Google’s(On2) VP8 and the video future

25
May/100

Google has made “public” and “open-source”, “patent-free” and “royalty-free” their recently acquired with the purchase of On2 video codec VP8. It’s an amazing step for the world of digital video. This step obviously was coordinated with other companies like Adobe, which ensured support for the new codec in further versions of Flash and Sorenson that presented an encoder ready to work with the new codec. They have also provided patches for the hugely popular transcoding programs ffmpeg and mencoder that

This is big news and consequences for web video may be really profound. It’s an amazing amount of events for a single month and they did in in a single day. Well done Google!

Google has also “created” a new video file format which is supposed to use the new codec. Well they didn’t create the format as it is Matroska container and the combination of VP8 video and Vorbis audio would be possible anyways, but a specification and a new name for it. It’s now called WebM. Cheers to all

Now why is this BIG?

First of all I’m not convinced yet that this is it, but for the sake of enthusiasm I will pretend to be blind and I will comment the VP8 codec and not talk much about patents.

We have a mature file container (Matroska) with a mature royalty-free audio codec(Vorbis) and now a royalty-free video codec for them. As a result that’s a good quality, royalty-free video format. That’s something many people dreamed for with Theora coming true. The weakest part of this combination and the most important patent/royalty wise is VP8. Let’s hope it holds out to the expectations. Matroska is a streamable format which is a great thing for live streaming in the future too.

The good stuff about VP8:

It’s a “modern” codec – it’s can actually provide better video quality/compression than the average we currently see and probably will see in the next 2-3 years. That’s sufficiently modern for a lot of uses. It’s not superior to h.264, but it’s supposedly better that just about everything else out there.
It’s “patent-free” – That remains to be proven but if it is, that’s huge. It’ll boost open and closed source, free and commercial software video support to new heights.
It’s “royalty-free” – This is directly dependent on the above. If confirmed this will allow games to fit in lesser number of CD/DVDs as they would not have to use Bink anymore.
There is a “mature encoder” readily available – On2 encoders have evolved for several years now and as VP8 is supposedly based on VP7 the encoder is supposedly mature. This is very important in order to make a codec popular as people actually need to encode their video with it.
Flash will support VP8 – That’s likely 60% penetration in a couple of months after implementation. To measure how big is this, consider that no microsoft codec has ever had such penetration, desipite it’s quality or the amount of money thrown for it’s popularization or the amount of time. People speculate that YouTube could have achieved similar penetration by itself, but that’s just childish. YouTube is a great resource, but it holds just about 25% of the videos played on the internet(which is still HUGE). Point is flash video rules the web for the time being.

The bad stuff about VP8:

No hardware support with VP8 – that’s one of the major achievements with h264. It improves efficiency and allows HD video to be actually played without a supercomputer. VP8 however is not compliant with existing functions and does not provide a way for such in the near future. I don’t really believe that VP8 will get hardware support ever, despite the talks with many hardware vendors, as it’s a question of architecture it lacks and that’s “FINAL”.
The specification is final. – This is really good if it’s a mature and tested specification, however it isn’t and google have already refused to patch up bugs that were found in the few days of it’s public existence. Many developers fear an IE6 like scenario when bugs become part of the specs.

Finally if the patent parts hold out and some flexibility is shown in the “specs are final” part this format may actually conquer the web. I hope it does!

There is a great article by Multimedia Mike about this whole VP8 opening, which is from the unique perspective of one of the developers of Flash at Adobe.

For the people interested in the actual code/algorythm quality a good article was written by Dark Shikari programmer of the x264 open source encoder from the libavcodec(ffmpeg) project.

Bookmark It