UT2K3 DM-Icetomb performance feedback needed ASAP!

[saulin]

OK if you say so

http://members.shaw.ca/navasa/aquamark.jpg


3Dmark nature test without the sooo called optimazations speed drops to hell for the Radeon.

http://www.guru3d.com/forum/showthread.php?s=&threadid=56310&highlight=cheating+3dmark

And come on UT 2K3 is one old benchmark it has been optimized big time by both Nvidia and ATI.


Oh yeah here is an image comparison on

ATI

http://www.guru3d.com/miraserver/images/articles/am3/4XAA-8XAF/Pass01Frame05000RADEON.png

Nvidia

http://www.guru3d.com/miraserver/images/articles/am3/4XAA-8XAF/Pass01Frame05000.png

More discussion on these here:

http://www.guru3d.com/forum/showthread.php?s=&postid=504330#post504330


But of course only Nvidia cheats right?

[reever2]

I hope you asked Balderdash to post that pic from his webspace, and btw go look in the thread where he posted that and was shown what the real problem was

[saulin]

The real problem is that both ATI and Nvidia use the optimizations... in order to achieve better fps or simply better benchmark scores. However most people just accuse Nvidia of cheating.


The real problem is that most people that uses ATI just bitches about Nvidia IQ and as we can see in the screenshot comparison. The Nvidia screenshot looks just darker nothing you can't fix with "Digital Vibrance" and yet the Nvidia shot looks better. AF looks better. Plus is not missing some details that the ATI shot is missing.


So there you have it.


As of right now there is no point for me trying to show you that Nvidia's IQ is not the only one that sucks and that nvidia is not the only one that cheats anymore.

Call it an optimization, bug or cheat. The bottom line is that both companies have some of the three options.

[reever2]

Did you even read the thread where that pic was posted? There is no way to see who is cheating or optimizing because there is no refrast photo

And on the missing objects, how do you know it's not a bug? Every object that is not being rendered on Ati cards all share the same characteristic, they are all under shadows

[saulin]

They are under shadows. Kinna funny isn't it. That way is not soo easy to be spotted.

Anyways I repeat what I just said.

Call it an optimization, bug or cheat. The bottom line is that both companies have some of the three options.

my 2 cents!!

[reever2]

Originally posted by saulin
They are under shadows. Kinna funny isn't it. That way is not soo easy to be spotted.


Dude wtf are you talking about? There is nothing to spot, the items are MISSING, anybody can clearly see what is not being rendered properly

[Spotch]

Originally posted by saulin
The real problem is that both ATI and Nvidia use the optimizations... in order to achieve better fps or simply better benchmark scores. However most people just accuse Nvidia of cheating.


The real problem is that most people that uses ATI just bitches about Nvidia IQ and as we can see in the screenshot comparison. The Nvidia screenshot looks just darker nothing you can't fix with "Digital Vibrance" and yet the Nvidia shot looks better. AF looks better. Plus is not missing some details that the ATI shot is missing.


So there you have it.


As of right now there is no point for me trying to show you that Nvidia's IQ is not the only one that sucks and that nvidia is not the only one that cheats anymore.

Call it an optimization, bug or cheat. The bottom line is that both companies have some of the three options.


I am very confused at the logo on the starboard fender. I have run the demo and in that shot and previous shots the little orange logo is drawn consistently. It never looks like the NV shot. In fact it looks completely different.
Whats confusing is that it does not look like a blurry or less detailed version, it IS a different logo and has a distict shape and color. Now could this be an alternate texture layer that is not detected or drawn by the NV card or vice versa meaning it is the ATI card not drawing or detecting the texture displayed by the NV card? We must find out.

[fivefeet8]

Come on. I want more people to look at the shots and say which drivers they are from.

[Spotch]

Originally posted by fivefeet8
Come on. I want more people to look at the shots and say which drivers they are from.

Your shots are flawed. It is easy to tell because the two driver versions render the health power ups very differently. Notice one is blue and the other is a very pale blue.

PS: sorry to ruin it for you... :D

[fivefeet8]

Originally posted by Spotch
Your shots are flawed. It is easy to tell because the two driver versions render the health power ups very differently. Notice one is blue and the other is a very pale blue.

PS: sorry to ruin it for you... :D

What are you talking about.. Which is which driver then if it's so easy? Actually, when the health power ups spin, they look different. Guess you didn't ruin it. And your neglecting to make a decision as well. So which driver goes with which shot?

[Spotch]

Originally posted by fivefeet8
What are you talking about.. Which is which driver then if it's so easy? Actually, when the health power ups spin, they look different. Guess you didn't ruin it. And your neglecting to make a decision as well. So which driver goes with which shot?

Perhaps I was mistaken. I just ran the game and noticed that they do change color when they spin. I just brought up the glaring difference in your screenshot thinking that anyone with an FX or ATI card could tell by running the game themselves.

Perhaps if you created identical screencaptures it would help people compare the two. Also the screenshots are no longer accessible. I could not see any real differences in image quality myself, however it is not the screenshots that look the same that is of importance here, it is the ones that are different.

[fivefeet8]

Originally posted by Spotch
Perhaps if you created identical screencaptures it would help people compare the two. Also the screenshots are no longer accessible. I could not see any real differences in image quality myself, however it is not the screenshots that look the same that is of importance here, it is the ones that are different.

Yeah, free geocities accounts have very small bandwidth limits per hour. I could make identical screen captures, but I was trying to test the claim made earlier by other people that saw differences in IQ from similar screenshots. Not identical. And has been noted numerous times, the only thing evidently different is the missing roof fog. Which by all accounts could very well be a driver bug and not an optimization or cheat.

Come on. Those before who said they saw noticeable texture differences, choose.

AFter which, I'll post up identical shots with different drivers.

[StealthHawk]

Originally posted by saulin
BTW Nvidia isn't the only one cheating.

ATI cheats in 3DMark and cheats in Aquamark and in UT 2K3 just so you know.

ATI cheats in like one game test(Nature) in 3dmark2001(who really cares? Is it even worth removing). Of course you beloved NVIDIA cheats in the Nature test...and the other tests too.

There is no proof that anyone is cheating in Aquamark. If anyone is cheating it's NVIDIA.

ATI doesn't cheat in UT2003 :rolleyes: Only NVIDIA does.

[fivefeet8]

Part of the experiment is over. I posted the same shots and topic at another message board and the answers were the same. When you already believe the 51.75's reduce image quality for performance, your probably going to see IQ loss even when there probably isn't any. As evident by this experiment, all the ones who chose the First DM-icetomb shot as the 51.75's were wrong. And anyone who chose the second Ice-tomb shot as 51.75's are also wrong. Both were done with 45.23 drivers.

The two rustatorium shots were done with different drivers though. But it's funny how hard it is to choose drivers for those two.

[Flavius]

You bastard! We should flay you alive for this! This is enough to make me post at Rage3D exclusively.

[hovz]

well if they were both 45.23 then u also havernt proved that there isnt a loss in image quality

[Spotch]

Originally posted by fivefeet8
Part of the experiment is over. I posted the same shots and topic at another message board and the answers were the same. When you already believe the 51.75's reduce image quality for performance, your probably going to see IQ loss even when there probably isn't any. As evident by this experiment, all the ones who chose the First DM-icetomb shot as the 51.75's were wrong. And anyone who chose the second Ice-tomb shot as 51.75's are also wrong. Both were done with 45.23 drivers.

The two rustatorium shots were done with different drivers though. But it's funny how hard it is to choose drivers for those two.

Since I am not interested in proving my loyalty to one card or another I am immune to your test. (I did say after all that the Ice shots looked the same.) The problem here is not to show my card is better than yours. It is to show that your card should be better. NV has misled and cheated the most important people of all in my opinion; THE CUSTOMER!

What we all want to do, is to force NV to be accountable for their misrepresentations regarding the entire FX line. DX 9 ? I dont think so! Cine FX? Not really. Industry standard Detonator Drivers? Bull-cookies!

They need to pull back and release a driver that does what the game tells it to do. Then release some hardware that can actually perform acceptably in those circumstances. Then offer a recall. All FX customers that request it should get a replacement part that works as advertised!

[OWA]

BTW, what is the best way to get identical screencaps between different systems? I was trying to do that last night with MOHAA Spearhead to demonstrate a bug with ATI and shadows enabled. I had a really hard time getting one that was identical...actually, I never did I just got one that was semi-close before giving up for the evening.

[fivefeet8]

Originally posted by hovz
well if they were both 45.23 then u also havernt proved that there isnt a loss in image quality

But it also doesn't prove that there was a loss either.

Remember, I said I'd post up identical shots soon. ;) The previous was just an experiment to test IQ quality subjectivity. It proves that even though many of you claimed noticeable IQ loss, it might just have been the slight shift in the scene. And even with the same drivers, you claimed noticeable IQ loss. Why is that? Because you already believe it before you look. Your expecting IQ loss, looking for it, and every little thing that is different, must be it.

OWA - YOu could make a demo with the game where all you do is stand at a certain place for a few seconds to take a screenshot. Then copy that demo into the other computer and run it. Run the demo on both computers and take screenshots. That's what I've done in Ut2k3 in the upcoming shots.

[fivefeet8]

Originally posted by Spotch
Since I am not interested in proving my loyalty to one card or another I am immune to your test. (I did say after all that the Ice shots looked the same.) The problem here is not to show my card is better than yours. It is to show that your card should be better. NV has misled and cheated the most important people of all in my opinion; THE CUSTOMER!


Nor am I, but this wasn't really the purpose of the test. And it wasn't to show which card is better.

Anyways, I posted identical shots from both driver sets. If the site is down, wait a bit. Damn bandwidth limitations. :D

[Spotch]

Originally posted by fivefeet8
But it also doesn't prove that there was a loss either.

Remember, I said I'd post up identical shots soon. ;) The previous was just an experiment to test IQ quality subjectivity. It proves that even though many of you claimed noticeable IQ loss, it might just have been the slight shift in the scene. And even with the same drivers, you claimed noticeable IQ loss. Why is that? Because you already believe it before you look. Your expecting IQ loss, looking for it, and every little thing that is different, must be it.

OWA - YOu could make a demo with the game where all you do is stand at a certain place for a few seconds to take a screenshot. Then copy that demo into the other computer and run it. Run the demo on both computers and take screenshots. That's what I've done in Ut2k3 in the upcoming shots.

Another problem with your shots is that the walls had running water. This made for a very inconsistent basis for comparisons. Also make sure you set your drivers to allow the application to determine Aniso and AA settings. This is one area the affected by the driver hacks made by NV.

PS: Please do not use hacked up Omegas. We are not interested in pot luck drivers only in what NV has released and are in question here.

[hovz]

there is proven IQ loss, there is no trilinear in ANY d3d games

[Spotch]

Originally posted by hovz
there is proven IQ loss, there is no trilinear in ANY d3d games

Yeah that used to be the primary gripe about ATI cards in many games.

[fivefeet8]

Originally posted by hovz
there is proven IQ loss, there is no trilinear in ANY d3d games

Weren't we talking about IQ loss in Ut2k3 in the 51.75's from the 45.23's? :rolleyes:

[StealthHawk]

Originally posted by fivefeet8
Weren't we talking about IQ loss in Ut2k3 in the 51.75's from the 45.23's? :rolleyes:

Actually, are we?