HTML5: What You Need to Know
One of the highest profile and most divisive issues in the streaming video marketplace today revolves around HTML5. In this article, I'll describe HTML5, discuss its pros and cons, detail how it's being adopted (or, more accurately, not adopted) in the marketplace, and conclude by telling you when you need to start thinking about it for your own website.
HTML stands for hypertext markup language, and HTML is the primary language used to produce most web sites. HTML is a standard set by the World Wide Web Consortium (WC3), a group which includes browser developers and other interested technology companies. HTML5 is the latest iteration of HTML to be introduced by the WC3.
What's particularly interesting about HTML5 is that it contains a video tag that enables browsers to natively play back video within the page, obviating the need for plug-ins like Flash, Silverlight, and QuickTime. HTML5 was quietly progressing and flying below the radar until Apple introduced the iPad without support for Flash, with video playback accomplished via HTML5. This created a firestorm of interest, and the inevitable Flash vs. HTML5 arguments.
HTML5—Sounds Great, but there's a Big Catch
At its highest level, HTML5 sounds great. Simpler is always better, and if you don't need a plug-in to play a video file, that's one less item for potential viewers to download. The problem is that—at least today—only around 50 percent of available browsers support HTML5, and more importantly, the W3C hasn't specified one codec that must play in all browsers.
For example, suppose you bought Apple's argument that HTML5 was better, dropped your current plug-in-based technology, and produced your website in an HTML5-compatible format that played on the iPad. To accomplish this, you'd have to encode your video in H.264 format, which is the only format that plays on the iPad.
Sounds good, until you realize that less than ten percent of those browsing to your website from their computers can play the file, since only Apple Safari and Google Chrome - both around 5 percent market penetration - can play HTML5 video in H.264 format. Those visiting your site via Mozilla Firefox could not play the video, since Firefox can't play H.264-encoded video. This is particularly significant since Firefox - at 23 percent market share - is by far the most popular HTML5-compatible browser.
To play in Firefox, and the Opera browser, you'd have to encode your video into the Ogg Theora format, which is much lower quality than H.264 and isn't supported in any of the more popular encoding programs like Adobe Media Encoder, Apple Compressor, Sorenson Squeeze, or Telestream Episode. This means that you'll likely have to use a command line encoding tool to achieve the best results.
Once you produce in H.264 and Ogg format, you still only have—at most—40 percent or so of potential viewers that can play your files. To satisfy the rest, you still have to make your video available using your plug-in based technology, which is typically Flash using the VP6 codec. Operationally, code on your web page would query the browser as to its capabilities—if it was HTML5 compatible, it would send the video encoded in the proper format, if not, it would revert to the plug-in.
Google's WebM technology, which is currently supported in Opera and will be supported in upcoming versions of the Firefox and Chrome browsers, does little to break the logjam, since Apple won't support it, and Internet Explorer 9 will only support it if already installed on the system. Of course, until WebM support becomes pervasive, you'll still have to encode in H.264 for Safari, Chrome, and Internet Explorer 9; Ogg for older versions of Firefox and Opera; WebM for newer versions of Firefox, Opera and Chrome; and potentially VP6 for Flash (or Windows Media for Silverlight).
By this point, you have to be asking yourself if the extra work is really worth it. This turns out to be a very good question.
Ring Around the Collar
Product marketers create "problems" that their products uniquely solve as a way to market their products. The "ring around the collar" that laundry detergent Wisk removed is a prominent example from my youth (see this YouTube video for a look at a commercial from the 70s).
To help explain his company's decision to promote HTML5, Steve Jobs was famously quoted as calling Flash a "CPU hog" rife with "security holes." To determine the veracity of the CPU hog claim, I ran some tests comparing the CPU cycles required to play back Flash vs. HTML5 on Windows and Mac notebooks. On Windows, Flash was much more efficient.
On the Mac, Flash was more efficient than Chrome, but less efficient than Safari running HTML5, primarily because Safari could use the chip on the graphics card (GPU) to accelerate video playback, a function that Flash couldn't access at that time. Later, Apple opened the GPU programming interface to Adobe, and the disparity in playback performance dropped by 73 percent.
Specifically, Safari playing a file via HTML5 required 13 percent of CPU resources on a MacBook Pro, while Flash-based playback using the same browser at 22 percent. While still a significant difference, I'm not sure that 9 percent warrants the CPU hog label, particularly since Google Chrome required 40 percent of CPU to playback video via HTML5, but only 34 percent for Flash.
More importantly, in a recent survey for StreamingMedia.com, which polled actual users of Flash, overall satisfaction was quite high, with less than 2 percent of respondents rating Flash as poor for performance, stability, security or end user satisfaction. Overall satisfaction with Silverlight was even higher, though the sample was much smaller.
Though the survey revealed that many respondents planned to migrate to HTML5, it also revealed significant concerns about HTML5 as shown in the following table.
In particular, the respondents were concerned about the browser and codec issues raised above, as well as the lack of a proven adaptive streaming facility or digital rights management, or the need to rewrite all Flash or Silverlight-based applications to Javascript, another HTML5 "gotcha."
The survey was published in June, and to get a read on current HTML5 adoption I visited 46 prominent media, B2C and B2B sites using the HTML5-compatible Opera Browser and Safari browsers on an HP workstation. Other than sites with a clear agenda, like YouTube and Wikipedia, none of the sites used HTML5 as the primary delivery mechanism to my computer, though 17 of the sites could deliver video to the iPad, presumably via HTML5. More specifically, virtually all media companies that I visited supported the iPad, with much fewer B2C companies and very few B2B companies.
These results shouldn't be at all surprising. Supporting HTML5 expands a web site's encoding chores and forces them to re-write their Flash or Silverlight-based applications in Javascript. At the same time, these sites lose their ability to protect their content via digital rights management, or deploy via adaptive streaming, which improves the overall viewing experience. Sites that are advertising supported lose access to some advertising networks, though this will likely resolve more quickly than some other issues.
What's the benefit of HTML5? Assuming satisfaction with Flash and Silverlight is high, which our survey established, there really is none.
When Does it Matter to You?
At some point in the future, HTML5 (or later) will become so established that major video distribution sites like the three and four letter networks and user-generated video sites will stop supporting Flash (or Silverlight) playback. Flash (or Silverlight) will become less important to viewers, so fewer will install the plug-ins. At that point, sites that don't support HTML5 will start losing viewers.
Until that point, which is clearly well off in the distance, the burdens of supporting HTML5 outweigh the benefits. If you're a media or B2B company, iPad support should be on your radar screen, but general support for HTML5 on your primary website doesn't need to be a priority.
Jan Ozer's article first appeared on OnlineVideo.net