http://www.tvtechnology.com/features/news/n_AVC.shtml
The technical literature seems to all agree on one thing - H.264 AVC will be superior in quality but VC-1 will require less processing overhead.
Other things the article mentions are AVC is more advanced in the standards process, and VC-1 components are built into Windows media. (hence Microsoft's purpose for proposing VC-1 rather than AVC (of which they are a member) as a standard seems to be to promote Windows media over competitors).
As for the differences between the two codecs, Goldman said, "from a high-level point of view, they have much in common in regard to coding efficiency The differences are deep down in the coding level, either with how or where a coding tool is applied in the processing chain, or in a few cases, with the choice of a different tool to perform a particular function. Both are excellent and they far exceed the capability of MPEG-2 in bit-rate reduction.
"AVC is more complex and may ultimately offer the potential for higher quality and more efficient coding. VC-1 is less complex and therefore, presumably easier to implement efficiently," Goldman said.
Examples of specific differences between AVC and VC-1 include AVC's use of six-tap filters versus VC-1's use of four-tap filters. AVC uses an entropy scheme known as CABAC (context-adaptive binary arithmetic coding) that is computationally intensive and therefore expensive to implement. VC-1, by contrast, uses high-order entropy coding that is less complex and less expensive to implement.
"We chose smarter algorithms that are less complex," Ribas said. "It is a significant advantage to be less complex. Rather than the kitchen sink approach, we were very careful in our design choices."
Modulus' Robinett, however, believes that CABAC is one of the elements of AVC that will provide video quality superior to VC-1.
It is standardization that drives acceptance of these new codec technologies. AVC has, in fact, been an international standard for about one year now. A Joint Video Team (JVT) composed of coding experts representing the ISO/IEC and the ITU-T organizations (from the moving picture experts group [MPEG] and the video encoding experts group, respectively), created this advanced coding standard.
These organizations are now poised to ratify extensions to the AVC standard that would result in four new profiles including High Profile directed at broadcast television applications, as distinct from the currently ratified Main Profile. The AVC High Profile is a superset of the Main Profile and adds a few new tools, including the adaptive use of an 8x8 block transform and quantization scaling matrices. The technical work is complete and as of this writing is awaiting formal ratification.
VC-1 is well into its standards process under the auspices of SMPTE. Based on the Windows Media 9 Advanced Profile, which is roughly equivalent to the AVC Main Profile, VC-1 includes tools such as the 8x8 block transform that appear only in the new extensions to AVC.
The technical literature seems to all agree on one thing - H.264 AVC will be superior in quality but VC-1 will require less processing overhead.
Other things the article mentions are AVC is more advanced in the standards process, and VC-1 components are built into Windows media. (hence Microsoft's purpose for proposing VC-1 rather than AVC (of which they are a member) as a standard seems to be to promote Windows media over competitors).