Beyond Standards: Moving Past MPEG to Create a Video-Friendly Web
Cisco and Mozilla have gone on to join the Alliance for Open Media (AOM), along with other industry heavyweights such as Amazon, Google, Intel, Microsoft, Netflix, and others. The AOM, which formed in response to the H.265 licensing morass, is working on AV1, a royalty-free codec that its executive director, Gabe Frost, says will contain some of the best ideas and features of both Thor and Daala, in addition to Google’s VP9.
Other companies, such as RealNetworks (which is currently offering its RealMedia HD codec) or even Google (with VP9) are now forced to use comparisons to H.265 as the standard by which they are judged. But the issue isn’t just quality. As the Cisco blog post shows, H.265’s licensing may preclude use of the standards-based codec.
All Things Mozilla
The move by Cisco toward royalty-free codecs brought to mind a similar situation, just a few years ago, when Mozilla attempted to eradicate H.264 use in the browser.
Mozilla is repeating history, albeit with the IETF rather than with a private group like it did around 2009. At that point, Mozilla backed the Open Video Alliance (OVA), which intended to push alternate codecs into the marketplace, putting its weight behind Mozilla’s efforts to see Theora emerge as an alternative standard. That was back when Firefox, the browser that Mozilla oversees, ruled most of the alternate browsing battles.
The OVA held an Open Video Conference from 2009 to 2011. Heavy hitters attended the conference, including representatives from Adobe, Blip.tv, Harvard, and Kaltura. The latter was a major player in the move to offer open source video alternatives to the then-uncertain licensing issues surrounding H.264.
One of the crowning achievements of the OVA was the move to have Wikipedia videos use an open source video codec.
“Video on Wikipedia is fully collaborative and free for anyone to view, modify, and share,” Josh Levy, OVA program assistant, told me in a March 2010 email.
“It is made possible by advancements in technology on the web,” wrote Levy, “notably HTML5 standards and the open video format Theora, which is patent-free and openly licensed.”
The OVA used a presentation at SXSW 2010 on the Let’s Get Video on Wikipedia project to tout the concept. Ben Moskowitz of the OVA and Leah Belsky of Kaltura even launched a “campaign headquarters” web domain (videoonwikipedia.org, now defunct) as part of the presentation, in keeping with the idea that the OVA was as much a political movement as an attempt to free us from the chains of licensed video codecs.
That campaign approach was further emphasized by the backing of the Participatory Culture Foundation, makers of the open source Miro video and audio player, as well as the sister company of the not-for-profit Participatory Politics Foundation and its OpenCongress spinoff.
In its short life, the Open Video Conference attracted significant political interest—including speakers such as Amelia Andersdotter, a member of the European Parliament and part of the Swedish Pirate Party, and numerous law professors—but the technical advancements for the streaming industry to work with just weren’t there.
As a result, the Open Video Conference held its last sessions in 2011. In addition, the OVA has long since stopped blogging, instead pointing OpenVideo Alliance.org domain visitors to the archives of the Open Video Conference.
Mozilla, for its part, now points developers toward a number of options for video and audio tag elements used in website and app development.
One of these is a set of open source audio codecs known as Vorbis and Opus, which are championed by the same community that brought us Ogg Theora. But the video codec used alongside these open source audio codecs, in a modified Matroska container format, is Google’s VP8 or VP9 codec. Known collectively as WebM, this combination is not yet supported on new browsers such as Microsoft’s Edge.
Mozilla states that it supports MP4 H.264 “in some cases, but only when a third-party decoder is available” for playback.
Yes, you read that right: Mozilla’s stance is that it requires an add-on to its browser to play standards-based H.264 in an MP4 container format. On the other hand, Mozilla does support Media Source Extensions, a JavaScript approach to using the HTMLMediaElement to deliver content in MP4 and WebM containers, as long as the underlying platform has the proper codec.
The Room and the Elephant
We’ve made it all the way through this article without discussing a company that has single-handedly avoided bringing its innovations into the MPEG or ITU folds: Apple.
That’s not entirely true, since Apple was part of the MPEG-DASH (Dynamic Adaptive Streaming over HTTP) specification working group. Apple’s contribution to the cause was to convince the working group to include an MPEG-2 Transport Stream (M2TS) implementation of MPEG-DASH, on the outside chance that Apple wanted to forgo its own HTTP Live Streaming (HLS) technology for adaptive bitrate delivery via web browsers. (Thus far, it hasn’t.)
In the meantime, Apple has continued to submit “drafts” of its HLS specification to the IETF, in the form of what the industry affectionately calls the Pantos spec. HLS is an example of a single company innovating, using its technology in its own enclosed ecosystem so consistently that it forced rivals—Google, Microsoft, even Mozilla—to support HLS across mobile and desktop browsers.
This brings us to the two other building blocks: transport and caching.
If you think back to our tug-of-war scenario, there’s always an unasked question in the air if the innovators win: What about backward compatibility?
In the case of HLS, the questions of backward compatibility and universal adoption have come up over time. The innovation was less about the codec or even the transport—Apple’s HLS uses M2TS, a common encapsulated video delivery package that broadcasters have used for decades—but rather about the use of “plain vanilla” web servers to deliver video streams.
Caching In
The advent of HTTP-based streaming, with its hundreds of thousands of video file segments served up sequentially, has single-handedly pushed the need for caching innovations.
To answer that call, the Streaming Video Alliance (SVA), which was founded in 2014, has announced approaches to what it calls an open caching standard several times over the past year. It did so first in April 2015, using guidelines from the Open Caching Working Group as best practices for caching.
In May 2016, the SVA announced another set of open-caching guidelines. “We’re making very good progress toward our goal of creating an industry-wide open caching standard,” said Jason Hofmann, Limelight Networks’ vice president of architecture and an SVA member.
Hofmann noted that the group has grown to more than 40 members in the 20 months since its November 2014 inception.
“We created the Open Caching Working Group and have already published a comprehensive set of Functional Requirements that were approved unanimously by the Members,” said Hofmann.
The size of the group, which is on par with the size of an MPEG Working Group, means decision making is slow. This may explain why it’s taken 13 months to move from the initial open-caching guiding principles document to an actual technical requirements document.
On the other hand, perhaps the SVA will ultimately move a bit quicker than formal standards bodies, to get to a point where open caching is a reality based on an ad hoc standard rather than a formal MPEG or ITU standard.
“We ultimately expect our work will result in a global best practice and standard for Open Caching systems in ISP networks,” said Hofmann. “This is a bold goal, but we have a movement of SVA members who share the deeply held belief that open caching will be an essential part of the internet infrastructure going forward. With the support of these like-minded members and given our progress to date, I’m quite confident we will reach the goal relatively soon.”
Summary
In the end, technologies alone are not enough to properly address the Netflix-ification of the internet we rely on today to deliver media. Several underlying assumptions surround the technology itself: backward compatibility, easy licensing, and less onerous licensing fees, as well as the chance for a technology to be universally adopted.
While those assumptions may most often be met with standards-based technologies, we’ve seen that innovations can come from single large companies (such as Apple) and even from industry consortiums themselves. Innovation can also come from small companies with a singular vision, although a small company seldom wins the tug-of-war against industry giants.
This article originally ran in the July/August 2016 issue of Streaming Media magazine as “Beyond Standards: Escape From MPEG.”
Related Articles
OTT companies innovate to differentiate their products from their competitors' products, but for the industry to reach its potential, they'll have to cooperate.
02 Oct 2017
Enough with the codec confusion. The online video space won't really thrive until industry standards are put in place and enforced by the government.
08 Jun 2016
The industry is turning away from plug-ins and embracing HTML5 everywhere. Here's how the vendor-independent streaming standard is gaining momentum.
29 Mar 2016
Just when Media Source Extensions and Encrypted Media Extensions are making HTML5-based video playback a reality, DASH royalties threaten to derail it.
27 Jul 2015
Companies and Suppliers Mentioned