[homerdog]

Sorry for the confusion, I was actually referring to the whole undersampling thing Novum has been on about. And that seems like more than just a bug in the R8xx hardware.

[Arnold Beckenbauer]

Originally Posted by homerdog

A hardware bug that they haven't managed to fix since R600?

There is a speculation that it's a texture cache bandwidth problem/limitation - since R600.

[Tamlin]

Tested with Catalyst 10.7 and it seems that nothing have changed.

[Rootax]

Originally Posted by Tamlin

Tested with Catalyst 10.7 and it seems that nothing have changed.

Maybe this is not a problem for ATI, so they won't change it ?

[homerdog]

Originally Posted by Rootax

Maybe this is not a problem for ATI, so they won't change it ?

Do you mean that ATI is not concerned with it? Or the problem does not exist?

[Rootax]

Originally Posted by homerdog

Do you mean that ATI is not concerned with it? Or the problem does not exist?

This

[homerdog]

Originally Posted by Rootax

This

Ha, well I'd wager they'll look into fixing it at least for the next refresh. It's not like they don't know about it at this point.

I'd also wager they won't mention they fixed it or that there was a problem in the first place. Which is understandable if a bit disappointing.

[Tamlin]

Originally Posted by homerdog

Ha, well I'd wager they'll look into fixing it at least for the next refresh. It's not like they don't know about it at this point.

I'd also wager they won't mention they fixed it or that there was a problem in the first place. Which is understandable if a bit disappointing.

Lets hope they do. Alienbabeltech have earlier fired off a message to Catalystmaker about these issues, so they are definetly informed about it.

[homerdog]

BTW Tamlin the thread title is redundant. Anisotropic Filtering filtering.

/nitpick

[Tamlin]

Originally Posted by homerdog

BTW Tamlin the thread title is redundant. Anisotropic Filtering filtering.

/nitpick

LOL! Nice catch.

Fixed.

[no-X]

Originally Posted by Tamlin

Lets hope they do. Alienbabeltech have earlier fired off a message to Catalystmaker about these issues, so they are definetly informed about it.

I informed Dave a quater ago. No response. Local reviewer informed local PR, who forwarded the mail to ATi about 2 months ago. No response. I'm not sure, but I believe ATi is trying to survive till the launch of next refresh...

I don't believe this is a software issue (whether the hard transition or sampling density)...

[homerdog]

Originally Posted by no-X

I informed Dave a quater ago. No response. Local reviewer informed local PR, who forwarded the mail to ATi about 2 months ago. No response. I'm not sure, but I believe ATi is trying to survive till the launch of next refresh...

Not surprising. If there's nothing they can do about it there isn't much to say.

I wonder if it is a glitch or by design?

[Arnold Beckenbauer]

Originally Posted by homerdog

Not surprising. If there's nothing they can do about it there isn't much to say.

I wonder if it is a glitch or by design?

Imagine: AMD releases a full trilinear fltering tool, which gives you full trilinear filtering quality without ugly samples savings, but it costs you eg 10% of fps. And then some hardware reviewers start to write their reviews using this tool...

[homerdog]

NVIDIA has offered full quality filtering for a long time now...

[no-X]

Originally Posted by homerdog

I wonder if it is a glitch or by design?

as for sampling density, Jawed mentioned cache bandwidth

As for the transition, I believe it can be related to the last minute changes of the gpu/architecture. It could be a bug arised from quick redesigning as well as an intention (transistor saving method). But I think the second variant (intention) is very unlikely. I can't believe that they would sacrifice milions of transistors to achieve perfect angle independent filtering and ruin it at the same time by using some strange tranzistor-saving-hard-transition-creating technique...

[Tamlin]

Originally Posted by no-X

I informed Dave a quater ago. No response. Local reviewer informed local PR, who forwarded the mail to ATi about 2 months ago. No response. I'm not sure, but I believe ATi is trying to survive till the launch of next refresh...

I don't believe this is a software issue (whether the hard transition or sampling density)...

Its a shame that there is no response. Thats a bad sign.

I'm not so into the programming part of AF, so I can only report my observations. Could anyone explain why this is not an issue with larger checkboard patters? If I use d3d-af-tester with checkerboard texture 4, there is no hard transitions and no grey mesh circles. I'll see if I can upload a picture of this later today.

Could it be something wrong with the minification/magnification?

[Arnold Beckenbauer]

Here is an official statement: http://www.computerbase.de/news/hard...xturfilterung/

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 company's 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 don't 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.

[Squilliam]

Hey Arnold! They are talking about the ancient R300+ past there...

[Arnold Beckenbauer]

Originally Posted by Squilliam

Hey Arnold! They are talking about the ancient R300+ past there...

But this statement is still up to date.

[Squilliam]

I guess you're right... Its just strange seeing something that old being brought up. Its funny how fast this industry moves...