“Open video” could beget the next great wave in Web innovation–if it gets off the ground.
In 2005, Michael Dale and Abram Stern, a pair of grad students in digital media arts at the University of California, Santa Cruz, decided it would be fun to make video remixes of speeches in the U.S. Congress. Their goals were artistic; Stern had notions, for example, of editing a Senate floor speech to remove everything but the pronouns. They would be following, loosely, in a tradition of video commentary that includes remixing speeches from the 2004 Republican National Convention to feature only the many utterances of terrorism or September the 11th by George and Laura Bush, Dick Cheney, Rudy Giuliani, and others. Aware that congressional proceedings are public–and that C-SPAN airs them freely–the pair went online to hunt for the raw material. But “the footage wasn’t there,” Dale recalls. While C-SPAN did offer archival material for a fee, he says, “if we wanted to pull together a few different clips of senators saying different things–there was no online repository for download.”
So they bought a computer and several hard drives, which they hooked up to a television, and started unabashedly copying C-SPAN’s congressional coverage. Then, in March 2006, they went live with a website called Metavid.org, hosted by the University of California, which offered the purloined legislative footage free for the downloading. Before long, C-SPAN–a nonprofit company created by the cable industry–claimed that the university was violating its copyright. When university lawyers learned that only the videos of committee hearings had been shot by C-SPAN’s cameras (proceedings on the floor of the House and Senate were recorded by government cameras), a compromise was reached: floor footage could stay up (with the C-SPAN trademark removed), but the committee footage had to be taken down. C-SPAN later liberalized its policies to allow free reuse of federal-government coverage–but it excluded commercial use. This is not something Metavid could promise, so the hearings remain unavailable on the site.
As they looked for alternative sources of committee footage, Dale and Stern encountered a thicket of technical problems. It turns out that many (though not all) congressional committees do make their own videos, and some of these committees allow you to play the videos on their websites. But the technologies involved reflect the chaos of competing formats that characterizes Web video today. To pick two examples: the Senate Commerce Committee offers videos in a Flash player but offers no download link. And the House Judiciary Committee still uses RealPlayer, a format that’s now largely obsolete. Any would-be users of these resources would soon run into trouble. Where download links weren’t provided, they’d need special software to copy the video from the government site. Once the videos were in hand, they’d have to buy software to do any necessary format conversions and editing. And finally, they’d have to upload the results. “All of these offerings are difficult to reuse in a video project,” says Dale.
Open Video in Practice
How a remix was made--and how it could have been easier.
By David Talbot
A similar transformation of video would not just allow trouble-free playback of any video you might encounter. It would also mean that any innovation, such as a new way to search, would apply to all videos, allowing new technologies to spread more rapidly. And it would make it far easier to mix videos together and create Web links to specific moments in different videos, just as if they were words and sentences plucked from disparate online text sources: imagine linking part of a politician’s speech to a contradictory utterance years earlier. “In 1993 people thought AOL’s newsrooms were mind-blowing, because that’s all they were exposed to,” says Dean Jansen, outreach director of the Participatory Culture Foundation, a nonprofit group that is developing an open-source video player called Miro. “Now they can write their own blogs and find and read hundreds of thousands of news sources and blogs, from all over the Internet. I don’t think it’s an exaggeration to say that this is the scale of change that would become possible if video [technologies] were totally free online, like text and images.”
Today, Dale works toward realizing that vision as part of an effort by the Wikimedia Foundation, which launched and operates Wikipedia, to create video companions to the online encyclopedia’s text entries. The idea is that you’ll be able to search the Web for snippets of video, import them into a Wikipedia article, and keep track of edits–all using open technologies that don’t require video plug-ins or software purchases. One hope is that Wikipedia, as the world’s seventh-largest website, will help drive video openness generally, says Chris Blizzard, director of technical evangelism at Mozilla, which is supporting the project. But the larger point is that efforts like these will make it far easier for anyone to innovate with video and for anyone else on the Web to enjoy those innovations. The results are impossible to predict, except through the example of what the open Web has provided so far. “Nobody is going to tell you they want something before it emerges,” Blizzard says. “Rather, the experience of the Web is: ‘Holy cow, I can do this other thing now!’ Open standards create low friction. Low friction creates innovation. Innovation makes people want to pick it up and use it. But it’s not something where we can guess what ‘it’ is. We just create the environment that lets ‘it’ emerge.”
Let’s Go Crazy
YouTube has helped make video a mainstay of the Web, thanks largely to its simplicity and user-friendliness. Anyone can open a YouTube account and upload videos, and anyone who visits YouTube can easily find and watch videos, all free. It has become the world’s third-most-popular website, with 41 percent of the video-hosting market. A recent analyst report by Credit Suisse predicts that YouTube will serve up an astonishing 75 billion video streams this year, to 375 million users. And every minute, YouTube’s burgeoning servers slurp up 20 hours’ worth of newly uploaded user videos, says the company’s director of product management, Hunter Walk. Susan Boyle, the Scottish songstress phenom? The latest footage from Tehran’s street protests? Bulldogs on skateboards? Your cousin’s baby video? It’s all there, available in a few clicks.
And as YouTube grows and adds features, it continues to stress simplicity and user satisfaction (much in the spirit of its current owner, Google, which bought YouTube in 2006 in a deal worth $1.65 billion). Among other features, it has introduced ways for users to add elements such as captioning to their videos, build on their social networks (by automatically alerting Twitter followers when they upload new videos, for example), and annotate videos with computer-readable tags to improve search results. Other new tools can help businesses manage their YouTube-hosted videos and learn who is watching them. “YouTube represents a unique place in the video ecosystem; the breadth, depth, and freshness of content is unparalleled,” Walk says. “The best years are ahead of us.” In 2009, uploads of videos from mobile devices are up 1,700 percent–400 percent just since the release of the new iPhone 3G, he says. And the only obvious price of such service is exposure to advertising.
Internet video is thriving in other respects, too. Not just YouTube but Apple TV, Windows Media Center, Hulu, and more are making it possible for computers and mobile devices to deliver programming normally associated with television. (YouTube, for example, in a bid for growth and revenue, is offering premium channels with short-form content from entertainment titans Disney, ABC, and ESPN.) Boxee, a New York City startup, is bringing things full circle with a browser that enables you to play any media available over the Internet on your TV screen; the interface is designed for easy use from across the living room.
Against this backdrop, there initially seems little to dislike about YouTube. But its sheer size makes it an easy and tempting target for filtering by national governments (Iran, for one, has done just that). The result is that video can, in some contexts, be censored more effectively than other forms of Web content. Similarly, YouTube is a convenient target for legal action by media companies trying to protect copyright, sometimes in ways that overstep the bounds of common sense. Two years ago Stephanie Lenz, a Pennsylvania mother, got an e-mail from YouTube announcing that it had taken down her shaky 29-second movie of her toddler son, Holden, giggling and dancing as the Prince hit “Let’s Go Crazy” played distortedly in the background. YouTube explained that it was responding to a request from the Universal Music Publishing Group, which owns the rights to the song. She argued that her video represented “fair use,” and it was reposted. But she decided to sue Universal Music, claiming that it was abusing the Digital Millennium Copyright Act. (Universal Music later said it had issued thousands of so-called takedown notices on behalf of Prince alone; the Artist himself is fanatical on the subject.)
While artists have every right to thwart wholesale copying, such crackdowns on incidental, noncommercial use–which is generally quite legal–can inflict collateral damage on innovation. “When people make their own mixes of existing material and YouTube takes that down, this is a huge inhibitor to this kind of commonplace creativity that the Web enables,” says Abigail De Kosnik, a professor of new media at the University of California, Berkeley, and author of Illegitimate Media: Minority Discourse and the Censorship of Digital Remix Culture. “What people need to realize is that too much of those kinds of protections and [technology] restrictions–and right now, without open video, we have too much of both–inhibits new genres from emerging.”
Finally, the need to generate revenue is driving YouTube further toward a centralized, television-like model, with advertising-supported premium content. In short, while it’s never been easier for the average Internet user to find and consume video online, you can’t easily adapt or reuse what you find in the vast body of video out there. “The video box that you see on YouTube is a whole bunch of different formats inside this plug-in that isn’t manipulable, transformable, or remixable in the way that everything else on the Web is,” says Mark Surman, executive director of Mozilla. You can’t even download videos you play on YouTube–at least not without help from third-party websites or software.
YouTube sees little need to add features such as downloading tools. “We haven’t gotten enough feedback that we need downloads,” Nikhil Chandhok, a YouTube senior product manager, said at a recent conference in New York City. “You are mostly connected all the time … and can access any YouTube video you want.” Even if you do go to the trouble of using third-party services to download videos, if you want to do anything creative with those videos, your work has just begun. You will need to convert various formats, buy video editing tools, and learn to use them. (Walk did not want to talk about open standards, except to say that the company has “been about a lot of kinds of openness early” in terms of expanding access to video itself. Of course, making video easier to work with outside of YouTube would tend to threaten YouTube’s dominance.)
If YouTube is the epicenter of the Web’s video revolution, Wikipedia is the epicenter of online collaboration. In the eight years since its founding, it has grown to become not just the dominant online reference but an increasingly important source of real-time news, with more than 13 million frequently updated entries, including 3 million in English. But these two hubs of free, user-generated content operate as if in separate universes. Wikipedia, which makes it easy to alter content, offers few videos to play (though about 3,000 videos can be found scattered around the site). YouTube, with millions of videos available, offers few options for editing or innovating with them. Generally, each site’s best qualities as an information resource are all but absent from the other.
But that could change as Wikipedia strives to add features that permit effortless open-source video editing and remixing. Michael Dale, the former Santa Cruz student, is leading the effort at Wikimedia under the sponsorship of Kaltura, a startup with offices in New York City and Israel. Kaltura is developing open-source technologies for playing, editing, and uploading videos. A major benefit of open video is that the video itself can be extracted from the player, just as an image can be extracted from a website when you right-click it. With the new version of HTML technology, HTML 5, an open-source player is included in the browser–no plug-ins required. Mozilla’s newly released Firefox 3.5 browser, Apple’s Safari, and Google’s Chrome (see “An OS for the Cloud,” p. 86) all support this feature, though Safari requires a plug-in to support a specific open video format, called Ogg Theora, that Wikipedia is using. And if history is any guide, these advances by competitors may goad Microsoft to follow suit with improvements to Internet Explorer. “Right now, when you post a Flash video, you are posting the video and also a plug-in player, and that can make it difficult to access the video file itself,” says Dale. “Once video is just another asset on the Web and something browsers can natively deal with, we can pull audio, video, images, and text from anywhere on the Internet and do the kinds of sharing and editing and remixing that you want to do, all in the open Web platform.” Can Wikipedia really change the way everyone uses video? “When Wiki started, people said it wouldn’t work, but it worked,” says Kaltura cofounder Ron Yekutiel. “The next question is: Why should it stop at simple media?”
The results should start to become visible this fall. If you are editing a Wikipedia entry, you will find an “Add media” button. Clicking it will bring up an interface that will, initially, allow you to search through three repositories of free licensed multimedia files. One is Metavid, the congressional archive started by Dale and Stern. Another is the Internet Archive, the San Francisco-based digital library most famous for archiving old Web pages; it also holds hundreds of thousands of old interviews, documentaries, and films contributed from various sources. The third is Wikimedia Commons, a multimedia repository operated by the Wikimedia Foundation itself.
Some observers think Wikipedia’s foray into multimedia will help move the entire Web toward open video standards. “To make video part of the fabric of Wikipedia will provide incentives to [video] producers to get their stuff out there and indexed,” says Jonathan Zittrain, a Harvard Law School professor and cofounder of the Berkman Center for Internet and Society at Harvard University. Producers who want their videos excerpted and linked on a Wikipedia page–drawing more traffic to their own websites–will not just have to put much less restrictive licenses on the material; they’ll also have to accept open standards rather than proprietary ones. “With no business model yet gelled, this is just the right time for Wikipedia to be experimenting, and possibly leading, the development of open tools and content for video,” Zittrain says.
Jimmy Wales, Wikipedia’s founder, sees the effort as the next logical advance in Web technology. “Today any computer programmer in the world can launch a website and have full-strength tools for creating new things,” he says. But he points out that this is not yet true for video. No collaborative video editing process is available to all Web users. “It’s a process that’s a lot harder to do if all I can do is download a 60-minute video to my computer, open up some [proprietary] software to edit the video, then upload it,” Wales says. “There’s no easy way for other people to give direct feedback. The record of the edits isn’t there. And if someone else wants to change it, they have to redo all the work on their computer.”
Wikipedia’s effort to promote open video standards isn’t the only one; the YouTube competitor Dailymotion, for example, is making 300,000 videos available in the Theora format. But whatever the catalyst, wide acceptance of such standards could have important implications even for people who don’t want to make their own video remixes. In particular, it could drive broader and faster advances in video search. Consider Blinkx, which has indexed 35 million hours’ worth of videos and devised a variety of ways to search them, from simple means–metadata, or computer-readable tags that literally describe what’s in a video–to advanced techniques involving speech analysis and facial recognition. One method devised by Blinkx allows searchers to draw a box around a face in a video, click it, and then search the Web for other videos containing that face. But for that trick to work with all Web videos, Blinkx must rebuild the interface code to accommodate each of a handful of dominant video formats and 80 lesser-used ones. “If open video works, then all the people doing these kinds of innovations within individual video formats–they can all talk to each other,” says Suranga Chandratillake, the company’s founder and CEO. “It means innovation isn’t split into separate groups in separate formats. Today the video Web is written in tens of languages, causing all the usual barriers when you want to switch from one to the next. With a dominant open format, everything will link to everything else; viewers will be able to freely watch content and jump between relevant clips.”
And on the copyright front, Creative Commons, the nonprofit organization that has provided usage licenses for 250 million copyrighted works, is helping to clarify what existing video works can and can’t be used. “Open licensing is a crucial part of this fairly multilayered ecosystem that will make open video take off,” says Mike Linksvayer, Creative Commons’ vice president. “If the video itself, and the components of the video, like music, aren’t actually openly licensed, then each of the other layers is hindered.”
Lately, Mozilla’s Blizzard and Surman have been showing off something a Mozilla developer cooked up with open-source video tools. In their video, the two men walk in and out of the camera’s field of view. A thought bubble dances over each head (tracking their movements thanks to face recognition software); inside each bubble, their real-time Twitter feeds are displayed. This was all done with Theora, HTML 5, and other new standards, Blizzard says. While such a stunt could be performed with proprietary software, it wouldn’t be so easy–or so easily shared. “This is what we mean when we talk about taking video out of the plug-in prison and allowing people to create things,” he says. The goal isn’t to make any one application possible but to bring about the next Internet revolution–one whose specific form is hard to foresee, except that it’s likely to be televised.
David Talbot is Technology Review’s chief correspondent.