Ogg Theora?
Much muttering is taking place on the question of whether the W3C should choose a default video compression algorithm for the Web. In particular should the W3C adopt Ogg Theora as a default CODEC for Web video.
As a practical matter H.264 and AAC have become established as the defacto CODEC of choice for delivery of online content. H.264 is deployed on cell phones, in Blue Ray and HDDVD players, You Tube and iPods. It is also the codec used in the new line of HD camcorders from Sony and Panasonic. The main reason for this is that H.264 is precisely positioned at the 'sweet spot' for delivery of video content over broadband. When the default video encoding on YouTube was H.263 the frame size was limited to 320x240. The recent switch to H.264 allowed the frame size to increase to 640x480, comparable to NTSC.
Ogg Theora may or may not deliver the same level of performance, finding meaningful reviews of CODEC performance turns out to be harder than it should be. The only review I could find that included Theora in its list of candidates rejected it for its quality comparison tests on the spurious grounds that it had failled to compress The Matrix sufficiently to fit on a CD-Rom, hardly a test that many would consider relevant.
The real value of Ogg is that it purports to be unencumbered. The inventors of Ogg have been granted patents but have granted an irrevocable royalty free license. Problem solved?
If only...
My skepticism is due to the fact that I have been here before with both GIF and JPEG. I was the principal instigator of the move to replace GIF with JPEG after the UNISYS patents claim covering Lempel-Ziv-Welch compression became public. Despite the intention to adopt JPEG as a clearly unencumbered format, other parties have since asserted claims in connection with JPEG.
If we had known that there were trolls we would have built the bridge in a different place. The sheer volume of patent claims related to compression techniques makes it impossible for anyone to ever be sure that they are building in a troll-free zone.
The only way we can be sure that an algorithm is not encumbered is to adopt a standard that is sufficiently old that any patent claims have expired. Far from encouraging development of new technology the patent system as it currently operates encourages us to use technology that is at least 20 years old. Even then we do not have perfect certainty as until 1995 the term of a US patent was 17 years from issue rather than 20 years from filling as at present. Applications made prior to 1995 have the advantage of the longer of the two terms. It appears that the only way to be certain that a video CODEC is unencumbered is to wait for the patents on MPEG-2 to expire.
The net result is that corporations are going to prefer H.264 to the alternatives because the one thing they like less than paying a royalty for use of an unencumbered technology is the uncertainty that adopting an alternative would create.
It should not be this way, but it is.
