Introduction Recently the BBC announcement that it was developing it's own video codec: Dirac. Named after the British physicist Paul Dirac, this project has raised more than a few eyebrows, as Yahoo News noted; "The BBC is quietly preparing a challenge to Microsoft and other companies jostling to reap revenues from video streams...The sums at stake are potentially huge because the software industry insists on payment per viewer, per hour of encoded content. This contrasts with TV technology, for which viewers and broadcasters alike make a one-off royalties payment when they buy their equipment." plugincinema.com is pleased to feature an interview with Thomas Davies, Senior R&D Engineer at the BBC who devised the Dirac algorithm.
The Interview  PC:From an audience point of view, how will they interact with Dirac? Will it be supported by any media players (e.g. Real, Microsoft Windows Media Player, Mplayer etc?)
TD:Audiences' experience of Dirac will primarily be through media players, I would think, although there's nothing to stop the technology being adopted in other arenas like mobile phones or PVRs. As far as players go, we'll be submitting a patch to Mplayer to allow it to play Dirac pretty soon. A DirectShow filter for Windows Media player is also on our route map, and we've had several volunteers to do this.
Dirac is still in a fairly early alpha stage, and we've got a long way to go. We've got to speed up the decoder and the encoder quite a bit, for example, although we can already play back CIF at reasonable speed. But people shouldn't expect Dirac to be 'there' for a little while yet - we're aiming to go to beta by Autumn next year.
PC:Will Dirac cover both video and audio streaming?
TD:Not directly - we're not developing an audio codec under the Dirac banner. There are codecs we can use already out there, like Ogg Vorbis and MPEG AAC.
We will need to sort out how Dirac is delivered with accompanying audio, though, since what people will download will be a multimedia stream. This means specifying a wrapping format, which will say how all the parts of a multimedia stream will be bundled together - it handles things like synchonization of the video and audio. It?s important to say that there can be lots of wrapping formats, and anyone is free to wrap Dirac in any way they like. We'll have a bitstream spec which they can take and do anything with they want to. For ourselves, we're going to use the Multimedia Exchange Format (MXF), which is an Open Standard which is widely used in broadcasting. But the Ogg folks might want to use Ogg and that would be good, too.
PC:In terms of technology development, how much new territory is this development project exploring?
TD:Well, Dirac is a mixture. Since we're intending Dirac to be license-free, we had two options: use old techniques that are out of patent, or invent new ones. When I started developing Dirac, I used tried and techniques, like wavelets, which weren't in standards at the time, and tried to develop them. And that's what we'll continue to do as the algorithm develops. So we've tried to build on some pretty well-understood technology, and also tried to do some new things with it. We're patenting the new stuff, quite a bit of which hasn't got into the software yet. The license means that these patents are licensed for free within the Dirac software.
 There's an obvious risk that we'll be wrong about some of the older techniques, and they'll turn out to be patented. So we're reviewing the technology as we go along, and it's one reason why the algorithm could still change. The other reason is that we've had quite a few new ideas that we want to get in!
PC:Have you had any interest from existing online media corporations - RealNetworks, Microsoft, Apple QuickTime, DivX Networks etc?
TD:We have had interest from some other technology organisations, but discussions are at a preliminary stage.
PC:Are the plans to use the same codec both for streaming and (assuming the BBC still goes ahead with the plans) for distribution of works using file trading?
TD:If your question is, is Dirac going to be used for these, then the answer is that I personally hope so (obviously), but we don't know. In releasing Dirac, I don't think the BBC is saying, "We're going to use it for all these thing in the future, and nothing else". Dirac is another option, which has some advantages, and adds an element of choice for the BBC and other broadcasters. It's based on different technology, for a start, and so enriches the compression ecosytem, whatever happens to the system itself. What the BBC will use will depend on a whole set of technological and non-technological criteria. And the beauty of software is that what you decide to use can change relatively easily.
As for the question of whether the choice of codec for content-sharing is tied to that used for streaming, I don't think it is.
PC:In your opinion, does working at a publicly funded organization have a different approach in terms of both the development and the licensing. How is it different to that of a commercial corporation?
TD:That's a hard question! I think the BBC has always had a very strong commitment to Open Standards, which have actually benefited the whole broadcast industry as well as other PSBs. But I don't think such a commitment is necessarily confined to publicly-funded organisations. There are many manufacturers and technology companies that share it. I think it's more a question of degree - the BBC is particularly concerned about public access and in the development of TV to date that's implied Open Standards.
As for licensing, BBC R&D does do commercial licensing. There's always an argument about whether the public is best served by our licensing things and selling them, thus bringing more money in to make programmes with; or giving things away, and allowing the public to participate at lower cost. Essentially we consider each project on a case by case basis.
PC:Is Dirac related to MPEG-4 in any way, or is it all fresh stuff?
TD:MPEG-4 is a pretty broad church, so there are always going to be some similarities. But using wavelets instead of block transforms is a fundamental difference, and it has knock-on effects throughout the codec.
PC:Has the BBC considered, now it is developing online media software, joining a trade body such as the Internet Streaming Media Alliance?
TD:We haven't considered it to date. As Dirac matures, we'll be considering more how we should promote it and the use of Open Standards generally in internet broadcasting.
PC:In your FAQ you say that the license that Dirac is released under is Mozilla triple license (MPL), but also that, "...allows for relicensing under the GPL or the LGPL." Could you clarify this - does it mean that Dirac will only be under the MPL, but that others developing using the source could release their work as GPL?
TD:Yes, they could do that.
PC:Have you, or other members of the team done much open source (or similar) development prior to Dirac?
TD:BBC R&D has released some other code Open Source, for the MXF file formats for example. Our Lead Programmer, Anu Suraparaju, was active in those and has also been involved in other projects like MPlayer and transcode. Personally, this is new to me, though. It's been quite an experience - and great fun. It's a good feeling to know that someone is using or modding your software, or even studying it for a Masters thesis.
A big thanks to Thomas for his time in answering our questions.
Further Links BBC Dirac Page
http://www.bbc.co.uk/rd/projects/dirac/
Dirac FAQ Page
http://www.bbc.co.uk/rd/projects/dirac/documentation/faq.htm#1
Dirac Source Code Page
http://sourceforge.net/projects/dirac
Yahoo News Report (on Slashdot.org)
http://developers.slashdot.org/developers/04/08/13/128249.shtml |
