HTML 5: Video Codec Options (Unfortunately) Thrive
While the HTML 5 specification notably supported by the just-released Opera v10.5 browser (and with support planned for Internet Explorer 9) may not obsolete Adobe Flash (or, for that matter, Microsoft’s Silverlight), in no small part to the longevity (and arguable necessity) of DRM-inclusive content, its usage will likely grow by leaps and bounds in the future. DRM-less video content, for example, might bypass Flash and natively render video windows (and associated surrounding player apps) within the browser, as Vimeo’s and YouTube’s experiments (the latter following in the footsteps of NeoSmart Technologies) foreshadows. But as my early August 2009 piece points out (and is unfortunately still the case), HTML 5’s video codec situation is a mess and shows no sign of sorting itself out any time soon.
The two codecs currently supported by the HTML 5 <video> tag are H.264 (aka MPEG-4 AVC, aka MPEG-4 JVT, aka MPEG-4 part 10), and Ogg Theora. H.264 backers rightfully point to Ogg Theora’s unproven and potentially infringing intellectual property situation, which even mighty Microsoft encountered when it pursued standardization of Windows Media Video at SMTPE as VC-1. Ogg Theora backers (which have lingering issues of their own regarding their preferred implementation) rightfully point to H.264’s delayed but likely inevitable royalty-payment potential to content developers, content providers and/or software developers.
And both groups point to supposed shortcomings in the other’s video quality results. Poor Jan Ozer, whose work I’m proud to say I’ve long admired and followed, is the latest victim of this crossfire. When Jan gave the conservative nod to H.264, the open-source zealots turned on their flamethrowers, questioning his methodology, his content choices, and his chosen utilities and their setting selections.
Long-time readers may remember that yours truly has also done some work in this area, and let me tell you, it ain’t for the faint of heart. Brass cojones and asbestos-coated Jockeys are prerequisites, and no matter how much time and energy you pour into the project, inevitably you’ll walk away frustrated. There are so many possible combinations of content (material, frame rate, frame size, color depth), compression configuration options and other variables, and they’re all so dynamic with respect to time, that results are inevitably not only incomplete but also obsolete by the time they’re published.
Meanwhile, the On2-developed VP8 lurks in the shadows (the company’s acquisition by Google finally closed last week, after—among other things—a renegotiation of the purchase price), and nobody knows for sure what Google’s plans are for it. Will Google open-source it, as On2 did with prior-generation VP3 (which ironically was the foundation for today’s Ogg Theora)? If so, will the HTML standardization body accept it as yet another <video> tag codec option? And if so, will it become the de facto standard for HTML 5 video? My own limited testing attests to the codec’s high quality, even in the form of its VP6 predecessor, which bodes well for its competitive position versus H.264. And On2 and its implementation partners have presumably done a thorough job of vetting and addressing VP8’s patent infringement potential
In addition to the earlier-mentioned Vimeo and YouTube HTML 5-based video players, Jilion is developing a product called SublimeVideo, specifically targeting WebKit-based browsers (including Internet Explorer with the Google Chrome Frame add-in). Nuanti is working on an Ogg Theora (video)/Vorbis (audio) enhancement to Microsoft’s Silverlight browser plugin. And the latest Xiph.org codec plugin pack for Windows includes experimental HTML 5 <video> tag support in Internet Explorer for Ogg Theora (but not H.264), as well.
Meanwhile, Adobe soldiers on with its Flash advocacy and development, dismissive of the handheld trends (specifically Apple’s looming market presence) that may neuter Adobe’s aspirations, along with users’ growing frustrations with plug-in vulnerabilities. Corporate blog posts criticize Apple’s anti-Flash stance. Meanwhile, hardware-accelerated Flash 10.1, initially unveiled last October, inches towards ‘gold’ release. Check out AnandTech’s benchmarking of the initial beta from mid-last November for a taste of the upcoming version’s performance and capability potential.