AMD working on XvMC for r300g?

[agd5f]

Meanwhile for me as an end user it looks like none of the HD 4xxx or 5xxx generation cards will be reasonably usable in the near future. Maybe a cast out 3xxx card from an upgrading windows user...

X1xxx = r500
HD2xxx, HD3xxx = r600
HD4xxx = r700
HD5xxx = evergreen

r600 and r700 are both well supported so HD2xxx-HD4xxx cards are working well already. evergreen support is just starting to roll out now.

[BlackStar]

you seem to miss the point?, its just an h.264 clip thats hard to decode, its nothing to do with "A Benchmark" app or whatever.

It's a clip that's intentionally hard to decode, in order to test decoding efficiency. If this doesn't fit the definition of a benchmark I don't know what does.

[Hoodlum]

It's a clip that's intentionally hard to decode, in order to test decoding efficiency. If this doesn't fit the definition of a benchmark I don't know what does.

Really? So you believe the they shot the clip with the intention that it will be hard to decode? That seems completely unrelated to the task of filming a nature documentary to me. I think it's more a case of someone noticing it was hard to decode and using it for that unintended purpose.

[pingufunkybeat]

I think it's more a case of someone noticing it was hard to decode and using it for that unintended purpose.

This is the case with most benchmarks.

[BlackStar]

So you believe the they shot the clip with the intention that it will be hard to decode? That seems completely unrelated to the task of filming a nature documentary to me.

Filming is completely unrelated to encoding. *Completely* unrelated.

You can encode the same clip on dvd quality (plays fine), typical hd quality (~40Mbps max, plays fine) or you can use the current encoding (<=100Mbps) which may fail even on dedicated hardware.

My Nvidia/VDPAU laptop can play every single full-hd movie fine but fails to decode this clip in real time. Does this mean that VDPAU sucks? No, it merely means that this clip is not representative of real-world hd content.

[bridgman]

Well, if you gave me a camera and asked me to film something that would mess up a typical video encode / decode stack...

... it's hard to think of anything worse than birds for blowing out the motion comp

[Hoodlum]

Filming is completely unrelated to encoding. *Completely* unrelated.

You can encode the same clip on dvd quality (plays fine), typical hd quality (~40Mbps max, plays fine) or you can use the current encoding (<=100Mbps) which may fail even on dedicated hardware.

Fair enough but the assertion that this was the intention is still illogical. Why would they wish you a bad experience with the bluray version intentionally?

My Nvidia/VDPAU laptop can play every single full-hd movie fine but fails to decode this clip in real time. Does this mean that VDPAU sucks? No, it merely means that this clip is not representative of real-world hd content.

The problem with this statement is the simple fact that it is real-world hd content. I should know, I own the bluray. I still don't see how this is any different than Prime95 being a superb stress test for an overclocked pc (completely unintentionally).

[Hoodlum]

Just played the clip on my other pc (which is slower than my main PC and not running Linux).

In the test PC I am using:
- Phenom II 2.8ghz Tri-Core
- ATI 5770
- 4GB dd3 ram

Pretty mid-range.

For the test I was using:
- Windows 7
- Catalyst 10.1
- MPlayer with the highest quality post processing (6)

Result:
It actually used less CPU than my own rip of the bluray (26% at most) This clip is using the [email protected] Profile. I use a higher profile for my rips. For comparison trying out a totally unrelated 720p nature program ([email protected] profile) video results in 20% CPU maximum in the couple of minutes I watched it.

The bitrate was posted earlier in the thread - 14.8mbps which is accurate. 14.4 of which is video. 14.8*122(duration)/8=225.7(the file size is actually 216MB). It is actually undersized even for 14.8mbps. This is not at all unrealistic for a bluray (which can go up to 40mbps).

I can provide a screenshot of this with a frame counter if you like.

[Hoodlum]

Well, if you gave me a camera and asked me to film something that would mess up a typical video encode / decode stack...

... it's hard to think of anything worse than birds for blowing out the motion comp

True. It definitely shows you if your acceleration isn't working properly, that's for sure!

[jrch2k8]

True. It definitely shows you if your acceleration isn't working properly, that's for sure!

no, it only shows that your decoder is prepared to a future bitrate standard, up today i havent see any BD or h264 movie encoded like that, is a good bench cuz well it helps you to be future proof, besides the obvious fact that you need one hell of expensive hardware to decode it. so until ssd is the commoners hard disk and we get more processing power in the masses i seriously doubt someone will use that mounstrocity of bitrate to the mass market.

in resume is a good reference not a must, for now. and that now could be quite far in time cuz i really cant see the difference between 1080p and 1080p 100mbit bitrate in my 1080p led 120hz tv, dunno maybe if thaters move 500" led screen maybe but for normal market 1080p is here to stay for many years.