“VP8 is an open source video compression format supported by a consortium of technology companies. This paper provides a technical overview of the format, with an emphasis on its unique features. The paper also discusses how these features benefit VP8 in achieving high compression efficiency and low decoding complexity at the same time.”
I guess the headline should mention it, since VP8 is basically the roots of webm.
Got off of it at the 3rd page
True but the nutshell was pretty much the weighing up of complexity versus other considerations. You can have an incredibly efficient compression but if it so complex that it requires an incredibly powerful CPU any of the efficiencies are off set by increased CPU utilisation. It reminds me very much back in the old days of hard disk based MP3 players and the weighing up of high bit rate music, the disk rotating for longer periods and higher battery use.
I found that VP8 at playback is less power hungry than H.264. Holds true on Windows, Linux, Atom N270, Athlon X2, Pentium M and C2D. By a considerable margin.
That was the idea from the beginning. Which means, if you were watching a VP8-encoded video instead of x264, you *should* be able to get longer battery life out of your notebook, which we all know is critically important for some people.
Unless you have hardware h264 decoding, which many laptops nowadays have. Geforce 8 and up have full hardware h264 decoding, as do 4500MHD, Arrandale and Sandy Bridge GPUs from Intel, and Radeons HD3xxx and up (I think, I don’t have much knowledge about radeons).
Only if you are dumpster diving my friend, as anything built in the last 5 years will have onboard H.264 acceleration which will drop the hardware usage to practically nothing. I don’t know how it is on Linux (and considering the mess with “free as in freedom!” zealotry I wouldn’t be surprised if it doesn’t work) but on Windows DXVA takes care of it in both Nvidia and ATI chips, and I have seen some truly weak systems play 1080p with a crappy $20 GPU card added.
The problem with WebM is it does nothing substantially better and a whole lot worse. it IS patent encumbered (Take one look at the 2000+ H.264 patents and tell me how to make ANY codec that can avoid all 2000) and Google will NOT indemnify so you are hanging in the breeze, it sucks MORE CPU than H.264 for LESS picture quality at anything less than 1080p, which is the vast majority of web videos. Oh and Google like Oracle and Sun are the final say on what is allowed, no “free as in freedom” there.
So in the end the split being rammed home by Google over WebM VS H.264 will leave one winner, and that is Adobe Flash. Flash will play everywhere except on iDevices which will be H.264 (and Adobe is already coming up with a wrapper) and if you think Jobs will EVER let WebM play on iDevices natively I have a bridge to sell you.
Nope, sadly like Theora and Vorbis before it WebM will be a teeny tiny niche, with my prediction that Google will go strictly Flash or H.264 within 2 years. if they do as they’ve been claiming and switch to solely WebM, when so many can’t play it natively or have the codec? Well I predict it’ll be a fine time to invest in one of Youtube’s competitors as there is nothing there that can’t be replicated. Instead of being a day late with a worse format FOSS needs to get ahead of the curve and start patenting 3D video and ultraHD NOW, instead of trying to play a game of catch up years after the boat sailed. If “free as in freedom” was all it would take to sell you’d have Theora and Vorbis players everywhere.
Do you truly think that 3D video was not patented from the moment those horrible red and green glasses were out ?
If you can vaguely picture yourself something and have a lot of money at hand, you can patent it. This is why the current patent system must die.
Oh, and just one thing… I’ve said it before and will say it again, but…
GPU-accelerated H.264 rendering will drop *CPU usage* to nothing, sure. However, considering that the difference between a modern Nvidia GPU powered on and off (no calculation being done, just turn it on and off) is 1h40 vs 2h30 of battery life under normal use, I want to see some numbers to back up the claim that HW acceleration can save power for light tasks like video decoding.
Edited 2011-04-25 10:13 UTC
Oh my, where to begin with correcting all the mistakes? *sigh*
Because a separate computation unit is free, right? Advanced power management means that this unit is powered off when not in use, balancing your calculations a bit.
You just won’t find out about this in your task manager, so it “looks” free.
The problem with WebM is it does nothing substantially better and a whole lot worse. it IS patent encumbered (Take one look at the 2000+ H.264 patents and tell me how to make ANY codec that can avoid all 2000) and
MPEG-LA won’t indemnify you either – they merely have a package to sell. Just look what happens to MP3 player vendors happens every year at Cebit, when Sisvel calls customs to clean up: MP3 is theoretically backed by a patent pool, too.
I doubt that h.264 can steer free of all the TrueMotion/On2/VPx/Google patents, so when MPEG-LA (or any of its partners) go wild, Google can, too.
You can force products out of the market entirely (no forced licensing or any such stupid matter), so if _one_ of the TrueMotion patents apply to h.264, Google could make it drop dead (and vice-versa, of course).
Huh? Oracle and Sun have nothing to do with WebM. And the licensing terms are clear: “free as in freedom” code licensing, clear and permissive patent license that covers said code.
Rhethorical question: What video codec format is “Adobe Flash”?
Rhethorical question again: What video codec format is “Adobe Flash”?
Macs used to support Firewire only. When USB came along, they adopted it (after some bad “look how much it sucks” press by Apple before).
That’s all show for the gullible masses. Should WebM take off, iDevices will support it. Given that they license a whole lot of their SoC functions from elsewhere, the next generation (or the one after that) will probably contain the decoder (and maybe even encoder) already, so they can enable hardware accelerated WebM at any time with a software update to the iOS version running on the iPhone 6 and iPad 4.
Nope, sadly like Theora and Vorbis before it WebM will be a teeny tiny niche, with my prediction that Google will go strictly Flash or H.264 within 2 years. if they do as they’ve been claiming and switch to solely WebM,
In two years Adobe will probably support WebM (not just VP8, but the whole thing) in Flash, providing a suitable fallback in case this is still needed.
They’re not the h.264 police, but support whatever codec brings in money (and h.264, which is the main “Adobe Flash” codec these days costs them, but brings in only so much money).
Except the userbase and the brand. Which is.. uhm.. screw it, you’re right – that’s worthless.
It doesn’t require patents to push some feature “everywhere”. What Theora and Vorbis lack is market power.
For WebM, Google provides reference designs for hardware decoders and encoders for free, liberal/free licensing of the known patent set (which is better than MPEG-LA, where it’s expensive licensing for the known patent set, both are without indemnification), and professional support for companies willing to deploy WebM.
Theora and Vorbis only provide an SVN repository with code. That’s okay for an open source project, but not enough to really nail down a heavily contested market.