OMG!!! ATIisntCHEATING!!!

[VIAN]

There has been a lot of discussion about our trilinear filtering algorithms recently.

The objective of trilinear filtering is to make transitions between mipmap levels as near to invisible as possible. As long as this is achieved, there is no right or wrong way to implement the filtering.

We have added intelligence to our filtering algorithm to increase performance without affecting image quality. As some people have discovered, it is possible to show differences between our filtering implementations for the RADEON 9800XT and RADEON X800. However, these differences can only be seen by subtracting before and after screenshots and amplifying the result. No-one has claimed that the differences make one implementation better than another.

Our algorithm for image analysis-based texture filtering techniques is patent-pending. It works by determining how different one mipmap level is from the next and then applying the appropriate level of filtering. It only applies this optimization to the typical case specifically, where the mipmaps are generated using box filtering. Atypical situations, where each mipmap could differ significantly from the previous level, receive no optimizations. This includes extreme cases such as colored mipmap levels, which is why tests based on color mipmap levels show different results. Just to be explicit: there is no application detection going on; this just illustrates the sophistication of the algorithm.

We encourage users to experiment with moving the texture preference slider from Quality towards Performance you will see huge performance gains with no effect on image quality until the very end, and even then, the effect is hardly noticeable. We are confident that we give gamers the best image quality at every performance level.

Microsoft does set some standards for texture filtering and the companys WHQL process includes extensive image quality tests for trilinear filtering and mipmapping. CATALYST passes all these tests and without application detection, which could be used if you wanted to get a lower-quality algorithm go undetected through the tests.

Finally, ATI takes image quality extremely seriously and we are confident that we set the bar for the whole industry. We dont undertake changes to our filtering algorithms lightly, and perform considerable on-line and off-line image analysis before implementing changes. This algorithm has been in public use for over a year in our RADEON 9600 series products, and we have not received any adverse comments on image quality in that time. If anyone does find any quality degradation as a result of this algorithm, they are invited to report it to ATI. If there is a problem, we will fix it.

 

[Dman877]

ATI press-release? It's gotta be true .

I'm glad I don't waste my life examining screenshots to find tiny flaws.

EDIT: Of course I'm glad someone does it to keep the corporate goons honest .

 

[VIAN]

This is ATI's response to Driver Heaven's inquiry.

Makes sense.

I feel bad for Nvidia. ATI are better business people-they know how to persuade.

 

[imported_Aelius]

Assuming this is true, it works for me.

I don't care what they do as long as it doesn't hurt image quality that I can see.

 

[Robor]

This "cheating" crap is getting old. I think this forum if full of fanboys who couldn't tell one manf from another in an identical rig at real game speed. If that's the case then all this back and forth is a waste of time.

Both sides "tweak" their drivers to run their hardware at the best IQ/speed as possible, period.

 

[BFG10K]

Our algorithm for image analysis-based texture filtering techniques is patent-pending. It works by determining how different one mipmap level is from the next and then applying the appropriate level of filtering.

That's exactly what I suspected was happening and it certainly explains all of the evidence we've seen so far. This is not a cheat.

 

[VIAN]

Did you also suspect this when Nvidia did it? - even though they never bothered to tell us exactly what they were doing.

 

[BFG10K]

There was evidence before that nVidia was doing it because some of their colourmip benchmarks were lower too and Carmack also commented that both vendors were doing it.

 

[Shamrock]

it degrades IQ, it is a cheat. ATI is trying to cover it up. over at NVNEWS.net they have several posts about it, and even the ATI "pre-release" meeting documents in pdf format, explaining "tricks" on AF filtering. Yes, I linked to nvnews, but that doesnt make it a biased post, the forums are littered with ATI folks, and they even feel betrayed.

You cant have Tri-linearing filtering with this optimization. Not even a selection to toggle it on or off.

 

[VIAN]

It doesn't lower IQ, it just makes them do less transitions, thereby, cutting the workload.

 

[Description]

link?

 

[Description]

We encourage users to experiment with moving the texture preference slider from Quality towards Performance

There is a way to turn it off.

 

[VIAN]

BTW, I JUST MADE THE ARTICLE UP. THERE IS NO OFFICIAL RESPONSE.

 

[AnnoyedGrunt]

I think the issue is that even in "Quality" mode it isn't using Trilinear filtering. Instead, it will automatically choose based on it algorithm. Therefore, you never know quite what you'll get.

I don't have a problem with this, as long as reviews inform you when they are testing AF speed and as long as ATI advertizes the "feature". The fact that ATI has kept it hidden does not make it seem like a great new technology, but rather a way to make people think they are getting one thing but instead give them another.

From a users standpoint, the technology is very good, since you don't really know the difference and because the IQ is comparable to the highest quality images. However, from a benchmarking perspective it really throws off the results when comparing the X800 to the 6800. The X800 gained it's advantage when AF was enabled, but perhaps it was because you were never getting Trilinear filtering like you were on the 6800.

I'd be interested in seeing benchmarks with the optimized filtering algorithms on both cards to get a more "apples-to-apples" comparison.

IMO, the raw power of the X800 series is less impressive now that we know it wasn't running the same tests. However, the fact that you can get the "same" IQ and great FPS is still very nice.

I would like to see ATI include a way to actually force Trilinear if you wanted to, since you currently can't do that. If they did that, then I think people would have very little reason to complain.

-D'oh!

 

[Shamrock]

who cares if it's an "official response" as said in my thread from Chksix that it's PR BS, you actually think they would admit to it? and as far as lowering IQ, according to computerbase.de it DOES lower IQ

 

[Shamrock]

you are correct AnnoyedGrunt, even in high quality mode, it does NOT do FULL tri-linear filtering

 

[alexXx]

annoyedGrunt, it is clear that you are an NVIDIA fanboy. If they can create the same result as trilinear filtering with a more refined algorithm, then excellent and more power to them. They claim to pass MS's trilinear test or watever, so that is good enough for me, and should be good enough for you.

 

[BFG10K]

Have people missed that fact that nVidia does the same thing?

If you want to slam it as a cheat then fine but be sure to slam both vendors instead of just singling out ATi.

Like I said before, I don't have a problem with either vendor doing it.

 

[VIAN]

I'd be interested in seeing benchmarks with the optimized filtering algorithms on both cards to get a more "apples-to-apples" comparison.

There are benches out there where Nvidia's card say "optimizations on" - that's "brilinear" - same as ATI.

 

[Shamrock]

Originally posted by: BFG10K
Have people missed that fact that nVidia does the same thing?

If you want to slam it as a cheat then fine but be sure to slam both vendors instead of just singling out ATi.

Like I said before, I don't have a problem with either vendor doing it.

Both vendors might do it, but at least NVidia doesnt deny it, nor try and hide it. The drivers actually turn it off and on automatically when an application is detected. :\ at least with NV's drivers there is an option to turn it off by going to high quality.

and now you actually admit it's a cheat...it's just "you have no problem with it now" except for the 400+ posts arguing with Rollo about it

 

[Ackmed]

Originally posted by: Shamrock
it degrades IQ, it is a cheat. ATI is trying to cover it up. over at NVNEWS.net they have several posts about it, and even the ATI "pre-release" meeting documents in pdf format, explaining "tricks" on AF filtering. Yes, I linked to nvnews, but that doesnt make it a biased post, the forums are littered with ATI folks, and they even feel betrayed.

You cant have Tri-linearing filtering with this optimization. Not even a selection to toggle it on or off.

Show me a pic where it degrades IQ?

btw, just because its posted at nvnews, doesnt mean anything. That site is worthless.

 

[nemesismk2]

Originally posted by: Ackmed

Originally posted by: Shamrock
it degrades IQ, it is a cheat. ATI is trying to cover it up. over at NVNEWS.net they have several posts about it, and even the ATI "pre-release" meeting documents in pdf format, explaining "tricks" on AF filtering. Yes, I linked to nvnews, but that doesnt make it a biased post, the forums are littered with ATI folks, and they even feel betrayed.

You cant have Tri-linearing filtering with this optimization. Not even a selection to toggle it on or off.

Show me a pic where it degrades IQ?

btw, just because its posted at nvnews, doesnt mean anything. That site is worthless.

Typical ATIdiot, just because a website doesn't fully support ATI video cards it's worthless! :disgust:

 

[alexXx]

typical nvloser. The website is named nvnews. Think about it.

 

[Marsumane]

For once i totally agree with you. They both cheat.

Who is faster now? I'll say what ive been saying. I say were dead even and it is all in future drivers.

 

[nRollo]

Originally posted by: BFG10K

Our algorithm for image analysis-based texture filtering techniques is patent-pending. It works by determining how different one mipmap level is from the next and then applying the appropriate level of filtering.

That's exactly what I suspected was happening and it certainly explains all of the evidence we've seen so far. This is not a cheat.

[Faints dead away again]
LOL, BFG, I'm shocked you're quick to forgive ATI and their Brilinear filtering.
Myself, I'll wait for more from B3d, etc..
I'm sure ATI has all our best interests in mind, but somehow I still think they're just spinning this to cover their muddy footprints.