Unreal 'optimization' a non-issue?

[Ruined]

http://www.hardocp.com/article.html?art=NTAw

HardOCP just did an in depth filtering comparison of the 5900Ultra 44.03 vs. 9800PRO in UT, and came up with the following conclusion:

"In-Game Still Screenshot No AF = 5900 Ultra SLIGHTLY Sharper Textures with no mipmap transitions that are distracting.

In-Game Still Screenshot 8XAF = 9800 Pro SLIGHTLY Sharper Textures with no mipmap transitions that are distracting.

In-Game Movement No AF = Can’t tell any difference.

In-Game Movement 8XAF = Can’t tell any difference."

...

So, basically, it appears the optimization many have been complaining about is a non-issue.

[extreme_dB]

What Unreal optimization issue? Nevermind, it's a non-issue being discussed in countless threads across the internet.

[reever2]

Originally posted by Ruined
http://www.hardocp.com/article.html?art=NTAw

So, basically, it appears the optimization many have been complaining about is a non-issue.

If you take one sites word over everybody elses, sure...

ahh barons post just went away

I thought we had 60 bajillion posts about this. And we probably do.

Oh well.

[Ruined]

Originally posted by reever2
If you take one sites word over everybody elses, sure...

ahh barons post just went away

Well, I think HardOCP is coming up with a sensible conclusion - even if there are minor differences in filtering (slightly favoring fx5900 w/o AF, slightly favoring 9800 w/AF), you can only see those differences with zoomed in Photoshopped screenshots, but in actual gameplay (according to them), you can't tell the difference. In this case, they seemed to go to great lengths to detect differences as well, but in the end could not during gameplay. Therefore, it's a moot point.

HardOCP has been very pro ATI in the past so I wouldn't say it's an nvidia biased site, though they seem to be approaching it in a sensible way - compare the games to each other while you are actually playing, as opposed to studying photoshopped screencaps...

And yeh there are a lot of posts but I thought this was a pretty cool article because its a brand new article done by one of the 'big' sites, and is very in depth. Plus, it specifically addressed the filtering issue and optimizations everyone has been talking about recently. I know I read in many threads people were stating that Nvidia was 'cheating' in UT and/or offering inferior filtering, but this article is a good counter for that point, since their direct A/B comparisons of actual gameplay found no noticable differences, and direct A/B comparisons of screenshots they studied went one way w/o AF and the other with AF.

In other words, both cards play Unreal just as good regardless of how the drivers work.

[reever2]

Originally posted by Ruined
Well, I think HardOCP is coming up with a sensible conclusion - even if there are minor differences in filtering (slightly favoring fx5900 w/o AF, slightly favoring 9800 w/AF), you can only see those differences with zoomed in Photoshopped screenshots, but in actual gameplay (according to them), you can't tell the difference. In this case, they seemed to go to great lengths to detect differences as well, but in the end could not during gameplay. Therefore, it's a moot point.

HardOCP has been very pro ATI in the past so I wouldn't say it's an nvidia biased site, though they seem to be approaching it in a sensible way - compare the games to each other while you are actually playing, as opposed to studying photoshopped screencaps...

It has been said over and over on numerous sites how they chose a place to take a pic of that would make the situation less apparent. It is clear as day on any of the shots Beyond3d posted. And dont talk about them being sensible. Is it sensible to ban people who do nothing wrong like DaveBaumman(?) and doomtrooper?

[Sazar]

Originally posted by Ruined
Well, I think HardOCP is coming up with a sensible conclusion - even if there are minor differences in filtering (slightly favoring fx5900 w/o AF, slightly favoring 9800 w/AF), you can only see those differences with zoomed in Photoshopped screenshots, but in actual gameplay (according to them), you can't tell the difference. In this case, they seemed to go to great lengths to detect differences as well, but in the end could not during gameplay. Therefore, it's a moot point.

HardOCP has been very pro ATI in the past so I wouldn't say it's an nvidia biased site, though they seem to be approaching it in a sensible way - compare the games to each other while you are actually playing, as opposed to studying photoshopped screencaps...

And yeh there are a lot of posts but I thought this was a pretty cool article because its a brand new article done by one of the 'big' sites, and is very in depth, and specifically addressed the filtering issue and optimizations everyone has been talking about recently. I know I read in many threads people were stating that Nvidia was 'cheating' in UT and offering inferior filtering, I would say this article is a good counter for that point.

their conclusion is flawed... comparng bilinear/trinlinear filtering is the problem that is being raised... and has been proven in screenshots other than [H]'s across the web...

have a read @ www.amdmb.com for their version of events.. even kyle linked to it in the thread started concernig this topic on [H]

[Ruined]

Originally posted by reever2
It has been said over and over on numerous sites how they chose a place to take a pic of that would make the situation less apparent. It is clear as day on any of the shots Beyond3d posted. And dont talk about them being sensible. Is it sensible to ban people who do nothing wrong like DaveBaumman(?) and doomtrooper?

Yeh, but is that sensible, finding a particular screenshot that when studied might show some differences for one card or another? Maybe in another screenshot the Nvidia output would look better than ATI? If HardOCP ran through the levels on both cards and couldn't see any difference, isn't that really what matters, since playing Unreal consists of running through levels shooting people, not taking screencaps of, studying, and zooming in on far away textures?

You can kind of draw a (somewhat faulty but what the heck) analogy to DVD. DVD uses MPEG2 compression, and if you pause a frame of a DVD, you can often study it and find some sort of artifacting, be it edge enhancement, haloing, mpeg blocking, etc... However, when played back, you generally don't notice such things (unless the DVD is mastered very poorly). So, although you aren't getting the full uncompressed perfect image quality, and although when you pause the DVD you might be able to find flaws in particular frames, when you are actually watching the movie you don't notice these things, so they are irrelevant. What I'm saying is, if it takes studying a zoomed in still frame to find a difference in an in-game image, its rather irrelevant since you would never be doing that when actually playing the game.

I demand absolute fidelity in audio or video. I don't compromise it.

Most people who buy high-end cards agree with me. If they're going to do it, they should at least give me a choice.

[reever2]

Originally posted by Ruined

You can kind of draw a (somewhat faulty but what the heck) analogy to DVD. DVD uses MPEG2 compression, and if you pause a frame of a DVD, you can often study it and find some sort of artifacting, be it edge enhancement, haloing, mpeg blocking, etc... However, when played back, you generally don't notice such things (unless the DVD is mastered very poorly). .

That analogy doesnt work in this situation, it is the other way around. Mip map transition boundries, like aliasing is MORE apparent when in motion and to many is quite annoying

[GlowStick]

Actually im surpzied we havent seen any threads about ATi pushing for HardOCP to change their review.

Isnt nvidia handleing that waaaaaay better than ATi?

[Ruined]

Originally posted by Sazar
have a read @ www.amdmb.com for their version of events.. even kyle linked to it in the thread started concernig this topic on [H]

But that site only deals with the FX5600 - it is possible Nvidia used a different optimization for the FX5600 hardware to make it more playable with UT2k3 since it does not have the fillrate or memory bandwidth of the FX5900.