Trilinear Filtering Comparison(R420 vs R360 vs NV40 vs NV38)

here's ours if anyone missed it, got kicked out on the last login so I donno if we missed anything.

http://www.bjorn3d.com/#56340

 

How about a direct link, so we dont have to visit the dump that is, nvnews.

 

I'm not sure that this has been independently proven True of False yet. That really isn't the point. They should have an option to enable "Legacy Trilnear" (using thier words not mine). There obviously is a difference or else they wouldn't make a point like they have that thier method is different.

It still remains to be seen if thier method is better, I'm not going to just take thier word for it.

 

More chat stuff

 

P.S. I did just get the above a second ago it seems it is still going on.

 

Yep. I dont't think I've seen this much pointless messages in B3D's thread ever before. And this is only 10 pages long..

 

More Chat

 

Ignore my last and let me ask this. Since the Drivers won't allow full-time enabled "legacy Trilinear". How is it that anyone else can confirm that thier method is better on an X800? True you can do some comparisons to a 9800 or Nv card but that is not the same. So how could anyone possibly prove to ATI that they should enable it when they can't even see the alternative on that hardware?

 

yay someone posted the answer (i got dsiconnected before it was answered)

 

uh oh :

We provide trilinear filtering in all cases where trilinear filtering was asked for

 

More...

 

which means either legacy, or the new one. As their both trilinear.

 

hb i wasnt being serious, just trying to get a rise out of you

 

what's this supposed to mean ? i really don't get it.... , can someone translate please ? :?

 

People have way too much time on their hands, 80 pages of nothing. Go grab a beer, take the wife out for dinner or something. The internet was just waiting for some smell of a ATI conspiracy, then it is like a pack of wolves smelling a wounded animal. :?

 

Since I am seriously techically impared, unless I'm missing to notice something very important, it looks to me as the guy is trying to avoid a clear answer.

 

Why would you think it might be worse? Think they'd want to decrease IQ, which is one of their big advantages over nVidia, simply to garner a bit higher frame-rates in benchmarks? I don't think it's likely, and that instead what they've done is to optimize their drivers to provide what they consider to be acceptable IQ in line with "legacy trilinear," while running faster at the same time. In short, this is what all legitimate optimizations are supposed to do. Maybe they should have called it "extra-trilinear" or "super-trilinear" or "faster trilinear" for the folks hung up on the terminology...

I mean, I think you've got it backwards: the first thing they looked at was "legacy trilinear" and what's come of it is an optimization congruent with the R4x0 hardware which employs optimized trilinear in certain cases only where the algorithm tells them it will work without IQ degradation. From what I understand of what they've said, in cases where the IQ requires legacy tri, the optimization switches off. The goal of the optimization seems to be, therefore, to only employ the optimization where it does not affect IQ adversely. That's a perfectly reasonable thing for an optimization to do. A "cheat" on the other hand would deliver demonstrably worse IQ while providing better performance, which is why it would be a "cheat" as opposed to an optimization.

The problem I think is that as ususal the idea of "cheating" is being confused with "optimization"--and we all know who to thank for that bit of throwback propaganda, don't we?...

 

Quote:

Coloured mipmaps naturally show full trilinear - statement

our image quality analysis reveals that there could be visible differences in the image. - issue

It should be noted that trilinear was originally invented to smooth transitions between mip-levels, and in the original definition mip-levels should be filtered versions of each other, as coloured mip-levels clearly are not. - statement

Despite this, we understand that people often make use of hardware for purposes beyond that originally envisioned, so we try to make sure that everything always functions exactly as expected. - issue

:?

 

Uhm, Volt works over at Bjorn3D now...

 

I may have missed it but have we determined the reason for filtering differences between R360 and R420 or have such differences even been confirmed? Outside of the COD screenshots is there any evidence which points to reduced filtering quality of the R420 vs the previous generation ATI products?