[gillianseed]

There's no doubting gasset-glaser's (DarkShikari) knowledge in these matters, but he is hardly an objective party.

Apart from being one of the developers of open source GPL licenced x264 encoder, he and the other developers are also licencing it for proprietary use (nothing wrong with that, I think it's a great way of making money out of open source development). This means that a open source royalty free codec like VP8 (and later VP9) could cut into their revenue stream.

I'm not saying that this affected the outcome of his test, but it's still context. Also his test back then was also criticized for only using one clip, here's another test with a larger set of comparisons:

http://qpsnr.youlink.org/vp8_x264/VP8_vs_x264.html

I can't vouch for the validity of this anymore than the validity of any other test.