ATI Brilinear ?

Cowboy X said:

With the R300 series of cards and the 8500 series ATI adopted an angle dependant anisotropic filtering method . This saved framerate while giving identical image quality to 'full' af (which worked on all angles) except for the cases in a game where there are textures at the angles left out by ATI . Many people did'nt like it and wanted to see those angles rendered properly as wel , but for most it was a non-issue . ATI then redeemed itself by offering full trilinear high quality af when the application asked for it . [...]

Click to expand...

8x Aniso (set in control panel for both)

Click to expand...

As it has been mentioned before, I'll mention it again.

If you select your AF level through the ATI control panel, only the first texture stage is trilinear filtered and all others are bilinear filtered. If the actual application makes the request for trilinear AF, all texture stages receive trilinear filtering. The control panel settings are for those games that have no AF level setting in-game. This has been known, and publically announced, for a long time.

Since UT actually has a setting in-game for AF, you should use the built-in setting versus the control panel. This will likely solve the issue, because as clearly stated:

Click to expand...

If you are talking about something occurring in "lines" in the texture ("mip map transitions"):

"Naive" bilinear prominently causes that issue of "crawling" above a certain level of detail and contrast level for the textures (a level of detail that is very common nowadays, and contrast level that is pretty likely in a wide variety of situations).

"Naive" trilinear works to solve that problem directly by working to do extra blending. It pretty much removes the issue completely.

Theoretically, you could, as an alternative, do "tricks" that might not be "naive" to try and solve the problem with "naive" bilinear, but without actually doing "naive" trilinear. "Fast trilinear" can be thought of as such a "trick", that (AFAIK, though extensive testing hasn't been established yet) is successful enough to be a (faster) substitute for trilinear in the general case.

"Brilinear" is a name termed for an intermediate step between bilinear and trilinear...better than bilinear, but not as complete as trilinear. Whether it will have the "line crawl" that bilinear has depends on how "naive" it is and where it falls in between them...

If done in a "naive" way, and with the position in between them that seems to be done in order to get performance benefits, the success of "brilinear" tends to fall significantly short of "naive" trilinear, and so far hasn't had a good track record in established observation of being sufficiently close to trilinear in effectiveness. Therefore, a "naive" brilinear is quite ill-suited to silently replace trilinear, and this is why the burden is on ATI for an explanation to show that the indication of brilinear (proven) is not "naive" (unknown).
A "naive" brilinear appears to be what was identified for nVidia in the past over a period of time (well, for the FX series...the 6800U's "brilinear" hasn't yet been evaluated to see if it is different, AFAIK), where the deficiency to trilinear has been shown to be evident in "moving" video and even to still have noticable issues in still screenshots which aren't good at showing the issue. We don't have either for the current issue at the moment.

However, what seems to be the case is that the 1) X800 brilinear is not quite "naive", and that RV360 (9600) brilinear, also assumed to be "non naive", has gone unnoticed as being lacking in comparison to trilinear by many gamers. This doesn't mean there still isn't a significant problem with silently replacing trilinear with it...the lack of a problem with that is something that would have to be proven. It just indicates a possibility that the "non naive" aspects of it put it closer to "naive" trilinear than past experience with "naive" brilinear.

This is why there is still a lot of questions, and what the article's analysis didn't happen to address completely, though the article calling it a cheat based on past experience is based on a quite logical set of assumptions.

Click to expand...

jvd wrote:
Okay . I don't have some of these fancy testing programs

But i got a few people together . We had an x800pro , 9700pro , 9600xt 5800ultra
We had unreal 2k3 , 2k4 , cod , quake 3 all running , highest quality and highest aniso we could run it at .

No one for the life of us could see a diffrence between the radeons. My little sister commented on the fx though. She said it looked blurrier .I couldn't see a diffrence though.

SithSolo1 wrote:
Did you run it on the same monitor

jvd wrote:
Not the same monitor. Same type . all set up to screen defaults (bought 20 of them got a really good deal ) they are 19 inch sony trinitrons .

Click to expand...

Here´s the official ATI response:

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.

Click to expand...

Bar81 said:

OMG, ATI is the cheaterz like nVidia!!!!!!!!!!! Hopefully, this will get some of the ATI lovers who bash nVidia to stop or bash both

Click to expand...