Web Video Chaos Intensifies Despite Progress on Standards

Dean Hachamovitch, corporate VP, Internet Explorer

Dean Hachamovitch, corporate VP, Internet Explorer

June 4, 2011 – The long-raging battle for platform supremacy in the Internet video streaming era has taken a new turn where growing support for a Google initiative challenging MPEG-4 H. 264 as the de facto encoding standard is threatening to overshadow progress toward achieving interoperability across disparate H.264-based streaming formats.
 
The good news/bad news picture complicates the strategic planning process for stakeholders in next-generation multi-device video services, including network service providers, consumer electronics manufacturers and content suppliers, all of whom have been working in various forums to break the proprietary barriers to benefitting from IP adaptive streaming. Even without Google’s push for what it’s calling WebM, the challenges that confront these efforts to untangle the proprietary streaming mess are immense. With WebM in play there’s reason to wonder whether the upheaval will ever end.

After a year in the background, WebM has begun to gain traction with chip-based implementations, browser support from the leading suppliers, including Microsoft, and growing use on the Web, including Google’s recent launch of the platform in YouTube. Google’s aim is to create a royalty-free video streaming environment by leveraging the VP8 video encoder it owns as a result of acquiring On2 last year together with the Vorbis audio codec, an open source product of the Xiph Org Foundation, and an open-source transport container produced by the Matroska Multimedia initiative.

Or course, sites like YouTube that are starting to accommodate WebM videos are continuing to run H.264 and other coding formats using legacy HTTP (Hypertext Transfer Protocol) –based adaptive streaming systems to distribute the content. But the domains are completely separate from a content management perspective.

WebM is incompatible with the H.264 systems with respect to not only the encoders but also the container format and streaming mechanism used to assure quality. Rather than employing the adaptive streaming mode pioneered by Move Networks and adapted for use in proprietary systems by Microsoft, Apple, Adobe and others, WebM applies variable bit rate encoding techniques to maintain minimum quality requirements across multiple streams.

Meanwhile, the International Standard Organization’s Moving Picture Experts Group has made significant progress on a standards initiative that bridges the incompatibilities between Microsoft’s Smooth Streaming and Apple’s HTTP Live Streaming (HLS) with a formatting structure that also simplifies use of multiple digital rights management (DRM) formats. The MPEG Dynamic Adaptive Streaming over HTTP (DASH) format, derived from the mobile industry’s 3GPP Adaptive HTTP Streaming specification, also supports trick modes with random access, dual-streams for stereoscopic 3D, including the Multiview Video Coding used with Blu-ray 3D, and multiple content management schemes. MPEG anticipates the standard will achieve final draft status next month.

While HLS and Smooth both use H.264 and rely on HTTP, the communications foundation for the World Wide Web, they use different container formats, which dictate different file sizes and timing sequences in the adaptive streaming bit-rate adjustment process. DASH defines a server-to-client communications format known as the Media Presentation Description to instruct device clients how to set up the streaming mechanisms, enabling them to interoperate with adaptive streaming segments delivered from servers in either MPEG-Transport Stream containers used by HLS or the Fragmented MP4 containers used by Smooth.

The format also creates a way for servers to communicate to devices what the specific DRM requirements are for a given piece of content. By allowing any DASH-compliant DRM to be implemented automatically with no need for intervention by the content supplier, the architecture greatly simplifies distributors’ efforts to meet protection requirements of multiple content owners.

Clearly, making it possible to store content in a single file for distribution to any Apple HLS or Microsoft Smooth compatible device with multi-DRM flexibility would be a major step toward cutting the costs and hassles of creating IP-based multi-device premium services. But Adobe, despite its embrace of H.264 in recent versions of Flash, is using incompatible modes of distribution, including its long-running format and the more recently introduced adaptive streaming mode it calls Dynamic Streaming.

One place where the market is coming together to bridge some of these gaps, including the new one introduced by VC8, is on the browser front in conjunction with development of the next-generation Hypertext Markup Language Web page format HTML5 under the auspices of the World Wide Web Consortium (WC3). HTML5 creates a multimedia-friendly environment for video on Web sites so that browsers can play video back natively within a Web page without requiring plug-ins like Adobe Flash, Apple QuickTime and Microsoft Silverlight.

HTML5 will employ HTTP-base adaptive streaming and the concept of video tags that are used to identify content that’s enabled for HLS streaming within the Web page, making it incompatible with WebM-based streaming. However, WC3 has dropped original plans to make H.264 a de facto component of HTML5 in deference to the VC8 juggernaut. This makes sense, given that both Mozilla and Opera have allied with Google’s Chrome on a policy of supporting VC8 exclusively in all future versions of their browsers.

Another sign of the gathering momentum behind VC8 was Adobe’s recent confirmation that it will be supporting the codec in a forthcoming release of its Flash Player. The move, coming in response to Google’s release of VP8 as open source code in mid May, “will help provide users with seamless access to high-quality video content on all of their Internet-connected devices,” says Michelle Perkins, product marketing manager for mobile and devices at Adobe.

Discussing the strategy in a recent blog, Perkins adds, “By including VP8 as part of the Flash Platform, we’re providing companies with a choice as to how they can work with and deliver great experiences to the Web. We are excited to work with Google and others to ensure Web video continues to evolve and better serve content publishers, Web developers and end users.”

Google is promoting VC8 and other components of WebM on a royalty-free basis on the assumption that as Internet development becomes ever more video centric the licensing costs associated with H.264 could impede growth. This dovetails with the WC3 policy favoring royalty-free technology.

But just how royalty free VC8 turns out to be remains to be seen. MPEG LA, the royalty collection body for H.264, is taking a hard look at whether any of the VC8 components draw on its members’ patents.

In a February blog, Microsoft Internet Explorer vice president Dean Hachamovitch, citing his firm’s failure at trying to make Windows Media Video (standardized by SMPTE as VC-1) a royalty-free codec, suggested Microsoft would only support WebM if Google showed a willingness to indemnify users against patent claims or come up with some other way to resolve uncertainties. “Asserting openness is not a legal defense,” Hachamovitch said.

But after initially limiting Internet Explorer support for VC8 to instances where Windows users download WebM clients, Microsoft announced in March that it was supporting playback of both H.264 and VC8 from HTML5 Web pages. “IE9 is the only browser today committed to supporting both formats directly,” the company said in a blog post.

Microsoft also said it was taking another step to bridge the H.264/VC8 chasm by introducing an HTML5 extension for Chrome in Windows Media Player so that Windows 7 customers using Chrome can play H.264 video. This is similar to what Microsoft did to allow Windows customers to play H.264 in Firefox, which has never supported H.264.

As these efforts to accommodate the Google initiate take shape, the company, in cooperation with Matroska and the Xiph.Org Foundation, has moved to shore up protection for users of WebM by creating what it calls the “WebM Community Cross-License Initiative.” The 19 founding members include industry giants AMD, Cisco Systems, Ericsson, Huawei, LG Electronics, STMicroelectronics and Texas Instruments along with the Google allies and several smaller firms, all of whom have agreed to “license patents they may have that are essential to WebM technologies to other members of the CCL,” according to the new community’s announcement blog.

But there’s still no indemnification guarantee, as requested by Microsoft’s Hachamovitch. And, of course, there are many major players missing from the cross-licensing initiative, including Intel and Microsoft.

Nonetheless, hardware implementation of VC8 and other WebM components is well underway, with more than 20 licensees worldwide now developing WebM chipsets, some of which are already in commercial production, according to the WebM hardware development team in Oulu, Finland. Many of these chips support VC8 at up to 1080p resolution and 60 frames per second, the team said.

In May the team released a second version of the VP8 hardware encoder, dubbed “Blueberry,” with a focus on improving the encoder for video calling use as well as enhancing overall performance quality. According to a blog announcing the release, the new encoder delivers 30 fps, 720p video at 1.2 megabits-per-second with a peak signal-to-noise ratio of 33dB, compared to a data rate of over 1.5 mbps for a comparable performance level on the previous encoder release.

Clearly, WebM is a force to be reckoned with. DASH will be a big help in efforts to streamline IP content management, but no one can afford not to take WebM into account as new content distribution plans take shape.