
Yahoo! Search

P.S. There a lot of funny stuff out there like Monkey Karate :-)

Comments

The link under the Yahoo! Search Video Beta actually ends up at http://search.yahoo.com/search/options. Quibble, quibble.

Very cool. As a tech geek, may I ask how the streaming media is downloaded into the search repository? Is open source software used?

That's cool, guys. This tool proves that there's still a lot of work to do in the Search Engine World.

It kinda looks like ATW results for me, but it's nice though.

Well done. Results were pretty accurate in my initial tests.

Nice, but how come an "early beta" of Yahoo Video produces exactly the same results as the established Altavista Video Search? Apart from the obvious Yahoo-Overture-Altavista ownership chain, I mean. Did you merge the Altavista engine with Yahoo's or viceversa?

Nice! I obtained really interesting results. Makes me want to enter the digital world of video and release little movies on the internet as well. Keep up the good work !

This is a great way to pull up tutorial videos for software applications.

What's the next big thing? Videoblogging?
http://en.wikipedia.org/wiki/Videoblogging
http://groups.yahoo.com/group/videoblogging/

What about Lifetime Digital Memory?
http://divedi.blogspot.com/2004/06/lifetime-digital-memory.html

Blog + Photos + Audioblogging + Videoblogging = Lifetime Digital Memory

See also:
The New Moore's Law
http://divedi.blogspot.com/2004/10/new-moores-law.html

Wow! This is the best online video search I've seen yet. This thing has tons of content. Thanks Yahoo.

Gracias por este nuevo instrumento de búsqueda y difusión de contenidos. Supongo que gracias a él los contenidos cientificos de algunas web de difusión quedarán más cerca del público. A mi me gusta.

muNice, but how come an "early beta" of Yahoo Video produces exactly the same results as the established Altavista Video Search

I think one of the more important features of this will be the categorizing of data by format, region, style, etc, even more so than type of player. While it is easy for a bot to "search" a document for keywords, searching and categorizing this sort of content (and audio) is proving to be a much more labor intensive battle. I've been up to my ears in it for years trying to categorize radio stations and podcasts.

as far as video search goes, i am loving blinkxtv. has anyone tried it out? they have 22 channels and show action videos! i was so disappointed with googles version...wow, they are lagging

I've tried Google, Yahoo, Singingfish and Blinkx. So far, my money is with Blinkx. It is really straight forward and gives me transcriptions of the video as well as the video. I mean it is pretty basic. I don't know what else Google is going to come up with, but for now, it's all about Blinkx.