Pretty catchy headline, eh? Well, I wrote it because I’m tired of seeing headlines like this one in CNET – “Mefeedia: HTML5-compatible video is on the rise.” Kind of makes you think that websites are adapting HTML5 en mass, doesn’t it? Well, the reality is, they’re not.
The second paragraph of the CNET article states:
A whopping 63 percent of all videos on the Web are now HTML5-compatible, compared to only 10 percent just a year ago, according to video-sharing site Mefeedia. Instead of relying solely on Flash to display their videos, many more Web sites are adopting video formats that can run directly in HTML5-compatible browsers.
Sounds pretty newsworthy, no? The next graph tells the tale.
The majority of the sites uncovered by Mefeedia are using H.264, the most common video format since it’s also compatible for playback using Flash. Google‘s VP8, or WebM, video codec is second on the popularity charts, followed by Ogg, aka Ogg Theora.
Actually, Mefeedia states that the “vast majority” of videos were in H.264 format, but let’s not be picky. Unstated, but undoubtedly true, is that the majority of those H.264 videos were encoded for playback in Flash. What the Mefeedia article does say is “H.264 is still the most common format, as it is compatible for playback in Flash as well as the browser natively.”
Either way, the authors of both the Mefeedia and CNET articles are selling the fiction that H.264 support equals HTML5 compatibility. That’s like arguing that if you speak English, you can chat with everyone in Canada, an assertion that many folks living in Montreal and Quebec would vehemently disagree with.
Why is that? Because there is no single compression technology (or codec) that all HTML5 compatible browsers play. By way of background, HTML5 compatible browsers don’t use plug-ins like Flash or QuickTime to play video, they have their own native codec support, which means that the browser vendor has to build playback support for that codec in their browser. There are three primary HTML5 compatible codecs, H.264, Ogg Theora and Google’s WebM, and for various reasons, no single codec is supported by all relevant browsers.
Let’s discuss this with some March 7, 2011 browser share numbers from NetMarketShare for perspective.
The three versions of Internet Explorer, comprising 54.33%, won’t play videos via the HTML5 video tag, H.264 or otherwise.
Firefox dominates the rest of the top ten, and has the largest share of HTML5 compatible browsers (54%), but neither version 3.6 or 3.5 play H.264 video via the HTML5 tag, and it’s likely that no version of Firefox ever will. Instead, Firefox versions 3.5 and 3.6 play Ogg Theora, with version 4.0, currently in beta, also playing Google’s new WebM video format. So if you buy CNET’s and Mefeedia’s “logic,” you can make your site HTML5 compatible by using a format that won’t play on 54% of the current installed base of HTML5-compatible browsers.
CNET and Mefeedia are pointing to videos encoded into
H.264 format for playback in Flash,
and calling that newsworthy proof of HTML5’s acceptance.
That’s a bunch of crap in any language.
To complete the picture, while Chrome 8 and 9 (totalling 10.3%) do include H.264 playback via the HTML5 video tag, Google recently announced that future versions of Chrome will not, so that number will decline over time. Instead, Google will support only Ogg and WebM.
Safari versions 4 and 5, totaling about 5% share, do play H.264 video via the HTML5 video tag, and future versions of Safari will as well, though Safari doesn’t support Ogg Theora and won’t support WebM.
Otherwise, Microsoft has shipped Internet Explorer 9, which is H.264 compatible, but does not include native support for WebM. Instead, according to a Microsoft blog, “We will provide support for IE9 users who install third-party WebM video support on Windows and they will be able to play WebM video in IE9.” That’s nice, but it still means that large percentages of IE9 users won’t be able to play WebM.
In the meantime, MPEG-LA, who administers the patent rights to H.264, has issued a call for patents essential to WebM, which could lead to royalties for anyone using WebM. In a bizarre twist, on March 4, the Wall Street Journal reported that the Justice Department is investigating MPEG-LA to determine whether it’s “unfairly trying to smother a free rival technology for delivering online video that is backed by Google Inc.”
The facts are these. HTML5 video support is confusing mess. What we do know, however is that to be fully HTML5 compatible, and support all current HTML5 compatible browsers, your site would have to include video encoded in three codecs, Ogg, H.264 and WebM. Citing video encoded in any one of these formats – particularly H.264 – as proof that HTML5 is gaining ground is a complete fiction.
Why the headline for this article? Recently, I toured over 100 high profile broadcast and corporate sites, testing to see which, if any, were fully HTML5 compatible, and could play video in all HTML5 compatible browsers without a plug-in. The only one I found was Wikipedia.org. The vast majority of sites — particularly broadcast sites – used Flash as their primary interface, with notable exceptions of Microsoft and Apple on the corporate side. Visit Microsoft without Silverlight installed, or Apple without QuickTime installed, and you couldn’t play any of the videos on the site, even from an HTML5-compatible browser that supported H.264.
The reality is that plug-ins like Flash, Silverlight and QuickTime are not going away anytime soon, and that complete support for HTML5 among high profile commercial broadcast and corporate sites has little perceptible momentum. CNET and Mefeedia are pointing to videos encoded into H.264 format for playback in Flash, and calling that newsworthy proof of HTML5’s acceptance. That’s a bunch of crap in any language.
If Mefeedia really wanted to portray the reality of HTML5 acceptance, they would disclose the percentage of Ogg Theora and WebM encoded video in their findings, including how much of the WebM was produced by Google property YouTube, who can hardly be called a neutral indicator. I’m guessing that CNET wouldn’t find those results quite so newsworthy.