A Nice Overview Of The VP9 Codec By A GNOME Developer

[phoronix]

Phoronix: A Nice Overview Of The VP9 Codec By A GNOME Developer

For those interested in learning more low-level details about Google's open-source, royalty-free VP9 video codec, GNOME developer Ronald Bultje has provided a nice overview...

http://www.phoronix.com/scan.php?page=news_item&px=VP9-GNOME-Overview

[karasu]

also related : the AV1 codec from the Alliance for Open Media : http://www.streamingmedia.com/Articl...V1-111497.aspx

[jacob]

also related : the AV1 codec from the Alliance for Open Media : http://www.streamingmedia.com/Articl...V1-111497.aspx

It's not exactly related, AFAIK although some concepts of the VP10 design (which was supposed to be the successor of VP9) have were carried over into AV1. But AV1 is really something to keep an eye on. The open-source, royalty free multimedia stack is almost looking fantastic, we have an excellent lossless audio codec (FLAC), the best lossy audio codec ever (OPUS) and the best container format (Matroska). Now we need a killer video codec. VP9 seems more or less on par with H265 quality-wise (depending on whom one asks) but encoding is unusably slow and it lacks hardware accelerated decoding. If AV1 manages to gain some traction, which it could with Google (i.e. Youtube) and Microsoft on board, we may finally be able to tell MPEG-LA where to stuff their patents.

[geearf]

It's not exactly related, AFAIK although some concepts of the VP10 design (which was supposed to be the successor of VP9) have were carried over into AV1. But AV1 is really something to keep an eye on. The open-source, royalty free multimedia stack is almost looking fantastic, we have an excellent lossless audio codec (FLAC), the best lossy audio codec ever (OPUS) and the best container format (Matroska). Now we need a killer video codec. VP9 seems more or less on par with H265 quality-wise (depending on whom one asks) but encoding is unusably slow and it lacks hardware accelerated decoding. If AV1 manages to gain some traction, which it could with Google (i.e. Youtube) and Microsoft on board, we may finally be able to tell MPEG-LA where to stuff their patents.

You also have lossyflac (or lossywav + flac) for lossy codec, which should work on more platform than opus (not sure of the difference in compression though). As for h265, x265 is not that slow anymore, I'm able to use it up to slow preset, just one preset away from the one I use in x264, but I have no vdpau for it :/

[Serafean]

You also have lossyflac (or lossywav + flac) for lossy codec, which should work on more platform than opus (not sure of the difference in compression though). As for h265, x265 is not that slow anymore, I'm able to use it up to slow preset, just one preset away from the one I use in x264, but I have no vdpau for it :/

I disagree : h265 is painfully slow. I'm trying to use live transcoding, and for that unaccelerated h265 is unusable (2 cores).

btw : any (preliminary) info about AV1 hardware support? When might it come?

[andrei_me]

Michael any news related to Xiph.org Daala?

[erendorn]

Michael any news related to Xiph.org Daala?

Mozilla is now part of the Alliance for Open Media, which is working on the AV1 codec quoted above, which includes bits of Daala (and some bits of Cisco's Thor, but is mostly Google VP10).
As such, Daala will never become a codec by itself (there would be no point). It will probably continue as a research project looking for techs and improvements to include in the future versions of AV1.

[Veerappan]

Interesting read. I am fairly familiar with the internals of VP8, but I hadn't taken the time to dig into VP9 yet (with an eye towards GPGPU acceleration).

[jacob]

You also have lossyflac (or lossywav + flac) for lossy codec, which should work on more platform than opus (not sure of the difference in compression though). As for h265, x265 is not that slow anymore, I'm able to use it up to slow preset, just one preset away from the one I use in x264, but I have no vdpau for it :/

I think the main advantage of lossyflac is that it can be played by any app that supports regular FLAC. AFAIK the compression rates for a given output quality are nowhere near as goog as a true lossy cosec though.

[geearf]

I think the main advantage of lossyflac is that it can be played by any app that supports regular FLAC. AFAIK the compression rates for a given output quality are nowhere near as good as a true lossy codec though.

Makes sense, thanks!