Startups Rush to Pave Way for Web Video
By Peter Svensson (Monday, January 28, 2008)
NEW YORK - Video on the Internet has gone from being the next big thing to the current big thing. But murky YouTube videos are just the start — there's a lot of room for improvement. A raft of startups are rushing to supply the tools to make better and more profitable video available. Nearly a dozen video-related startups will be presenting at the DEMO 08 technology conference, starting Monday in Palm Desert, Calif.
The biannual conference, which gives startups and more established companies six minutes on stage to pitch their new products to investors and media, has been the launching pad for several successful ventures over the years, including the Java programming language, TiVo Inc. and Half.com.
Several of the 77 companies presenting this time have been tackling the problem of taking video quality to the next level. It's quite possible to send high-definition video over the Internet, but the cost of doing it at scale is daunting, because it requires about 40 times the bandwidth of a YouTube-quality video.
"If you run any infrastructure that lets people share video, it's really, really expensive," said Dan Putterman, chief executive of San Francisco-based Squidcast Inc.
The company is launching a service that allows users to send video they've shot with their high-definition camcorders to friends and relatives at full resolution, for free.
To do this, users will take help from other users, in a manner similar to peer-to-peer file sharing programs like BitTorrent and KaZaa. Each file that is uploaded gets distributed in small chunks among the computers of many users (who won't notice the chunks or be able to look at them). The intended recipient gets an e-mail with a link. Clicking it starts the download, which pulls the chunks together from the network of user computers like a squid pulling in its tentacles.
In other words, Squidcast itself doesn't need to devote computers or buy bandwidth to transfer user's files. It will finance the service by showing short video ads to the recipients while they download.
"If someone attempted to do this as a hosted platform they would simply go out of business," Putterman said. "It can't be done and that's why it hasn't been done."
Atlanta-based Asankya Inc. is trying to solve the same problem, but for Hollywood rather than home movies. CEO Scott Ryan puts the current cost of distributing an HD movie online at about $3.
Considering movies rent for $4 to $5 and the creators have to be paid, there's no real money in it for distributors.
Asankya's solution deals with a fundamental problem of the Internet, which wasn't designed for transfers of HD-size files. Under regular Internet protocols, all the little parts that make up a file take the same route over the network, even if that path becomes congested during the transfer. Asankya's Hypermesh service, which it is previewing at the show, can send individual parts of a large file over different routes, then reassemble them in the right order.
"We just use the network much more efficiently," Ryan said. "We get cost advantages out of that."
The technology was developed by a Georgia Institute of Technology professor and uses servers placed at strategic points on the Internet in addition to information sent in by receiving computers.
The catch is that Asankya is mucking around with fairly basic networking technologies, ones that are built into operating systems. For now, receiving computers need software that only runs on Windows XP.
BitGravity Inc. is aiming for the same market, with a network of servers designed to deliver high-definition video. It launched the service three months ago, and at the show, the Burlingame, Calif., company plans to announce that it will be expanding the service to deliver live, streaming video.
Live streaming has been a "luxury" only the big media companies could afford, said Perry Wu, BitGravity's CEO, leaving a lot of unsatisfied demand, for instance for regional sportscasts. "Whether its basketball or field hockey or water polo or swimming, the number of people who want to watch that in real time is tremendous."
Demonstrations on BitGravity's Web site show high-resolution — if slightly jerky — video that starts almost immediately.
Two other startups, Eyealike Inc. and Visible Measures Corp., are tackling other problems close to the heart of content providers. Eyealike of Bellevue, Wash., will be demonstrating software that can scan videos submitted to Web sites to see if they contain copyright material.
Such filters are already in place, for example at YouTube, but Eyealike President Greg Heuss said the startup's product is better than the competition in that it can identify video that's had its audio stripped out or been cropped to as little as a quarter of the original frame.
Boston-based Visible Measures will be touting its service, which lets Web sites track how viewers play their videos: where they pause, what they rewind to see again. That should help the sites figure out which videos and ads that actually hold the viewers' attention, said Matt Cutler, the company's vice president of marketing.
"The challenge is: historically, no one's really known what happens after the play button is pressed," Cutler said. "What we're introducing is the ability to sort of peer beyond that."
Originally appeared in Associate Press and Yahoo.