The Quest For Open Video Standards
Adobe Flash declining and the rise of mobile platforms without Flash, HTML5 on the rise, H264 forever royalty-free for non-commercial use, WebM released by Google… There have been many new developments in the web video industry in 2010.
Hope, concern and lack of visibility for the future, this is where we’re at. We offer you an overview of the situation, so that you can find your way around and understand what’s at stake in this big shift.
Overview of web video
In the early 2000s, heavy clients like Windows Media Player, Real Player and Quicktime were fighting for supremacy over video playing on PCs. When Flash Video technology arrived in 2005 video became a standard type of content for websites and flooded the web. In 2005 YouTube was born; by 2006, 97% of the Internet users had Flash installed.
In 2010, 5 billion videos were seen every day on the Internet. On average, each Internet user watches 150 videos a month, up to a total of 14 hours, compared to 2 hours in 2008 (source: Comscore). YouTube holds the record, serving an average 4.5 hours per user. Video on the Internet is becoming a common communication tool both for individuals and companies, carrying anything from video clips to full length movies.
Flash technology was so far the main solution for broadcasting video on the Internet. But in 2010, with the rise of HTML5 standard and especially the <video> tag, expectations about free and standard video technologies became higher. YouTube and Dailymotion took the plunge and now offer their content in HTML5.
Is the Internet ready to free itself from Flash technology and its royalties? What does migrating to open video imply, and at what cost?
Well established proprietary technologies
Video on the Internet requires the deployment of resources on both client and server-side. On both sides the most efficient and widespread technologies are proprietary.
First, the video must be encoded in a format the client can read: the codec. The widespread codec today is H264 (with 65% market-share), held by MPEG-LA consortium on behalf of beneficiary companies such as Apple and Microsoft. H264 is currently supported by Flash Player and natively by Chrome and Safari browsers.
On the server-side, the architecture has to be able to handle streaming, and if necessary video converting. The most widespread video server to date is Adobe Flash Media Server, which uses the proprietary RTMP protocol and features offered by the Adobe technology. Microsoft and Wowza are trying to place themselves as competitive alternatives to the leader.
H264 and Flash Media Server are technologically ahead of their free alternatives. For example, free codecs like Ogg Theora and WebM-VP8 need more bandwith for equal image quality. Open Source video servers don’t allow continuous adjustment to connection throughput, secured streaming, etc...
Proprietary technologies imply the payment of royalties. For codecs, using the H264 is free of charge when broadcasting videos for a non-commercial use. However, royalties are required for home playing/recording devices, encoding software, VOD, etc. As for Flash Media Server, it requires the purchase of a license a server.
Open Video promises
Several candidates are fighting for the part of “the” alternative technology to proprietary software. As for the codecs, Ogg Theora was born in the early 2000s, and Google announced the birth of WebM (a container for the VP8 codecs) in May 2010. On the server-side, the prominent solution is the one developed by the Kaltura Open Source community (Kaltura also has a commercial hosting and broadcasting video solutions offer).
The use of these codecs shouldn’t be subject to royalties, even if some manufacturers maintain uncertainty about the legality of WebM in relation to existing patents. Infrastructure costs, of course, remain.
Open video technologies promise a greater freedom in content processing. The streams are not encoded within a black box anymore, but through well-known standards. It makes the final front-end integration (for example, for the customization of the player design) but also the audiovisual stream procesing much easier. The Mozilla Foundation produced a technical demonstration where a WebM encoded video stream is sent to the browser WebGL API to play videos on the faces of a 3D rotating cube.
Another important fact about Open Video is that it’s based on standard transport protocols, starting with HTTP. It lets us hope for internet video rid of connection issues due to corporate and private firewalls.
A long way to go
- 1. Compatibility and unified standard: still experimental
- Open video isn’t mature enough to dethrone proprietary technologies yet. The first obstacle is agreeing on the technology to be used. Without this agreement, no standard can exist: Internet Explorer and Safari support H264 for the <video> tag, while Firefox, Chrome and Opera support or are on the way to support WebM. Firefox and Opera also support Ogg Theora. Every role-player seems to stand firm, which makes the change to a Flash-less world delicate.
- 2. Make up for the technological gap
- Though open video technologies have gone past the experimental stage, they remain young and suffer from technological backwardness compared to proprietary solutions. It’s particularly visible in the video stream size/quality ratio, the ability to broadcast DRM protected streams, to adapt video bit rates to the available bandwidth or to send only parts of a file. If these specifications are critical for your business, it’s not time to adopt open video yet.
- 3. Adapt infrastructures
- Current server-side infrastructures are mainly meant to broadcast video via Flash Media Server. In order to move to open software on the server side, you’ll have to turn to video server open technologies such as Red5 or Erly Video, with the related system administration costs. The capacity of these technologies to sustain growing workloads is still uncertain, with performances considered to be lower than proprietary solutions. Beyond the technical binds, it is also important to consider human resources available to led the migration, as it may be a decisive factor to choose between technologies.
- 4. Unclear savings
- Migrating to open video will save money on server licenses and video encoding royalties as well. On the other hand, migration has a cost, both from infrastructure and human resources standpoint, that might offset savings in the short run. Due to limited feedback and the lack of experts for massive implementations of Open Source solutions, it will be necessary to plan an important safety margin to set up the new environment. A first round of tests on a small scale will allow an assessment of the necessary workload to achieve your goals.
- 5. Don't be late
- All that said, the world will transition, and soon enough to open video standards. In a fast paced enviroment, if your business relies on delivering video to consumers, you do not want to be late. You should probably not commit yet to a single solution, but start planning for it now. Identify your migration cost, your lock-in barriers and prototype and test your systems for the day after.
Unlike the H264 that can incorporate DRM and still be read by Flash Player, WebM and Ogg Theora don’t feature any content protecting system. Besides, with download or http progressive download (aka “fake” streaming), data streams are saved to the disk for caching and can easily be copied. Some open source supporters claim the lack oh DRM is actually a feature, since a stream has to be open to be standard, letting any resource free to interact with it. Nevertheless this is currently the main obstacle to its adoption by the media industry, which ensures Flash several more years of hegemony.
It’s always complicated to compare codecs. Each one has different results depending on the video type: static vs. action scenes, plain colors vs. high contrast, etc. Another obstacle is that Ogg Theora continuously improves, but doesn’t have as much a standardized release schedule as its proprietary competitors. As a result, testers don’t always know how to get the latest recommended version and set the complex options required for a fair comparison. Therefore, it’s difficult to settle the issue; the common shared feeling is that at a high bit rate, codecs are all acceptable. On the other hand, when looking for the best size/quality ratio, H264 remains at the top, ahead of WebM/VP8. Currently open codecs display a quality usually acceptable for general public though.
HTTP Live by Apple
In 2010, Apple created HTTP Live Streaming, an Open Source protocol based on HTTP, meant to serve as an alternative to Adobe Flash RTMP or Microsoft Smooth Streaming. The main benefit is the ability to broadcast video with a casual web server. The technology is based on existing standards and lets the client receive a video file in the form of a downloadable segment list. The client then selects the segments that fit best with their connection speed. At the present time, the only software to support it on the client side are Quicktime X and iOS (for Mac, iPhone and iPad), and it appears as an umpteenth technology to support in order to broadcast on Apple devices. However it could gain popularity if it were eventually accepted as an Internet standard by the IETF.
Kaltura is an Israeli start-up created in 2006, offering an Open Source platform for multimedia content based collaboration. The company also offers a hosting platform and video content management paying services. Kaltura brings together the biggest open video community to date and manages many projects meant to ease its integration: playing/recording/editing videos with client-side widgets, bricks for the biggest Open Source platforms such as Drupal or Joomla, but also libraries for PHP, Java and other languages, and finally an Open Source video server.