ATi is ch**t**g in Filtering

[Zengar]

Does it mean that extreme filtering performance of ATI is faked?

What's about Nvidia? Are they using trillinear or also brillinear?


This would change the situation slightly...

 

[digitalwanderer]

Highland's only an hour or two away by car, I'd be happy to take them off your hands for $20 each if they bother you too much

I got them for $50 each at me local bi-monthly PC show a few months back, I highly recomend the place for deals! (They always have some great deals on used monitors, but you really gotta hunt and kind of know what you're looking for.)

I'll keep it in mind though, I was looking at some nice 19 & 21"ers at the show this weekend and I'm thinking it's almost affordable enough to justify bumping up from me 17"....but 17" really are more than enough for me still.

 

[Hanners]

What's about Nvidia? Are they using trillinear or also brillinear?

Well, reviewers who used 61.11 were unable to turn off brilinear as the option to do so was broken in the drivers, so it seems like (for those sites at least) both sides were using brilinear.

 

[L233]

Well, reviewers who used 61.11 were unable to turn off brilinear as the option to do so was broken in the drivers, so it seems like (for those sites at least) both sides were using brilinear.

IIRC, most reviews used an earlier driver version though.

 

[Geo]

So, both the 9600 and X800 are doing 'brilinear' now in later drivers, even on the highest settings on the texture quality slider?

No.

Bit comparisons of Trilinear vs Bilinear appear to show that X800 is doing "Brilinear" filtering by default - because the colouring of mipmaps appear to show that full trilinear is in operation the assumption is that when you look at normal images "Brilinear" is forced on by default but "removed" when coloured mips are in use.

However, 9600 shows similar behaviour under bit comparisons, and appears to have done so since its release (but without anyone noticing any "Brilinear" effects in games, as yet). But, when you turn down the texture slider on 9600 and the early drivers the texture quality is still lowered in comparisons of Bi-vs-Trilinear and the coloured mip maps show "Brilinear".

Would it be fair to call this evidence of "intent" (as opposed to "whups!")? Even if not conclusive evidence?

 

[flick556]

So, both the 9600 and X800 are doing 'brilinear' now in later drivers, even on the highest settings on the texture quality slider?

No.

Bit comparisons of Trilinear vs Bilinear appear to show that X800 is doing "Brilinear" filtering by default - because the colouring of mipmaps appear to show that full trilinear is in operation the assumption is that when you look at normal images "Brilinear" is forced on by default but "removed" when coloured mips are in use.

However, 9600 shows similar behaviour under bit comparisons, and appears to have done so since its release (but without anyone noticing any "Brilinear" effects in games, as yet). But, when you turn down the texture slider on 9600 and the early drivers the texture quality is still lowered in comparisons of Bi-vs-Trilinear and the coloured mip maps show "Brilinear".

If Brilinier is removed when color mipmaps are used than wouldn't there be some fps penalty, otherwise there would be no point in doing something like that.

 

[Stryyder]

Well, reviewers who used 61.11 were unable to turn off brilinear as the option to do so was broken in the drivers, so it seems like (for those sites at least) both sides were using brilinear.

IIRC, most reviews used an earlier driver version though.

Many did, but not sure I would say most.

 

[digitalwanderer]

Just an aside really quick for poor Dave, did anyone else find his "Don't quote me on that" sig particularly humorous in light of the way his posts were put up as news again all over the web?

If it's any consolation Dave, I finally get your sig.

 

[flick556]

Maybe Ati is doing full tri and Nvidia's opinion that brillinier is good enough is valid

 

[Miksu]

So Ati is now doing the same thing that NVidia has been doing the last six months or so (detecting the colored mipmaps that is) ?

And no, I'm not defending Ati's work. This is kinda sad. Lars' benchmarks are interesting, so hopefully you can figure this out soon.

 

[pocketmoon_]

If Brilinier is removed when color mipmaps are used than wouldn't there be some fps penalty, otherwise there would be no point in doing something like that.

I may be way off base here, but by forcing trilinear when colour mipmaps are detected, then this would appear to give correct visuals when running those synthetic AA/AF/mipmap usage tools. Then under gaming conditions we go back to brilinear. Of course no-one would be interested in the FPS displayed by a tool to display IQ.

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

 

[nutball]

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

Uh huh. Don't you think that that "feature" would have some sort of marketing name (eg. IntelliFilter or SmoothSample) and be slapped all over the box

 

[pocketmoon_]

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

Uh huh. Don't you think that that "feature" would have some sort of marketing name (eg. IntelliFilter or SmoothSample) and be slapped all over the box

Well lets wait a day or two, I'm sure ATI come come up with one

 

[flick556]

If Brilinier is removed when color mipmaps are used than wouldn't there be some fps penalty, otherwise there would be no point in doing something like that.

I may be way off base here, but by forcing trilinear when colour mipmaps are detected, then this would appear to give correct visuals when running those synthetic AA/AF/mipmap usage tools. Then under gaming conditions we go back to brilinear. Of course no-one would be interested in the FPS displayed by a tool to display IQ.

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

No one would be interested before today.

 

[DrawPrim]

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

Most likely you are correct. The Microsoft conformance tests use colored mip maps all the time to insure a driver is displaying the correct result. More than likely the driver does some intelligent "peek" at the mips to determine how to optimize the mips.

 

[Exxtreme]

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

No. If you're using textures with one colour then you don't need a texture filter. If there would be an adaptive algorithm which analyses the content of a texture, this algorithm would set the "optimisation" far more aggressive because a coloured mipmap is a texture with one colour.

 

[no_way]

oh for f***s sake. There used to be a time in history when craphics cards rendered what you asked them to via API. Looks like those times wont be coming back.

Does this mean that when in my application i choose to apply decals or detail textures in second or third texture stage, im basically screwed in terms of filtering quality ?

 

[DrawPrim]

Of course this could be the drivers displaying intelligent adaptation, i.e. analysing the content of texture mip-maps and setting filtering based on some measure of visual difference.

No. If you're using textures with one colour then you don't need a texture filter. If there would be an adaptive algorithm which analyses the content of a texture, this algorithm would set the "optimisation" far more aggressive because a coloured mipmap is a texture with one colour.

Your mip filter would be LESS aggressive because the textures are vastly different between mip levels.

 

[Martillo1]

craphics cards

 

[Colourless]

What I think is, yes, they are doing some 'brilinear' trick and on textures that it would show it they are doing standard trilinear. Now one thing i'm curious about is what exactly the threshold between using it and not using it is. Quite possibly it's not purely intended to hide what's going on with coloured mip levels, but could also be intended to be used with any texture that has mip levels that are quite different. Need some sort of filtering test application that lets you set an blending level for the coloured mip levels to see if you can notice the change.