ati cheating in crysis benchmark?

[flippin_waffles]

that 24 aa is super crazy just love it best aa mode imho as for comparison we need close ups its impossible to understand this way they all look same

yep, superior aa for sure. BTW, have you seen these? af is incredible!


GTX 480


5870

I wonder if nv will ever be able to get close to this kind of AI AF. They've got A LOT of catching up to do!

[Manicdan]

how come no one builds a test that properly shows AF? is it that hard to have a pattern on different shapes and it shows you how each shape handles it?

[Chickenfeed]

The irony is ATI has their driver setting default to max texture IQ yet Nvidias dont. If anything by default Nvidia is cutting themselves short in the IQ deparment ( but my bet is this is done for performance reasons which could be considered equally retarded ) but as mentioned in the article they offer comparable texture quality at their default setting of "quality". The moment you use the "high quality" preset, the ball is in their court ( not by much but it is regardless ), simple enough. Ive owned / used / tested more Nvidia cards in the last 2 years ( although the bulk of my personal time has been with a 4870x2 and my current 5870 ) and I'd argue subjectively that even the GT200 cards provide better texture quality (AA quality is another debate of which I still feel this is in ATIs court but thats neither here nor there ). Now I'd wager if someone were to test a range of games, and focus on AF / texture quality alone, that you'd see the differences are subtle, if there at all, most of the time ( excluding this obvious concern with regards to Crysis )

What I want to know now is if this effects both Warhead and the first game ( don't have warhead installed currently )

[Manicdan]

i use to use atitool or ati tray tools, whichever it was, to turn mipmap detail to the lowest setting to get an extra few hundred points out of 3dmark06, even though it looks like crap

[Jowy Atreides]

how come no one builds a test that properly shows AF? is it that hard to have a pattern on different shapes and it shows you how each shape handles it?

Done years ago.
Aftester.exe

allows custom image af testing.

[Manicdan]

Done years ago.
Aftester.exe

allows custom image af testing.

so i guess no one uses it then huh, ill test it out when i get home see how my 4000 card does

[ElSel10]

I wonder if nv will ever be able to get close to this kind of AI AF. They've got A LOT of catching up to do!

Yeah, even though their AF looks better in practice, they have a lot of catching up to do.

[Jowy Atreides]

so i guess no one uses it then huh, ill test it out when i get home see how my 4000 card does

...umm, that's exactly what was used in the comparison pictures

Yeah, even though their AF looks better in practice, they have a lot of catching up to do.

No.

It looks worse in practice.

[ElSel10]

No.

It looks worse in practice.

You couldn't be more wrong. This needs to stop because it's getting annoying.

RV870 @ 16x AF + 8x SGSSAA:



GF100 @ 16x AF + 4x4 OGSSAA:



Not only is that particular area noticeably better on GF100 at highest quality AF setting vs. the RV870, but the entire image. If you go to this page, it allows you to compare the images by putting them on top of each other, and you can clearly see the image is better on GF100. So please, stop saying it isn't.

ATI has sacrificed filtering quality in order for their angle independent algorithm to work with a reasonable hit on performance, and this translates to lower overall AF quality than GF100 in practice.

[Chickenfeed]

Do note super sampling AA effects texture quality so making any texture quality comparisons when super sampling is used is somewhat pointless. AF comparisons are best made without any form of anti aliasing.

[Chumbucket843]

yep, superior aa for sure. Btw, have you seen these? Af is incredible!


I wonder if nv will ever be able to get close to this kind of ai af. They've got a lot of catching up to do! :d

lol. nvidia is really bad at games that are tunnels of colored circles.

how come no one builds a test that properly shows af? Is it that hard to have a pattern on different shapes and it shows you how each shape handles it?

i have done an AF test here with a 8192x8192 res texture. nvidia wins for about 90% of angles. not until it is extremely angled does ATi look better.
http://www.xtremesystems.org/forums/...7&postcount=12

[flippin_waffles]

You couldn't be more wrong. This needs to stop because it's getting annoying.

RV870 @ 16x AF + 8x SGSSAA:

[*img]http://img15.imageshack.us/img15/4818/16xaf8xsgssaahd5870.png[/img]

GF100 @ 16x AF + 4x4 OGSSAA:

[*img]http://img687.imageshack.us/img687/4026/64xafvia4x4ogssaa.png[/img]

Not only is that particular area noticeably better on GF100 at highest quality AF setting vs. the RV870, but the entire image. If you go to this page, it allows you to compare the images by putting them on top of each other, and you can clearly see the image is better on GF100. So please, stop saying it isn't.

ATI has sacrificed filtering quality in order for their angle independent algorithm to work with a reasonable hit on performance, and this translates to lower overall AF quality than GF100 in practice.

Hmm, looks like the image is better on the 5870 than the 400.

[XS2K]

Hmm, looks like the image is better on the 5870 than the 400.

We must be looking at different images...
Take down your ATi goggles

[annihilat0r]

Hmm, looks like the image is better on the 5870 than the 400.

[zanzabar]

Hmm, looks like the image is better on the 5870 than the 400.

i concur, the gf100 looks smoother but the closer section is not smooth so it looks strange (for the above images)

on a side note why dose it even matter i would bet that no1 who is arguing ether way would notice this while playing, and its subjective on what looks better

[Final8ty]

What a sad lot making such a big fuss over tiny differences as if you will notice in game as i thought most games are not a about looking at paused stills zoomed.
What a bunch of girls.

[Luka_Aveiro]

What a sad lot making such a big fuss over tiny differences as if you will notice in game as i thought most games are not a about looking at paused stills zoomed.
What a bunch of girls.

So, if it was nVidia cheating here, you would be saying:

KILL THEM ALL!!!

[Sadasius]

So, if it was nVidia cheating here, you would be saying:

KILL THEM ALL!!!

Nope....Only their families!

[Final8ty]

So, if it was nVidia cheating here, you would be saying:

KILL THEM ALL!!!

I said nothing about cheating, I would not care either way if it was NV seeing as your asking because you are noway telling more or thinking you can put words in my mouth would you now.
All i see is a bunch of girls arguing over the slight shades of make-up.
Some of you seem to have ATI v NV stamped in your forehead as that's all you seem to see & say no matter what is said.

[Sadasius]

I said nothing about cheating, I would not care either way if it was NV seeing as your asking because you are noway telling more or thinking you can put words in my mouth would you now.
All i see is a bunch of girls arguing over the slight shades of make-up.
Some of you seem to have ATI v NV stamped in your forehead as that's all you seem to see & say no matter what is said.

[flippin_waffles]

LMAO


Just like I said, I think ATi has a better image than nv in that screen. Looks more realistic.

[cegras]

Oh my god the differences are infinitesimally minute. You people are insane.

Someone give me a detailed explanation of why the GF100 image looks better in the red circle.

[ElSel10]

Oh my god the differences are infinitesimally minute. You people are insane.

Someone give me a detailed explanation of why the GF100 image looks better in the red circle.

Go on the website and compare the two images directly on top of each other and it would be obvious, as would the whole image, the textures are not as blurred.

As for the red circle, the black line between the tiles is straighter and not as blurry as it is in the ATI image. Seriously, go compare them on top of each other as you are supposed to do and it will be obvious. As would the image of the ground texture in Crysis being sharper and crisper on GF100 and the zones being not nearly as visible.

[zanzabar]

that line shouldn't be straight its a jagged line of bricks with offsets, like hand made walkway. if u look one has integrity and one is a blurred straight line that looks out of place. in both the mortar line is close to content width but on the NV its missing curves that classic hand made and laid masonry would have. but its subjective, to really compare u need something with clean straight lines like a flight sim

[flippin_waffles]

It looks like nv might be cheating and not rendering that joint correctly.

[edit]