Flash Player 10.1 RC Arrives But Still Not In Tune

[dacresbu]

video accel != graphics accel

I've apparently completely missed on these threads the reason for OpenGL video accel rendering not being a good API in comparison to the others. Windows almost exclusively uses D3D for video accel, don't they?

it sounds like your confusing video decoding acceleration with graphics acceleration.

I'm just surprised then that OGL hasn't tried to become the dominant video accel API in response to that. What does VA and VDPAU have that OGL lacks?
.

I guess you could use OpenCL but not openGL, i dont see how there's anything in there that could be good for video decoding or incoding. SIMD would be good for this, they already published libraries to access the CPU SIMD capacity with SSE (altivec on PPC).

Nvidia and ATI published their own api's for video acceleration. there is only one in windows but that's because DriectX does more than graphics acceleration, the GL in openGL stands for Graphics Language. DirectX does more. Apple has their own video decoding library

[bridgman]

Yeah, the important distinction is between D3D (which is not used for video very much) and DirectX, which includes D3D, DXVA (the video acceleration stuff) and a bunch of other things starting with D.

I imagine OpenGL could be useful for the motion comp and filtering parts of decode, and you could probably do IDCT with OpenGL but not sure it would be worth it. I haven't seen any good algorithms to make the entropy decode step (CABAC etc.) make good use of massively parallel hardware.

[Yfrwlf]

it sounds like your confusing video decoding acceleration with graphics acceleration.

I guess you could use OpenCL but not openGL, i dont see how there's anything in there that could be good for video decoding or incoding. SIMD would be good for this, they already published libraries to access the CPU SIMD capacity with SSE (altivec on PPC).

Nvidia and ATI published their own api's for video acceleration. there is only one in windows but that's because DriectX does more than graphics acceleration, the GL in openGL stands for Graphics Language. DirectX does more. Apple has their own video decoding library

You're right and I'm surprised then that Khronos hasn't tried to expand, not that everything needs to be under one umbrella but...might help. :P Or maybe include or support particular audio or video accel APIs to help push a total package like DirectX has become? *shrugs*

Well, as the article says, if VA and VDPAU really are becoming strong standards then all the power to them as long as they're good open standards...

Yeah, the important distinction is between D3D (which is not used for video very much) and DirectX, which includes D3D, DXVA (the video acceleration stuff) and a bunch of other things starting with D.

I imagine OpenGL could be useful for the motion comp and filtering parts of decode, and you could probably do IDCT with OpenGL but not sure it would be worth it. I haven't seen any good algorithms to make the entropy decode step (CABAC etc.) make good use of massively parallel hardware.

VLC has "OpenGL video output" as an output option, is where I got that from. And you're right, I abbreviated DirextX as D3D when I meant DX hehe. :P

[gbeauche]

It will do away with the YouTube dilemma except for the fact that the patented H264/MP4 codec is what Google wants to use for YouTube, which will never be supported by Firefox. It's a shame Google is supporting that closed standard instead of pushing for and helping alternatives like Dirac, Snow, and Theora out, not to mention pushing for the abolishment of software/math/art/idea patents in the U.S..

You know, Google engineers are not stupid, they won't use inferior codecs that would increase their costs...

You're right and I'm surprised then that Khronos hasn't tried to expand, not that everything needs to be under one umbrella but...might help. :P Or maybe include or support particular audio or video accel APIs to help push a total package like DirectX has become? *shrugs*

Khronos has OpenMAX specifications but this is mostly used in the embedded space (TI, NVIDIA Tegra, etc.).

Well, as the article says, if VA and VDPAU really are becoming strong standards then all the power to them as long as they're good open standards...

How can they be more open than being Freedesktop standards?

[smitty3268]

You know, Google engineers are not stupid, they won't use inferior codecs that would increase their costs...

It's funny you picked today to post that. Google is clearly looking closely at Theora, even if they haven't made any commitment to using it yet.

http://google-opensource.blogspot.com/2010/04/interesting-times-for-video-on-web.html

What is clear though, is that we need a baseline to work from - one standard format that (if all else fails) everything can fall back to. This doesn?t need to be the most complex format, or the most advertised format, or even the format with the most companies involved in its creation. All it needs to do is to be available, everywhere. The codec in the frame for this is Ogg Theora, a spin off of the VP3 codec released into the wild by On2 a couple of years ago. It scores quite well on both the quality and compression fronts, standing up respectably against it?s more popular rivals such as MPEG4, while actually being much simpler to decode. The overwhelming feature that makes it stand out from its rivals is the fact it?s free. Really free. Not just ?free to use in decoders," or ?free to use if you agree to this complicated license agreement," but really, honestly, genuinely, 100% free.

This is where Google's grant comes in - by helping fund the development of TheorARM (a free optimised ARM version of Theora), they are helping to hasten the day when video works everywhere on the web, for everyone.

[Yfrwlf]

It's funny you picked today to post that. Google is clearly looking closely at Theora, even if they haven't made any commitment to using it yet.

http://google-opensource.blogspot.co...eo-on-web.html

Good to hear, I guess actions will be louder than words once Google Chrome finally incorporates support for Theora, even though it's not the best codec as mentioned by gbeauche, since it is still 100% free and BECAUSE of that deserves to be the fallback or actual real standard which is used when there is doubt about the patented algorithms.

It's good to see Google support Theora, but what I'm also saying is they should look into better algorithms and codecs to support as well because I thought that Dirac and Snow were a) better than Theora and b) also free, unpatented math.

Of course, Google should also help fight software patents, too, as they are monopolistic, horrible, wasteful, stupid things.