I touched on Google's purchase of On2 in my last post very
briefly, but it's something that has the potential to enable some quite
interesting change in the way web video is delivered - not least because of
Google's open sourcing of the VP8 video codec (compressor/decompressor) code and specification in mid-May
2010.
You'll have seen the work of On2 over and over and perhaps not
known it. For a good few years, their VP6 codec was used in Flash Video (FLV) format,
offering an improvement over the previous H.263 Sorenson Spark codec and more recently
being replaced by the popular H.264 codec used to efficiently compress much of
high-definition content on sites like YouTube, Vimeo and elsewhere.
VP8 is in effect a newer VP6, with a lot less fame than its older
sibling and a handful of new or improved features. Joining VP8 is the WebM (aka Matroska) video container format
- providing the mechanism to wrap a VP8 video stream together with audio and
other data.
But why acquire a video technology company whose most
popular IP is ... out of date? Enter the minefield of patents, licensing and
royalties. H.264 is made up of a lot of different pieces of IP (aka the MPEG-LA
patent pool ) from a lot of different parties. Whilst implementing an H.264
encoder or decoder with free software is possible (FFMPEG and x264 both do this), the
video format is not always free to encode into or distribute. I'll leave you to
read
through the detail - but suffice to say it's a little on the dense side.
With VP8, Google get something close to the quality of H.264 (and competing formats
like VC-1) with similarly small file sizes, without having to cross that
minefield. With VP8 web video users get a format that capable of good quality
output and is royalty free.
Or do they? When it comes down to it - VP8 and H.264 are technically pretty
similar - meaning VP8 is unlikely to be wholly patent unencumbered. What
does that similarity really mean? It means that some of the FFMPEG developers
managed to implement
a VP8 decoder using 1400 lines of their own code, and drawing from the existing
FFMPEG H.264 functionality.
So why does this all matter? Because this is the format with designs on ousting
other video formats as the defacto standard to deliver content into the HTML5
<video> tag. With 4.67 billion+ video views a month and growing, and the
prospect of GoogleTV - avoiding licensing costs for producing that material and
pumping it out to consumers is, as far I can see, a very desirable place to be.