Wasn't the case for Dirac that it sucked on low bitrates, and was very slow to encode?
Keeping in mind that the following remarks are not of an Lawyer but that of an inventor that has been down the Patent road before several times I will say the following...
If you're doing something like h.264, you owe royalties up front- and if the rights holders choose to not license to you in a RAND manner (witness the back and forth with Apple and Nokia recently...) you can't implement- and if you do, you're guilty of infringement. If you're found guilty of infringement, you have no recourse whatsoever since it was clearly shown there were patents involved.
In the case of Dirac, you don't know if there's patent coverage or not- the providers thereof have done a good faith effort to investigate possible infringements and came up with nothing at that time, however. There is a chance that you could be infringing on something. There is, however, some small recourse in the case of infringement if there is one found and the rights holder for the infringing tech chooses to not license in a RAND manner. You can stop implementing the infringing technology and walk away. No damages can be assessed because you did a good faith effort in determining whether you infringed or not and didn't know you were infringing at the time of the implementation. Moreover, depending on the date of the patent and the date of Dirac's release, there may be a bar to enforcement with regards to their patent and Dirac's use thereof by their delay in enforcement. This is a bit different than in the case of h.264. If you infringe with h.264, you KNOW you're doing it and that tidbit detail will be used to it's maximum effect on you in a court when they choose to litigate your actions there.
In the case of Theora, the known patents have been provided an irrevocable license to it's implementation and it's derivatives that can play that format properly- from the rights holder, On2. While there is risk of overlapping patent coverage from someone else, they would have shown up by now and any attempt to enforce their rights respective to Theora would likely be barred due to an excessive amount of delay by the rights holder (Theora and VP3 have been around for quite a while now...someone, if there's overlapping patent coverage should have stepped up to the plate by now.). If so, there's much, much less risk with using Theora, actually, as a baseline codec, than anything else as it is very unlikely to have issues with regards to patents than anything in the same class right at the moment.
I'm not saying that Dirac's better or worse overall to Theora, mind- just that the comparison you linked to doesn't go far enough in comparing things... :D
Probably dirac does not violate software partens. But if it does (Noone knows, thats the being of softwarpatents) it's easy to write the code 'around' the patents: http://diracvideo.org/wiki/index.php...any_patents.3F
In the wiki there is also a test dirac video. Just test it if you like.
I think the point is that dirac is not really ready, yet. And theora is. You cannot really put dirac into mkv containers. Ogg is available longer and "just works" in most cases. And the qualitiy should be good enough in most cases. For general web video at least. JPG has also its disatvanteges and there are better alternatives, but for the general use it is enough. And it just works, so the situation will not change in the near future.
We need a free video standard for web video now. And dirac can be used in contrary to theora also for HD content. But it's not so well tested and optimised as theora, yet, but promising for the future.
And that we can decode H.264 till 2016 for free does not solve the fundamentally problem with H.264: http://www.h-online.com/open/news/it...16-921828.html
The reason dirac is relatively "safe" from patent disputes is because it uses some of the oldest and most widely used techniques (which would have expired if patented, but never were) for video encoding. If a patent claim was made against it there are near endless examples of prior use. It's pretty damn safe.
That said it's focused on high quality at high (sometimes extremely) bitrates. This isn't really suitable for streaming video.