BBC Wants Help With Dirac Codec

Number Ten Ox writes "According to The Register the BBC wants help to develop their open source video codec Dirac. '[Lead developer Dr. Thomas] Davies said the codec could live on anything from mobile phones to high-definition TVs but not before a lot of further work is completed. For one thing, Dirac doesn't currently work in real-time. Davies also reckons that the compression offered by the technology could be further optimised. The BBC is working on integrating the technology with its other systems, but the corporation would welcome more help in developing Dirac.' Sounds like something worth helping with."

Comments

BBC rules!

[orangeguru]

Compared to many other broadcoasters the BBC has a long and excellent record of producing great programms AND embracing the web/technology.

Certainly a good 'partner' to support ... compared to companies like Real ...

Re:BBC rules!

[matt_wilts]

Don't be so sure - if that was the case, they'd be streaming their current content using MPG or perhaps OGG. As it stands, they use Real!!

Matt

Re:BBC rules!

[Mike McTernan]

I think the only reason that the use Real is that the streams are more proprietary and harder to rip (for the novice in anycase), and it probably makes some copyright holders happier to let the BBC re-webcast certain content.

See here:

"What's the problem with Windows Media Player?

When the BBC began publishing audio and video content Real Media was the most secure form of streaming. Unfortunately Microsoft no longer supports Real content. Consequently, many of the later versions of the Windows Media Player will not play our clips. This may change in the future. NB: Some World Service clips are also streamed for the Windows Player. "

Re:BBC rules!

[0x461FAB0BD7D2]

What's wrong with the Real codec?

I've seen things compressed with RMVB which are on par with DivX and Xvid in terms of quality, but RMVB produces smaller filesizes.

Do not confuse the codec itself with the designated player. Real Alternative works too, without spyware, if that is what you're insinuating.

Re:BBC rules!

[junklight]

Indeed. In case you where wondering *why* they want to make a codec take a read of this:

http://eff.org/IP/BBC_CMSC_testimony.php

The Creative Archive is a really exciting venture and one of the projects that gives me small hope that the British Government may yet get the hang of copyright and online content

Re:BBC rules!

[LordK2002]

The Creative Archive is a really exciting venture and one of the projects that gives me small hope that the British Government may yet get the hang of copyright and online content.

The BBC almost certainly has got the hang of online content and copyright, but the BBC is not the British Government: it is an entirely independent organisation funded by the TV License (which is authorised and enforced by the government).

By contrast, the government is all too happy to jump onto Corporate America's IP bandwagon, with its Super-DMCA laws*, support for software patents and other such nonsense.

K

* In fact the Copyright and Regulated Relations Act 2002 passed in the UK makes it illegal to do anything that bypasses copy protection, not just traffic in "devices" as in the US. I guess marking a CD with a magic marker is now a criminal offense in the UK.

Re:BBC rules!

[junklight]

I work for another Establishment organisation and we and others are working on improving the situation. Having the BBC as a shining light really does help.

What am I missing?

[cslarson]

If they want to make an open source video codec, why don't they just support and help further develop the ogg video codec? Would the two codecs be so different that they are both needed?

Re:What am I missing?

[Ikkyu]

The Theora codec is a discrete consine transform, while dirac codec is a wavelet based. They are completely diffent ways of looking at video data and wavelett coding is showing promise as having higher compression rates and better quality.

What we really need is something that is scales with bandwidth, the more you receive the better your quality.

dirac vs. theora?

[crayz]

Have there been any comparisons? Do we really need two fully scalable open-source video codecs?

Also - the BBC is funded by the British government. When did they get a mandate to spend money developing video codecs. I don't have a problem with government-funded "arts" but this seems a bit beyond the normal scope of things

Re:dirac vs. theora?

[onion2k]

The BBC is funded by government, but thats where the relationship ends. The UK government has absolutely no say whatsoever in what the BBC spends its money on. If the BBC wants to develop video codecs then theres nothing the UK government can do about it. Thats one of the reasons the BBC news is able to remain impartial, and often reports on the UK government making a mess off things. See the Hutton report for details. :)

Re:dirac vs. theora?

[91degrees]

The BBC is not funded by the government. It's funded by the public through the licence fee. The government never gets to see it.

Re:dirac vs. theora?

[Anonymous Coward]

No the BBC is NOT funded by the UK Government.
The BBC has tax (i.e. the TV Licence Fee) raising powers of it's own - and is entirely independent of funding from government.
If the BBC *was* funded by government it wouldn't be considered trustworthy. It wouldn't be the "gold standard" of news reporting world wide that it is.

Re:dirac vs. theora?

[serviscope_minor]

Have there been any comparisons? Do we really need two fully scalable open-source video codecs?


Dirac is a next generation codec. It is also the only one using wavelets (like JPEG2000). Is there an argument for developing new codecs which compress better than current ones? Very much I'd say, unless you want all technological progress to stop here.

Also - the BBC is funded by the British government. When did they get a mandate to spend money developing video codecs.


They are a broadcasting organisation. Video codecs are very much part of broadcasting. They also did a lot of development on digital TV, which is soon going to replace all analogue TV by law in the UK. If they use this codec to put their archives up on the internet, then they certainly do have a good reason to do this development.

I don't have a problem with government-funded "arts" but this seems a bit beyond the normal scope of things

Is it? What about all that government funded science and tech research?

Re:dirac vs. theora?

[JohnGrahamCumming]

> Also - the BBC is funded by the British
> government. When did they get a mandate to spend
> money developing video codecs. I don't have a
> problem with government-funded "arts" but this
> seems a bit beyond the normal scope of things

Really? The BBC needs to stay up to date with technology in order to do the best job possible under its mandate. So that means that they are going to start out doing radio, spend money making television work the way they like it, then start promoting teletext (in the form of Ceefax), brand their own computer, and now they want to do the Internet their way (through an open codec).

It's worth reading their own history
for a perspective on just how much technical work the BBC has done since 1920. See also here.

John.

Re:dirac vs. theora?

[martinthebrit]

The BBC has a long history of R&D, based at Kingswood Warren in Surrey. Many important developments were made, under the funding of the BBC charter and through private industry. I'm sure the BBC's development of an open source video codec can only be good.

N.B I used to work for a broadcast equipment manufacturer, Snell & Wilcox, alongside many ex BBC engineers, and they employ some very good people.

Re:dirac vs. theora?

[provolt]

That has to be the funniest thing I've read in a long while. I think it's even funnier because it's moderated as "Informative".

For those who don't get the joke, read the wikipedia entry for the Hutton Report.

Re:dirac vs. theora?

[IWannaBeAnAC]

You are American, right?

There seems to be a cultural difference between the USA and the rest of the Western world, in that Americans are unable to conceive a government funded entity (directly funded or indirectly via 'license' fees) that is substantially free from Government influence. Possibly because there are apparantly no such entities in the USA. But in this matter, the USA is the exception rather than the rule, with respect to democratic governments.

Re:dirac vs. theora?

[aldoman]

Er no, becuase the government _can't_ stop the BBC from doing anything. They obviously have limitations like what frequency they can broadcast on.

Every 9 years (IIRC) the government reviews the BBC's progress and what funding method it should have.

Basically what I'm saying is the .gov.uk can't censor, change or stop the BBC from doing anything directly. They do not go to the government to approve TV shows, nor do they go to the gov to approve technology research.

This is in direct contradiction to social security in the US where the government controls it and could (probably) stop paying out tomorrow.

Re:dirac vs. theora?

[4of12]

Americans are unable to conceive a government funded entity (directly funded or indirectly via 'license' fees) that is substantially free from Government influence.

Only Americans that don't listen to NPR are under that misconception.

There was a transition in public thinking in America from the 1960's to the 1980's (Reagan was a big force in this movement) that government could do good for the public to a belief that anything the government does could only do things badly (inefficient, red tape, bureaucracy, fraud-infested).

As usual, the truth is never so simple: government is capable of doing good or evil just as much as the people that comprise it.

Re:dirac vs. theora?

[mike260]

That's like saying that Medicare/Social Security aren't paid for by the government, but by US citizens. True in one sense, but pedantic and moronic, especially since the relationship is understood.

But not by you, evidently. Medicare and social security are paid for (and run by) the the US government. The BBC is paid for by a license fee which comes directly from TV owners.

If it was a government funded body then it might have thought twice about attacking the government over their made-up WMD/Iraq claims, so I reckon the distinction is quite significant. Does that make me a pedantic moron too?

Re:dirac vs. theora?

[gowen]

Well hang about. The BBC said something about the government. The government got very upset about one specific allegation ("The 45 minute claim was inserted by govt spin doctors against the advice of the JIC") which Gilligan inserted off the cuff and which no-one believes to be true (even Gilligan admitted that was wrong).

The government then said "Will you retract that, as it isn't true". The BBC asked Gilligan, he stood by it. The BBC said we won't retract that.

Flash forward ... Hutton says: "The BBC's processes in checking Gilligan's story were woeful" (undeniably true; they asked Gilligan, then based their defense on the assumption he hadn't lied to them, which he had).

People think Hutton was a whitewash, because almost no-one's read it, and every newspaper in the country felt the need to stand up for their journalistic brother and pretend that the kerfuffle had been caused by something other than one specific lie in Gilligan's story.

Re:dirac vs. theora?

[imroy]

The Australian Broadcasting Corporation is government funded. It has specific rules about non-partisan bias, especially during election campaigns (like right now). Although its very position (non-commercial, etc...) tends to give it a slight bias towards the left, which the current right-wing coalition government has been whinging about on occasion. The youth-targeted Triple J radio regularly pays out commercial radio too.

Re:dirac vs. theora?

[gowen]

No, it isn't. Medicare is paid for through general taxation. The BBC isn't, its entirely funded through a license fee, paid by people for the right to receive television broadcasts.

Re:dirac vs. theora?

[Malc]

Furthermore, that 9 years is longer than one and normally two terms of government (a term being 0 - 5 years, IIRC). If the BBC pisses off a government, there's a high chance that they will not be able to do anything about it! The BBC seems to often take an anti-government stance too just for the sake of, or so it seems.

A system I don't like is the one on Canada where the CBC are completely at the mercy of the government. In the US, PBS is kept in its place by being poor and constantly having to go on begging sprees.

Re:dirac vs. theora?

[dpuu]

Almost right. The "World Service" of the BBC is actually funded from the foreign office budget -- which comes from tax payers. BBC America is an independent organisation, part of one of the big american networks (in a group with Discovery and TLC).

The domestic BBC has two sources of funding: the license fee and "commercial ventures". For example, they sell cheesy old series to PBS for american viewers. And, of course, they sell DVDs, etc.

Re:dirac vs. theora?

[Silburn_Luke]

Sort of, but its complicated. The BBC is an organ of the state, but it is not run by the executive arm.

First the BBC *is* actually responsible for collecting the licence fee. They farm the operation out to another entity, but its a statutory responsibility written in to their charter.

Second the BBC's grant-in-aid funding is paid from the the pot of licence fees but its level is set when the the BBC's charter is renewed every decade or so (of course the govt of the day has a large influence over that process when it occurs). So yes, the grant often diverges from what is in the common fund but the license fee which fills that fund is explicitly tied to this payment stream. And yes, the GotD has a big stick it can wave at the BBC - but a decade is a long time in politics and whilst theoretically, vide the Crown in parliament, the GotD can abolish the BBC (ie fail to renew its charter) if it gets uppity, the cost in goodwill would be horrendous. Even in her most eye-swivellingly megalomaniac stages, Thatcher never seriously considered doing that.

Addressing the way upthread post that started this off, the BBC is explicitly charged as part of its charter with conducting R&D into things like broadcast and storage technologies so this is exactly what they should be doing with the money they've been given. If they weren't, they'd be failing to fulfill their mandate. There's a lot of stuff out there that has come from the BBC Technology Divisions. Our gift to the rest of the world.

Regards
Luke

H-264?

[TiMac]

From mobile phones to HD, huh? Sounds a lot like the H-264/AVC codec that Apple is including as part of 10.4 Tiger that is an open standard that's been ratified.

What's the advantage to using Dirac over a standard?

Re:H-264?

[Xylaan]

That depends on if third party implementation of the codec would infringe any patents. One of the goals of Dirac was that it is not supposed to be patent encumbered.

Re:H-264?

[damiam]

H.264 is outrageously expensive to license. Dirac is an open, patent-unencumbered codec that can be freely used for whatever you want.

In the great tradition of geek-dom, ...

[reporter]

99% of the files used to test the new Dirac CODEC will be pornopraphy. Most of it will be weighted towards luscious, blonde lesbians engaged sexual acts that almost defy gravity.

Someone explain

[stratjakt]

What the major difference with this codec is. Why is the BBC developing their own codec instead of, for instance, throwing a few bucks towards OGM or XVid, or $YOUR_FAVORITE_OSS_CODEC?

Re:Someone explain

[imsabbel]

Because OGM is only a package format (like avi) and XVid is everything but legal (thus those "only for learning uses" disclaimers) because they simply decided to ignore the patents ( and divx is adware because divx-networks pays the royalities)

I say help

[AKAImBatman]

AKAIK, it's the only high compression video codec to not be encumbered by patents. (Although I've heard whispers from the OGG/Vorbis team.) That right there makes it worth development. Once the codec reaches a stable version, it can be integrated into free player solutions like HelixPlayer and VLC.

Cheaper patent licenses

[tepples]

Unlike licenses for MPEG standards, some licenses for the Dirac codec will be available royalty-free.

Re:Cheaper patent licenses

[tepples]

Dirac is available under "some licenses," namely the Mozilla.org tri-license of MPL+GPL+LGPL. The applicable patent licenses are granted royalty-free to any copyright licensee of the code.

Suggestions for Team Dirac:

[Power Everywhere]

• Release Dirac for QuickTime.
• Optimize compression algorithms for individual CPUs. Is Dirac running on a Pentium 4? HyperThread it. Is it running on a PowerPC G4/G5? Optimize for AltiVec. Same applies for Sun's VMX, MIPS' MME, etc.
• Release the codec under an Open Source license but one that will disallow forking or total appropriation (re: Not BSD or GPL).
• Start a web community/forum and accompanying mailing list for it.

Re:Suggestions for Team Dirac:

[THEbwana]

...err like .. perhaps the entire archives about to be published? http://news.bbc.co.uk/1/hi/entertainment/tv_and_ra dio/3836941.stm

Re:redundant

[Spad]

But none supported by an entity as large or influential as the BBC.

Codecs like Theora are great, but it's unlikely they'll enter the mainstream in the same way as something like DivX has - just as Vorbis is lagging behind other closed source audio codecs.

If the BBC started using Dirac for all its streaming video feeds, for example, then suddenly millions of users will have an excellent incentive to download the codec and if people already have it on their machines then others can produce Dirac based media without having to worry that people won't want to view it because it means downloading something extra.

Re:BBC + Codec = Not Free

[Anonymous Coward]

The BBC isn't a business, and isn't trying to make money. It's a state-funded channel, and is paid for by a license fee. I think it is actually a violation of their charter to get money from anywhere else: I know they're not allowed to run commercial advertisements.

Re:BBC + Codec = Not Free

[stratjakt]

American's don't get the concept of a Crown corporation. They don't act like any other "rational business". They don't really have to make a profit, and the way most are run, making a profit is a secondary objective.

I remember the sucking money hole that was Air Canada before the government chopped it up and sold it off. All of a sudden it's a profitable business, turns out they didn't need to be sending 737s to Beaversnatch, Alberta thrice a day.

Links to sourceforge and BBC's homepage

[Anonymous Coward]

Sourceforge project
BBC's Dirac homepage

They already have...

[Phil John]

Check the sourceforge project

Why Open Source Codes are essential

[alistair]

Anyone wondering why we need more Open Source Codecs should read the excellent companion article on today's register, a long OP Ed piece on Steve Ballmer entitled Love DRM or my family starves: why Steve Ballmer doesn't Get It.

In it Steve explains why the Digital Home has to come from Microsoft and specifically Microsoft's committment to DRM everywhere. A facinating, if biased piece.

Re:BBC + Codec = Not Free

[onion2k]

The BBC, just like any other rational business, is out to make money off of this while the rest of the world could benifit greatly from it.

Nope. The BBC need the codec in order to save themselves a bucketload of cash in the future when they make their digital program archive available over the internet (something they have to do according to their Charter). They're not intending to make pots of money from the codec, they just want it to exist so they can use it themselves.

Reasons Dirac is Not Redundant

[TheRealFoxFire]

1. Not patent encumbered (compare to H.264 and MPEG2/4 including "open source" codecs like XviD)
2. Next generation coding techniques (wavelets vs traditional DCT coding) (compare to Theora/MPEG 4)
3. Capable of scaling down as well as up (compare to MPEG2)

Re:redundant

[FrostedWheat]

The only truly open video codec worth mentioning is Theora. XviD's source may be open, but the codec itself is a patent minefield. Theora is patent free, as is Dirac. Even if the BBC did take out some patents, the license Dirac uses means these patents would be harmless.

So yes, we do need this codec and others like it. Theora is nice but it dosen't hold up against any of the new generation of commercial codecs that are coming out now.

Re:redundant

[AstroDrabb]

Can you please list all of the Open and non patent encumbered codecs? I can only think of Theora. Of all the codecs out there just about every one is enbumbered by a patent or license fee or DRM which hinders thier usage for distribution of public content such as documentaries.

You're missing a lot

[Crosma]

Dirac is a wavelet codec. The technology is far more advanced than Theora's. In fact, until On2 came along, Ogg were working on a video wavelet codec called Ogg Tarkin. They want with open sourcing VP3 because it would be quicker and easier, nothing more. As the BBC are demonstrating, putting together a competent wavelet-based video codec is non-trivial to say the least.

Put simply, Ogg Theora is already outdated. The source material (On2's VP3 codec) does not match any decent MPEG-4 codec. The BBC would be wasting their time by messing around with dated tech.

That said, Theora is usable and just about the only decent patent unencumbered video codec in existance. Until Dirac is finished, Theora will remain the sane choice for those who want to stay legal without paying through the teeth.

If and when Dirac is ready, it will blow everything else away. It will be worth the wait.

Re:You're missing a lot

[bullitB]

Dirac is a wavelet codec. The technology is far more advanced than Theora's. In fact, until On2 came along, Ogg were working on a video wavelet codec called Ogg Tarkin. They want with open sourcing VP3 because it would be quicker and easier, nothing more.

This isn't really true. Wavelet codecs are not necessarily better than non-wavelet codecs. This is especially true in the case of video, because, as of yet, no one has figured out a way to efficiently peform motion estimation in the context of a wavelet codec. While wavelets in the context of still images have done very well (see JPEG2000), most attempts in video have not been so successful (see Indeo 5 or...Tarkin).

I think it should say a lot that after briefly experimenting with wavelets in MPEG-4 "texture" compression, the smart people behind AVC (aka H.26L/H.264) decided to completely forget about wavelets in their next codec. In fact, AVC doesn't even use a classic DCT, it uses an "integer transform," which is generally considered of even worse quality than the DCT used in MPEG-1/2/4SP.

The most likely reason Xiph started video work on Tarkin with wavelets first is that wavlets are completely patent free. When On2 granted them rights to use their DCT-related patents from VP3, that no longer became an issue.

Put simply, Ogg Theora is already outdated. The source material (On2's VP3 codec) does not match any decent MPEG-4 codec.

This is a real oversimplification of matters. The Theora guys can tune their codec (a lot), and there is a lot of stuff a VP3/Theora encoder could do that an MPEG-4 encoder couldn't. There was a time when Vorbis was not even up to the level of MP3. A few years of tuning later, and now it's beating everyone.

If and when Dirac is ready, it will blow everything else away. It will be worth the wait.

I've heard this one before.
Video compression is around 15 years old now. For maybe the last 10, "wavelets" has been a hot keyword that gets people thinking "Ooo, that'll change everything!" The confusion got even worse with JPEG-2000, since now everyone seems to think that the gains in efficiency from JPEG to JPEG-2000 will be directly applicable to video (ignoring the facts that a lot of that comes from JP2's arithmetic coder and improved predictor, both of which are already being used in video codecs). Point is, I'd look at Dirac with a lot of skepticism. The fact that it is currently unable to decode video in a meaning manner at normal speed concerns me greatly. This suggests that it's already 10-100x times slower than current generation video codecs. Frankly, I think making something 100x faster (needed for Dirac) is probably going to be harder than making it perform 50% better (needed for Theora),

Re:You're missing a lot

[BillNyeTheScienceGuy]

We are probably reaching the limits of compression technology. It doesn't matter what space (DCT or Wavelet) you do your compression in, after a certain point you just can't throw out any more information without seeing visual artifacts.

With the jump from MPEG2 to MPEG4 type codecs we can see a compression gain of about 5-10 times the compression effiency. But basically MPEG4 came out of the "lesson" of MPEG2 and as such you should expect that it does some things better and some things worse. MPEG2 was the first major leap and MPEG4 was able to jump off even further because MPEG2 was so immature. MPEG4 is both highly developed and optimized and non likely to get much better. H.264 is supposed to be the next big thing, but it turns out that it is only incrementally better in terms of quality (and not in all cases!), but is far more complex and less attractive to chip manufacturers.

Essentially we are reaching the hairy edge of compression and more advanced codecs will be orders of magnitude more complex and the incremental gains will be very small. It is unlikely that we will see any more revolutionary video codecs that are "the greatest thing since sliced bread."

It is starting to reach a point where the surrounding features are the selling points. For example the DivX certified program ensures that your certified DVD player can play your DivX movies. You can't do that with real, wmv, or even Xvid. Our marketing department loves to say this, but DivX is the MP3 of video. Most codecs simply focus on compression, but it really is more than that! Video is the entire experience starting with the encode and ending at all the places you can watch the movie.

The bottom line is "Content is King." Even if you are 5% behind the best video codec, studios are going to look to the codec with the widest consumer base. PSNR be dammned, how many people can watch the movie is a far more powerful business case.

I am glad that the BBC is out there looking at this stuff as it is a breath of fresh air to have a studio actually understand how video compression works! However, I don't know that they will be able to "add value" in a way that will encourage people to use their codec.

Keep on encoding, people!

(NB- I have a slightly biased opinion as I am a codec developer at DivX... ;p )

Re:You're missing a lot

[Spy der Mann]

Just to inform you, in this article some guy applied the subband coefficient encoding algorithm used in wavelet image compression, to simple DCT coefficients (JPEG). The results: 16pixel-block DCT transform outperformed the wavelet image coding.

I'm sure that with a little research, much could be done with BBC's codec.

Re:You're missing a lot

[Bloater]

> The fact that it is currently unable to decode video in a meaning manner at normal speed concerns me greatly. This suggests that it's already 10-100x times slower than current generation video codecs.

Until recent optimisations, I haven't been able to decode broadcast resolution video realtime with any theora players. The issue is C/C++ vs vector assembler (ie, SSE/3dNOW) for the main transform.

The DCT has many fast implementations, the Mallet transform doesn't - lifting is one part of that, but the wavelet filters (along with the lifting algorithm) need implementing in assembler.

realtime lords

[Doc Ruby]

I will trade codec engineering time for TARDIS technology. In fact, that's where I got my TARDIS from!

Re:realtime lords

[AndroidCat]

I thought they wanted help with "Dalek Coding", oops.

Re:redundant

[skids]

Don't we already have enough codecs, including open source ones?

While I agree strongly that there are a lot of reinvented wheels in OpenSource that add nothing new or unique, audio codecs are a wide open area for innovation. There is a lot of complex mathematical theory involved and while many very smart people have more than just scratched the surface, we could see considerable improvement with more development. Each project serves as a test case for the methods it uses.

Personally, I'm dissappointed that the idea of using genetic programming (or related technology) to develop or improve CODECs has not, at least to my knowlege, taken off. Hopefully the people with the expertise in both fields will at some point come together. That would be a worthy use for the resources we have at our disposal these days, IMO.

I used to think this would only be good for lossless CODEC developement, but perhaps automated fitness tests for lossy CODECs could also be practical.

Re:Coders like to be in charge of code

[Smuttley]

If you knew anything about the way the BBC runs and operates you'd realise how dumb a comment that was.

Re:theres one difference between the off video cod

[magefile]

Wrong. Theora is nearly there, whereas Dirac isn't even working in realtime (RTFS). And, it lets them stay with one paradigm (I can't believe I just used that word) because Theora has an audio analogue (ogg) whereas Dirac doesn't.

And that's ignoring the benefit of being involved with an OSS project that, while rough around the edges, has a large development community already (both Theora devs and the potential pool of Ogg devs who could be enticed to work on Theora), rather than starting a new OSS project.

outsourced

[0xbeefcake]

I'm wary of the fact that this "call for help" comes just days after over 1400 BBC technology staff were out sourced to Siemens

Re:outsourced

[Blitzenn]

Apparently the managers feel that they can get their work done for nothing now with all this open source stuff going on. Are we putting ourselves out of work?

Re:outsourced

[oneiron]

Are we putting ourselves out of work?

Possibly? Which one is more important to you, your career in software development or the good of mankind being at the core of software development? Which do you think is more important to the rest of the open-source community? Can you have it both ways?

Tough questions... Is it even worth bothering to guess at answers?

codec modules?

[Doc Ruby]

Codecs are modules that fit into apps. Consistent with the three-tier architecture, they have APIs called by apps, and data access to the streams on which they operate, both of which are fairly generic (and ought to be completely standard). Their cores are different, depending on their transformation, their source/destination data formats, and their transformation technique, as well as metadata produced/consumed.

New codecs come along infrequently, and are usually too little, too late. There's a lot of duplicated effort across these projects. It seems a better strategy for everyone to share a skeleton that gets populated with codec core "plugins". An easy install mechanism might even let new datatypes deliver the smaller cores for codec'ing on the fly.

Re:BBC + Codec = Not Free

[AstroDrabb]

I think enough people answered your post about the BBC being "out to make money". So I will address this one.

Furthermore, if and when the BBC has a working codec will they be held responsible for copywrited material translated into it? I'd like to see the outcome from a lawsuit between the MPAA (Motion Picture Association of America) and the BBC reguarding thier codec being used as a good way to view movies.

Think about your question. Could the MPAA sue MS because someone violates a movie copyright and just happens to use MS's video codec? Of course not, that would be silly. It is no different with the BBC. If I distribute a movie encoded in Dirac and I do not have the rights to do so, then I am the one that would be liable for copyright violation, not the BBC.

saw them yesterday

[t_allardyce]

I saw them at the Linux expo at Olympia yesterday, it looked pretty decent and its still alpha, they said they sometimes get people helping and pointing out bugs, its pretty rockin that they're getting funding considering the direction the BBC is going, definately better than suns java desktop, but damnit they wernt giving away any penguins or anything >:(

Re:Lets stop paying for all software

[Technonotice_Dom]

I don't normally feed trolls... but what the hell.

a) the BBC isn't just a "company" - it's the highest quality broadcaster in the world. They always have done research and been at the forefront of new technology throughout their history. This is a project that anybody can help contribute to - as it'll benefit the community as a whole when it's complete.
b) they have put effort into it already - they've put out quite a few releases already (SF page) and have been working on it for a couple of years
c) although they want it to improve their online streaming services (currently done using Real technology), an open standard, no encumbrance from patents, with technology that other codecs at present don't use, is a very important project for not only the BBC, but for all of the computing community

Re:What about high quality profiles for a change?

[DaHat]

If you are seeing ATSC video that looks like crap, I'd be careful what you blame. It could be the MPEG-2 video, but it's more likely a low bitrate encode.

ATSC gives each RF channel 19.392658 megabits/second... very few broadcasters use all of that, in fact the majority tend to stay under 10 even with multiple programs in their broadcast.

Sporting events, like high motion/action movies often need more bandwidth to look good than a soap opera or day time talk show, it's likely that someone either was lazy and didn't up the rate for the game or they just don't do that ever.

Take a look at DirecTV sometime, you'll notice that many of the movie and tv channels are pretty low bitrate, but for major sporting events (ie super bowl, boxing, etc) they up it very high to ensure a pristine signal and picture.

Re:This CODEC is a good thing!

[Blitzenn]

Tell that to the people on the black list. YOU would have to ba a moron to not recognize what is going to happen there.

Evolving a codec is not going to work.

[Jim McCoy]

Personally, I'm dissappointed that the idea of using genetic programming (or related technology) to develop or improve CODECs has not, at least to my knowlege, taken off.

The problem is that the evaluation loop is too expensive. It is _trivial_ to develop a system that attempts to eveolve various mechanisms to encode data, but to iterate each generation you need some sort of way to determine the winners and the losers. If you could figure out a way to use a program to determine which was the better of two video encoding mechanisms it would be worthy of a PhD or two. The simple way to think about this is that if you could perform this evaluation with a computer you would have figured out a practical mechanism for general-application computer vision.

One possibility is to corral ten thousand or so friends and get them to view three encoded clips (the original source and the versions produced by the two population members you are testing) but keeping this up long enough to end up with a decent encoder is an unlikely proposition. Making this harder is that once your population of encoding algorithms weeds out the obviously broken solutions you need evaluators who can determine things like which codec produced the fewest artifacts and other details that would shrink your potential pool of human evaluators.

The hard part about using evolutionary computing techniques is not the evolution engine, it is all about figuring out how to test the population members your engine generates. If you do not already have a well-defined target that is easy to describe mathematically (or a test environment in which you can pit two population members against each other) you are basically screwed.

Re:Evolving a codec is not going to work.

[skids]

It is _trivial_ to develop a system that attempts to eveolve various mechanisms to encode data, but to iterate each generation you need some sort of way to determine the winners and the losers.

I am not so naive as to be suggesting human evaluation here, give me some credit willya? :-)

First off, as a side point, for lossless encoding evaluation is trivial.

Secondly, there has indeed been much work towards automated performance evaluation of lossy codecs. Not too much on video yet, but a lot on audio, right down to the level of modeling the resulting neural impulses generated by a waveform in the human ear. By using existing research which involved human viewing and listening surveys (Other people's PHd's), developing fitness tests is not as hard as you make it out to be.

Finally, while evolving a whole CODEC is probably not practical with today's CPU power, there are a lot of subsystems which could be optimized through GA/GP to improve their efficiency. Many times in an algorithm you have a subsystem who's functionality is well defined, but who's optimal implementation or parameters are not known.

For example, many algorithms use lookup tables, and I'm sure a clever mathemetician could come up with a family of symmetrical transform functions that vary across a set of coefficients. Those are probably the cases which GA should tackle first, because the search space is much smaller and represents a constant, a "coefficient" to use the term very loosly, of an algorithm rather than a whole algorithm.

The general idea here is not to magically create the best looking/sounding CODEC ever out of thin air. It is to take the goals which we suspect will result in good CODECs and find new algorithms to acheive them. Once we find optimal solutions to those, we either dissect them for insight, which improves our base of theory, or at that point we submit them for side-by-side human comparison with existing CODECs.

dirac

[xmp_phrack]

It's named after a Paul Dirac, a British scientist who worked on quantum mechanics.

Re:BBC + Codec = Not Free

[isorox]

The BBC need the codec in order to save themselves a bucketload of cash in the future

Yet ironically we're moving some DHCP servers from linux to windows, plan to move the central image server from Linux/SGI to Windows/SGI and then finally Windows/Windows, have just implemented a multi million pound project, in java, but put in windows servers for most of it, refuse to consider Open Office, refuse to have Mozilla as part of the standard desktop (and you have to jump through hoops to "legally" install it), and have half an intranet that's unavailable to the (few) Mac, Linux and Mozilla users, and the entire of Research and Development.

The BBC is a large company, some sectors are run my MCSEs living up Bill Gates' ass, others are at the forefront of technology.

Re:Good ol' tax-funded science at work

[Triskele]

I tend to think that the BBC is as great as it is due to it being very well managed and structured, not by the fact that it is taxpayer funded. The mere fact that taxpayers pay for it does not guarantee success.

Well if it was commercially owned I think it would guarantee failure. The BBC's culture is quite unique and probably not particularly understandable unless you're a Brit as they're a hangover from the days of the Establishment - a belief (however pompous and misguided) in public service for the good of the nation (and who's good has always been hotly contested).

Sex hurts compression

[0x0d0a]

Remember that the famous Lena image, which was cut from a Playboy magazine, was a *major* still image compression benchmark for a long time. It was a pretty bad choice -- it has a duplicate line at the top, it doesn't necessarily have the ideal color range, and worst of all, it was copyrighted.

The urge to benchmark with smut is strong, but should be resisted.

How does dirac compare to MPEG-4?

[Anonymous Coward]

How does dirac compare to MPEG-4 when it comes to compression? And how about performance?

Wavelets patent-free? I think not!

[TimoT]

The most likely reason Xiph started video work on Tarkin with wavelets first is that wavlets are completely patent free.

Hehe. ROTFLMAO. I have done data compression research and there are very few mathematical ideas as patent encumbered as wavelets. A quick naive search of USPTO patents with wavelet and data compression brings up about 250 patents and just wavelets about 10x as much and that's not even looking very closely.

TimoT

Re:Wavelets patent-free? I think not!

[bullitB]

I have done data compression research and there are very few mathematical ideas as patent encumbered as wavelets.

True, my statement "wavlets are completely patent free" is errant. (And not just because I spelled wavelets incorrectly. Ouch.)

Wavelets are, however less patent encumbered in the context in which they are used in Tarkin and Dirac, which is...why they're being used in Tarkin and Dirac.

Breaking the law, breaking the law

[tepples]

Do not confuse the codec itself with the designated player.

Unless the copyrights and patents in the codec itself are licensed only in conjunction with the designated player.

Re:This CODEC is a good thing!

[francisew]

I'd think that the argument someone made in an early post about the BBC not being a software development company applies. It only makes sense for the BBC to be involved in media distribution development to a very limited extent. Hiring a single person (or a few people) to coordinate the CODEC development makes sense. Hiring a full blown programming team wouldn't. They will need a continual progression of work over a long-term. They will also not be licensing the technology or getting a revenue for it. So why would they hire anybody to develop it?

I would be surprised to find any BBC worker who was laid off from a BBC 'software CODEC programming job' because of OS development. If anything, it will boost the productivity of employees working on the CODEC, by allowing them to develop the CODEC more quickly and robustly. This is a matter of asking the community to help develop a tool that it would like to use, without footing the full bill.

Besides that, why shouldn't I be allowed to give my own time? I can volunteer in a hospital, a shelter or as a tutor, why shouldn't I be able to volunteer my high-tech skills for a cause I believe to be worthwhile. Isn't it worthwhile to reduce the cost of disseminating the 'free press'?

"BEGIN TROLL FILTER" The US & britain are bombing other countries in the hopes of bringing them freedom. Why not support a better dissemination of information to them, to help distribute a picture of what is going on in the rest of the world. Change can be voluntary instead. "END TROLL FILTER"

Should a large portion of the BBC IT staff be paid to develop CODECs? I believe not, it's not the BBC's task to develop distribution media. However, you are raising a completely extraneous point, since I'm in no way replacing in-house IT staff. But if I felt that I could volunteer time for replacing IT staff, why should I feel bad about it?

I'm not quite sure what upsets you about this whole issue, but please feel free to explain to me how developing a useful, novel open-source CODEC puts IT staff out of work.

Depending on the aircraft

[csmacd]

They may actually be weightless! Zero-G aircraft, private spaceflights....the possibilites are endless!

Re:redundant

[mconstable]

If the BBC started using Dirac for all its streaming video feeds, for example, then suddenly millions of users will have an excellent incentive to download the codec and if people already have it on their machines then others can produce Dirac based media without having to worry that people won't want to view it because it means downloading something extra.

But if the BBC started using Theora now then that would help Theora get off the ground for exactly the same reason you outline above. A move like this would give the open source movement confidence the BBC were serious and exposure to the concept of a government agency sponsoring truly open source video codecs. Then in a few years we would all have two excellent and (more) widely used open source video codecs. More power to them, and us. As long as they keep using Real and WMA they are not endearing open source developers. I use an AMD64 machine and can't be bothered setting up a dual 32-bit environment, so I can't use Real, therefor I don't get to view any BBC content at all. Not that this means anything other than I'm one extra open source zealot that can't view BBC video.