The Player's the Thing...But So Is The Codec, Format, and Protocol

Article Featured Image
Article Featured Image

Glenn S. Pedersen, senior business developer at TV 2 Sumo in Norway, converted from Windows Media to Silverlight in late 2008. Pedersen says the response from TV 2’s viewers was "very positive. We have actually seen a drop in support cases at our help desk after users started to install Silverlight."

After chatting with Pedersen, I checked Silverlight’s penetration in Norway at RIA Stats, which reported that it was at 63% for all computers, 70% for computers running Windows Vista, 49% for Windows XP, and 51% for Macs running OS X—no samples were available for Windows 7. This is pretty persuasive evidence that, given compelling content, web viewers won’t hesitate to download the Silverlight player.

All that said, while player penetration isn’t critical to all content producers, it is to some. For example, Goldstein says that it is against MTV policy to require a client install, which is why MTV eschewed Silverlight and Move Network’s streaming technology. Inzerillo added that forcing viewers to download a plug-in was a big negative for MLB’s adoption of Silverlight, particularly because many subscribers were watching on corporate desktops, and they could run into issues attempting to download a new player.

Another major concern for both organizations was the investment in Flash development and programming skills. Goldstein says that MTV had invested heavily in Flash-based infrastructure surrounding the video player, including advertising, reporting, and social media modules and that moving to Silverlight might force them to port these modules as well.

Inzerillo felt that the Flash development environment was more robust and mature. "If you’re writing in Flex," he says, "you can Google any problem and find lots of answers. In contrast, Silverlight is still an island, and every problem is a new problem."

On the other hand, if you’re converting from Windows Media, at least the first issue isn’t relevant because you likely have little investment in Flash technology. In addition, the fact that your libraries of Windows Media content can play in Silverlight and not in Flash may go a long way in pushing you toward Silverlight. Not surprisingly, one of the reasons TV 2 Sumo chose Silverlight was because "all our existing content was in VC-1," Pedersen says. "Silverlight can play back [VC-1] without [a] problem."

Of course, the reverse argument is also true: If you’ve been distributing Flash streams for the past several years, a substantial portion of your library is likely in VP6 format, which can play in Flash but not in Silverlight.

So that’s where we stand today. Both Flash and Silverlight can get the job done for most streaming applications, with Flash enjoying near ubiquity and Silverlight less than 40% penetration. Conversely, Silverlight has the current advantage of integral HTTP delivery, although Adobe will likely catch up sometime in 2010. Your choice of technology should depend on the uniqueness of the content you’re delivering, your current investment in either Flash or .NET, and the format of your existing libraries of content.

Which Codec?
Now that Silverlight supports H.264, codec selection isn’t a decision factor between Flash and Silverlight, but you can still choose between H.264 and VP6 on the Flash platform, and VC-1/WMV and H.264 in Silverlight. Note that H.264 is currently the only codec supported by Apple’s HTTP Live Streaming. Regarding Flash, both the MLB and MTV sites have standardized on H.264 for new content, primarily because of quality but also because of configurability. This fact was particularly important to Inzerillo, who customizes encoding parameters based on the stadium, camera location, and even whether a game is played during the day or at night.

For Silverlight, TV 2’s Pedersen commented, "From our experience, H.264 is still too heavy for ordinary computers to decode if you go up in bitrate." Note that in my tests, which are documented on StreamingMedia.com (www.streamingmedia.com/article.asp?id=10776&page=1&c=3), I found H.264 easier to play back than Silverlight on all the computers I tested. If you’re choosing between VP6 and H.264 or Windows Media and H.264, you should run your own playback tests before making a decision.

Of course, the fox in the codec woodpile is Google’s acquisition of On2, which most people assume was primarily to acquire the VP7 and VP8 codecs. Industry reactions ranged from breathless wonder (Silicon Alley Insider’s "Google’s On2 Acquisition Could Revolutionize the Video Industry") to a more practical analysis (Dan Rayburn’s "Google’s Acquisition of On2 Not a Big Deal, Here’s Why"). But the reality is that no one knows the real reasons for the purchase except Google itself, and Google isn’t talking.

In my view, the most likely scenario is for Google to donate VP8 to the open source community in support of HTML5, while converting YouTube over to VP8 to avoid H.264 licensing costs and using VP8 (if possible) in video conferencing applications. A more interesting strategy would be for YouTube to stop accepting video in any format other than VP8 and then licensing the codec to the developers of video editors and other encoding tools—I think most encoding tool developers would pony up pretty quickly, don’t you?

Whatever Google decides to do with VP7 and VP8, should it impact your short-term planning strategies? Unless you’re MPEG LA, which administrates policies for the H.264 patent group, probably not.

That’s because VP8 is just a codec, which really needs a server/player/design environment infrastructure to be useful. If VP8 becomes open source, both Microsoft and Adobe can simply support it in their respective players and we can all move on with our lives. The contrary scenario (which assumes that Google will keep VP8 private and launch development and design tools to compete with Flash, Expression Studio, and associated tools) is highly unlikely given the investment and slow growth of those business models. Plus, the design tool business is different from any business Google has attempted to date. Google may attempt to license the codec as On2 did instead of going open source, but it’s unlikely that it’ll try to compete with Flash or Silverlight.

Google’s acquisition of On2 is the only event of the past 24 months that should make the patent group feel that it doesn’t have free reign on the licensing structure for distributing H.264 over the internet or to iPods and other devices. H.264 is the only technology that producers can currently use to distribute to the relevant four screens (computer, mobile, set-top box, device), which gives the patent group incredible leverage.

On the other hand, four screens sounds good, but set-top boxes are really a separate market (at least for now), and computer playback accounts for at least 85% of the other three screens.

According to Nielsen, more than 76% of the unique viewers watching streaming video in the U.S. visit YouTube. Google could convert them all to VP8 virtually overnight by encoding YouTube videos with VP8 and making a quick call to Adobe to convince them to add VP8 playback to the Flash Player.

We’ll probably never know whether On2’s acquisition impacted the proposed royalty. But the MPEG LA was scheduled to release details of the royalty structure by the end of January, after this publication went to press.

Multiple Streams?
Whether you choose Flash or Silverlight via RTMP or HTTP, everyone I spoke with for this article recommended distributing at multiple bitrates.

For example, Level 3’s Middleton reported that for the French Open, which was broadcast at multiple bitrates via Smooth Streaming, "dwell times" increased to 74 minutes, which was much longer than previous years.

Streaming Covers
Free
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

The Algorithm Series: QUIC Ways to Stream

Will 2022 be the year UDP finally shows its streaming mettle? The Quick UDP Internet Connections (QUIC) protocol might make the difference. First, though, OTT platforms need to make technical decisions about HTTP/3 that could further fragment the market.