I've been reading a lot about VP8 lately. It sounded great at first: A modern video codec acquired by Google in their On2 purchase and released as open source for royalty-free HTML5 video. Anything that gets rid of Flash for video playback would be awesome, and if WebM and HTML5 could do the trick, I was *so* on board. But before the press event hangovers wore off, warning flags were popping up.

Tests purportedly show VP8 to decode slower than H.264, which would be particularly bad for phones, tablets and similarly small devices. Especially since the code is thought to be pretty well optimized already. I'd like to see some more hard performance data, but there's not a ton of it out there yet. I suspect On2 of overstating the performance of VP8 (which would be typical).

And VP8 is apparently so similar to H.264 that MPEG-LA's lawyers are already mansion hunting. Google is not indemnifying users or developers, though. Even Microsoft offered legal cover for VC-1. But that codec was also very similar to H.264, they had to give it up rather than pay the royalties for everyone using it.

If VP8 survives the IP attorneys, it'll be a nice upgrade to the other open source codecs. But it'll probably need an overhaul to outperform H.264 if the early performance reports are verified. Google may go ahead and push it out anyway, but it's hard to believe the H.264 royalties are more expensive than the lawyers.

MPEG-LA should just ask Google for half the money they're going to spend this year on development, YouTube conversion, marketing and legal defense for VP8, distribute the windfall to the companies in the patent pool, and submit H.264 to SMPTE or IEEE. Everyone gets a free, high-quality codec, the patent pool gets a big buyout, and Google saves a lot of time and money. It's hard to imagine them actually doing that, though.