Streaming Video Trends by Jan Ozer

5/24/2010 3:50:51 PM
By PSN Staff

I gave a talk at StreamingMedia East in New York City earlier this month to a group of business executives assembled by California-based webcasting solution provider MediaPlatform. I was told that the execs were interested in "future directions of online video," so I put together a talk entitled "In Your Chair, I'd Be Asking Myself ..." The talk covered three main topics that correspond precisely with the major headings in this article.

Is My Video Competitive?


No sense looking ahead if your current video offering is behind the times. I had collected some streaming statistics for a video production seminar I taught earlier that day, so I shared them with the execs and present them below. Briefly, to collect the statistics, I captured streaming video files from 25 media sites like ABC, CBS and ESPN, and 25 major corporate sites like Nike, GE, Toyota and McDonalds. Then I analyzed the files and divided them into three categories -- conservative, mid-range and aggressive - by screen resolution.

As you can see, the days of 320x240 video at 300 kbps (kilobits per second) are long past. CNN streams most of their videos at 640x360 @ 800 kbps, while prime time TV replays are often at 848x480 @ 1.2 mbps (megabits per second). Apple streamed their largest iPad promotional video at 848x480 at a startling 2.7 mbps.

This tells me two things. First, these sites know that most of their viewers have sufficient download bandwidth to stream and play these videos in real time - obviously, they wouldn't use these configurations if their video consistently stopped to rebuffer when played by their target viewers. Second, in terms of viewer perception, this is the quality to which most Internet viewers are happily becoming accustomed.

If you haven't updated your streaming video configuration in a year or two (or three or four), you can almost certainly deploy a larger resolution, higher quality stream that your viewers can stream and play in real time. As with living room television sets, bigger is almost always better when it comes to streaming video - otherwise, the networks wouldn't have boosted the quality of their streaming video.

Like it or not, you're subtly judged by the size and quality of your streaming video, not only vis a vis the ESPNs of the world, but particularly by comparison with your competition, and other videos your target viewers frequently play. If you're still using that 320x240 @ 300 kbps configuration, you're almost certainly hurting the perception of your organization in the marketplace.

Adaptive Streaming

The next point related to a technology called adaptive streaming, which is offered by a variety of vendors, including Adobe (called Dynamic Streaming), Microsoft (called Smooth Streaming) and Apple for its iDevices (called HTTP Live Streaming). Briefly, adaptive streaming technologies encode your source video files into multiple versions at a variety of resolutions and data rates. Then, the technology distributes the optimal encoded file for your viewer's connection speed and playback device, and continually monitors connection speed during the broadcast to ensure that video playback is sustained.

If streaming throughput drops during the transmission for some reason, a lower bitrate file is sent, sustaining the broadcast. If throughput later increases, a higher bitrate file will be sent. All this switching is totally transparent to the viewer, who may notice some differences in quality, but the stream will go on. As an example, Major League Baseball encodes their games into 11 different versions for their subscription video service, using multiple systems to adaptively stream to their computer-based and mobile customers.

If you're casually serving a few video files on your site, adaptive streaming is definitely overkill. But if the quality of your streaming video is critical to your training, marketing or other activities, you should consider adaptive streaming, particularly if you're starting to think about HD delivery, or distributing video to mobile phones. More on mobile in a second.

Am I Reaching all Relevant Target Viewers?

The second question asked "Am I Reaching all Relevant Target Viewers?," and involved support for mobile transmission in general and the iPad in particular. Regarding mobile, early in 2009, The Nielson Company reported that the number of mobile video subscribers watching video on a cellphone had increased by 52% between the first quarter of 2008 and the first quarter of 2009.

In the June, 2009 update to their Visual Networking Index Forecast and Methodologyreport for 2008-2013, Cisco reported that Internet video now accounted more than one-third of all consumer Internet traffic, and that video on demand traffic will grow at a 53 percent compound annual growth rate beteween 2008 and 2103. Cisco further predicted that mobile traffic will double every year through 2013, increasing by 66x between 2008 and 2013, and that over 60% of that mobile traffic will be video by 2013.

These statistics simply confirm what we already know - that the iPhone and its smart phone competitors have had a profound effect on how we watch video. Where a mobile feed for an event like Wimbledon or the Masters was unheard of even three years ago, now the lack of a mobile access to high profile events is almost unthinkable.

Certainly, B2C companies need to be more aggressive with offering mobile video than B2B enterprises, since consumer news and entertainment are the videos most accessed by early adapters. For all organizations, however, if video is mission critical to your marketing, sales, communication or other enterprise pursuits, and you haven't already started thinking about a mobile strategy, now is the time.

Again, if you do decide to deliver to mobile devices, adaptive streaming should be part of the plan. That's because mobile connectivity can vary all over the map, from below 100 kbps to well over 2 mbps, especially if the viewer is ... well, mobile.

An example would be an iPhone viewer who starts watching a news video in the office, connected via Wi-Fi, then walks outside for lunch, where the connection switches to 3G and bandwidth drops significantly. Apple's HTTP Live Streaming adaptive technology can deliver a high quality stream for the Wi-Fi connection, then drop down to lower quality for 3G, providing optimal service for all connection types. If you don't deliver to these mobile customers adaptively, the alternative would be a lowest common denominator file that looks ugly everywhere, or a high bit rate file that only plays smoothly on the fastest connections. Adaptive is clearly the way to go.

As a caution, however, be aware that delivering to mobile devices at this point is a very frustrating pursuit, where you have to approach each major platform -- Apple, Google Android, Blackberry and Palm - as separate devices accessed by a different technology strategy. Fortunately, if you successfully deliver to the Apple iDevice and Android platforms, you've collected a significant percentage of potential viewers, and this selective strategy is the one used by most web sites serving mobile viewers today.

What about the iPad?


The iPad has a distinctly flavor of the month feel, but with 1,000,000 units sold in the first month, it may just catch on. Again, B2C companies will likely need to distribute to the iPad sooner than B2C vendors, but the iPad is a platform all web producers need to be thinking about. Fortunately, you can support the iPad multiple ways.

For example, you can create an HTML5 compatible web page or an iPad specific app your viewers can download in the Apple App Store. If you have an existing iPhone mobile site or application, you can direct all iPad traffic to the iPhone site, and advise iPad owners to download your existing iPhone app, which will work on the iPad. Obviously, video produced for the iPhone won't look that great on the iPad, but it will be functional.

Or, as discussed further below, you can use an Online Video Provider to provide player technology that works with both desktop computers an the iPad.

Now onto the third question, which was:

Am I Investing Wisely?


Not your IRA, silly, your investments in video-related technologies. More specifically, if your site is currently Flash or Silverlight-based, should you pull the plug on further advancements relating to those technologies and start switching over to HTML5, the next generation web markup language that features a video "tag" that eliminates the need for plug-ins like Flash or Silverlight (for more on HTML, see The Future of Web Video, Part 1).

Very interesting question. At one level, HTML5 seemingly has the momentum of a freight train, especially after Apple's iPad announcement, and Google's long anticipated open sourcing of the VP8 codec.

On the other hand, HTML5 clearly isn't ready for prime time. Depending upon which stats you use, the current installed base of HTML5 compatible browsers is between 40-50%, clearly too low to abandon your existing plug-in based technology. Even after Google's VP8 announcement, there still isn't one single HTML5 compatible codec, as neither Microsoft or Apple has committed to including the VP8 codec in their browsers.

In addition, though most sites don't care about digital rights management, many three-letter networks that pump out millions of streams a day need to protect their content, and DRM doesn't yet exist in HTML5. Neither does live streaming, or adaptive streaming, other than Apple's HTTP Live Streaming, which works only with Apple devices and the most recent Macintosh desktops running Snow Leopard.

Furthermore, converting over to HTML5 means rewriting all Flash or Silverlight-based applications into Javascript. For a simple video player, that's no big deal. But if you've integrated features like social media links, database front ends, rich internet applications and shopping cart functionality into your web site via Flash or Silverlight, the conversion costs could be quite high.

Conversely, proponents of HTML5 haven't really established any benefits of HTML5 that don't already exist in Flash or Silverlight. Sure, there have been some interesting demos, but no live sites that make web developers (or their bosses) point and say - "man, I really want that on my site and I can't get it with Flash." So beyond Teva wearing tree-huggers that want to abandon Flash or Silverlight because they're proprietary and "the web should be free," it doesn't feel like there's a cohesive, mainstream reason for web sites to move to HTML5.

All that said, the web is evolving towards HTML5, and sometime in the next two to three years, or maybe four or five, most sites will be cutting over to HTML5 and abandoning their plug-in based "legacy" technologies. So, what's the mean for you?

First, between the Apple iPad, and Google VP8 announcements, HTML5 has become  mainstream. Your boss, who probably doesn't know the difference between cascading style sheets and Egyptian cotton sheets, may actually ask why you're not cutting over to HTML5 this week. Feel free to use the browser statistics, feature deficits or conversion costs above as justification.

Looking forward, though, if you're adding major features to your website, or asking for serious streaming video-related budget dollars, you do need to monitor advancements in HTML5 technology and the penetration of fully HTML5 compatible browsers. In particular, 2011 will be a very significant year for HTML5 as Microsoft ships Internet Explorer 9 (IE9).

Why will IE9 be so significant? Because by far, the largest chunk of existing non-HTML5 compatible browsers are earlier versions of Internet Explorer, with IE6, which first shipped in 2001, holding on to a mind-boggling 17% share. Many of these older computers won't be able to use IE9, since it won't support Windows XP, but most other IE users will likely upgrade, dramatically increasing the installed base of HTML5 compatible browsers.

Long story short, by the end of 2011, the installed base of HTML5 compatible browsers will likely be well over 60 - 70%, with most HTML5-incompatible desktops too old and slow for many producers to care about. This will eliminate one of the major objections to implementing HTML5. By then, HTML5-based technologies for DRM and adaptive and live streaming should also be available, eliminating these objections. Who knows, maybe the browser vendors who formulate the HTML5 spec will agree on a single codec.

Bottom line is that if you're considering video related development projects that won't come to fruition until mid to late 2011, you better figure HTML5 into the analysis, because by then, it should be really hitting its stride.

DOY or OVP

The second question in the Am I investing Wisely category involved those hosting their own streaming server and funding their own streaming related development efforts. Certainly, if that's your core business, you need to continue as is. But if you sell shoes, umbrellas or consulting services, and started hosting your own streaming video because creating a player was so simple back in the day, you need to understand that a simple player isn't so simple anymore.

Just look at the technologies discussed in this article, adaptive streaming, mobile and iPad support and HTML5. Developing those on your own could cost literally hundreds of thousands of dollars. Perhaps it's time to consider a third party online video platform (OVP) like Brightcove or Kaltura. These services can host your videos, provide an HTML5 compatible player and distribution to mobile devices, and via agreements with content delivery networks like Akamai, deliver your videos much more smoothly to your target viewers.

For example, see Create an iPad-Compatible Video Site in 30 seconds or Less for details on how Sorenson Media's OVP enables iPad compatibility via their standard player. Sound interesting? For more background on OVPs, check out What You Need to Know About Online Video Platforms.

Jan Ozer has written more than 10 books related to the video world. He is an AVT Advisor and a frequent industry speaker.

Share This Post