Paper on technical details of WMV9

[FastMike]

I found this interesting paper on www.sciencedirect.com about Windows Media 9:

Windows Media Video 9: overview and applications
Pages 851-875
Sridhar Srinivasan, Pohsiang (John) Hsu, Tom Holcomb, Kunal Mukerjee, Shankar L. Regunathan, Bruce Lin, Jie Liang, Ming-Chieh Lee and Jordi Ribas-Corbera,
Signal Processing: Image Communication, vol 19, October 2004


It has some pretty interesting details on the WMV9 codec, such as variable transform size and in loop filtering. They also have some rate-distortion plots and some subjective testing results comparing against MPEG-2, WMV8, and H.264.

I am having trouble linking it, so goto Journals->Signal Processing: Image Communcation->Vol 19, Issue 9

[skal]

Originally posted by FastMike
I found this interesting paper on www.sciencedirect.com about Windows Media 9:

Windows Media Video 9: overview and applications
Pages 851-875
Sridhar Srinivasan, Pohsiang (John) Hsu, Tom Holcomb, Kunal Mukerjee, Shankar L. Regunathan, Bruce Lin, Jie Liang, Ming-Chieh Lee and Jordi Ribas-Corbera,



Wow. Nine (9!)(Neun!!) people from Microsoft
writting a paper about how great
WM9 is compared to X, Y *and* Z. Bill wasn't
available for co-signing the paper?


Skal

[Nic]

@skal: LoL :)

@FastMike: looks like you have to signup and/or pay for that paper... ?

-Nic

[gotaserena]

It's sort of expected that WMV9 would beat MPEG-2, but what is exactly the point of comparing it to H.264 Baseline for purposes of quality? Just because the reference decoder cannot decode H.264 as fast as the commercial microsoft application can do WMV9 doesn't make H.264 Baseline more complex.

Read through sections 4.4 and 4.5 and get the strange taste of "advertisement pushed as science" in your mouth.