Texture Seams Showing?

[edcwhite]

I am having a problem that is showing on both my desktop and laptop. The desktop is running a ATI 9800Pro with latest Omega Drivers (based on 5.7 cats)& the laptop an ATI Radeon Xpress200 running the ATI 5.7 cats. As you can clearly see in the attached images, the seams of the textures are showing. Aside from the fact this is occuring on two different machines, the fact that this only seems to really occur on Fern Bay & to a lesser degree Blackwood leads me to believe it may be a problem with the game not me. Could I please have some input on this?

[-SE1Z-]

ahhh
im having the exact same problem.
i thought it was my video card, as im having problems in other games.

running an AGP X800xt-pe - 256mb
running latest cats

http://www.2and2.net/Uploads/Images/LFS-lines.JPG

[Rotary]

I too have the same thing at Fern Bay... A6600GT 128mb, buggered if I can remember what driver though 7x.xx (what ever came with BF2 demo). It might be AA/AF related, which in LFS I have at 8/16. Although this doesn't happen on the other tracks with same settings, only Fern Bay from what I have noticed.

[atledreier]

That's an AA issue. I have it too, and also in RBR with AA on. And only with high-res textures. Weird, but I can live with it.

ATI Radeon 9700PRO with cat 5.7

[AndroidXP]

First (and slightly OT), it is reported that the 6600GT series has serious problems with AF, often causing texture shimmering or "pixel walking". AFAIK nVidia already works on a driver to fix these problems. A way to soften this shimmering is to turn the driver option of "Negative LOD bias" to "Clamp" and to set the MipMap LOD to a value of 0.3-0.5 (which can be done with the tool RivaTuner (http://www.google.com/search?q=rivatuner)).

It was also noted, that this shimmering seems to only occur when the anisotropic filtering is forced by the drivers, but not when set ingame (which isn't currently possible in LFS).

So much to the GeForce issue. What causes this on ATi cards (AFAIK) are hi-res textures combined with a sort of bad AF algorithm. You could try the same tweaks used for the GF, but I have no idea if that will help with this issue. Oh, and disable any driver tweaks to improve performance for "practically no visual difference" :rolleyes:

[edcwhite]

Good to see that I am not alone. I too have seen and heard about the problems with this in RBR, but this is different. First it doesn't occur on all tracks, second turning of AA & AF makes no difference. With AA & AF off the seam lines are harder too see (due to the fact that everything is shimmering) but are definitly still there. If we get a few more reports of this, I think I will post it in the bug reports section.

[ORION]

You can try to Clamp/Allow OpenGL comformant texture clamp also.

[Kegetys]

It happens because of the way the textures are mapped in S2. In those two screenshots for example the white lines appear to be the start line "checkerboard" from the bottom of the JA_roadxtras.dds texture. The texture is simply mapped in such a way that in the distance some pixels of the checkerboard get drawn because it is so close to the edge of the mapping. AF makes it much "worse" because it uses pixels from a larger area to draw one texel.

[xapexcivicx]

It happens because of the way the textures are mapped in S2. In those two screenshots for example the white lines appear to be the start line "checkerboard" from the bottom of the JA_roadxtras.dds texture. The texture is simply mapped in such a way that in the distance some pixels of the checkerboard get drawn because it is so close to the edge of the mapping. AF makes it much "worse" because it uses pixels from a larger area to draw one texel.
I envy this person.:nod::thumb:

[edcwhite]

Can anyone confirm that they are not having this issue?

[ThomasH]

I do not have those issues

[edcwhite]

What are your specs and setup? I would love to find out what I have that is causing this issue.

[Rotary]

First (and slightly OT), it is reported that the 6600GT series has serious problems with AF, often causing texture shimmering or "pixel walking". AFAIK nVidia already works on a driver to fix these problems. A way to soften this shimmering is to turn the driver option of "Negative LOD bias" to "Clamp" and to set the MipMap LOD to a value of 0.3-0.5 (which can be done with the tool RivaTuner (http://www.google.com/search?q=rivatuner)).

It was also noted, that this shimmering seems to only occur when the anisotropic filtering is forced by the drivers, but not when set ingame (which isn't currently possible in LFS).

So much to the GeForce issue. What causes this on ATi cards (AFAIK) are hi-res textures combined with a sort of bad AF algorithm. You could try the same tweaks used for the GF, but I have no idea if that will help with this issue. Oh, and disable any driver tweaks to improve performance for "practically no visual difference" :rolleyes:

Thanks!!

[Tyrion]

Pardon me for digging up this old thread, but I wonder if there has been any development regarding this problem?
I have tried changing the suggested settings in RivaTuner, but it doesn't seem to have much of an effect.

:schwitz:

[Ball Bearing Turbo]

Indeed I experience this sometimes too.... and RBR was brutal for it...

[keiran]

Never seen it as far as I can remember. Using a GeForce 4 MX and no AA or AF.

Keiran

[mrbogeyman]

i think it is exaggerated by low AA/AF levels. i noticed it when i put new GFX drivers in, and thus LFS had no AF or AA setup. FE is definately the worst for it, but it is greatly reduced by higher AA/AF levels imo.

i think it is either AA or AF in particular, but i canny mind :D

im on a GF6800 on 8xAA and 16xAF

[xapexcivicx]

First (and slightly OT), it is reported that the 6600GT series has serious problems with AF, often causing texture shimmering or "pixel walking". AFAIK nVidia already works on a driver to fix these problems. A way to soften this shimmering is to turn the driver option of "Negative LOD bias" to "Clamp" and to set the MipMap LOD to a value of 0.3-0.5 (which can be done with the tool RivaTuner (http://www.google.com/search?q=rivatuner)).
I'm way too lazy for that. I have my drivers forcing 4x AA and 8x AF on everything, and I only noticed such things in LFS.

[Ball Bearing Turbo]

Sorry for the second resurrection guys... but has anyone found a way to actually STOP this? I hate it, it's annoying.

[Ian.H]

Interesting thread! thanks to the people who've offered explanations here. I have this effect occassionally on Blackwood but also on my own stuff sometimes.. I thought it was down to bad mipping in the texture itself but sounds now like it might not be :)

Speaking of the way tracks are mapped, kegetys post and the current discussions in the hi-res texture threads, I tried slapping the texture I added to FEClub yesterday to BL1. I added some dirty marks to the track first that were done as a quick test, not very well for a tiled effect, but ignoring that.. the mapping seemed very strange!

http://tk.digiserv.net/tmp/bl1_tex.jpg

I'm gonna hazard a guess that the track's mapped in 3 sections.. left, right and centre because the track edge lines match, the oily lines match, but some how, in what seems to be random places, the centre part has been squashed or $SOMETHING (the few white pieces you see are the result of me not getting the grid line exactly the right size in the texture so it's overlapped ion a few spots). The part closest to the camera is what I had hoped to have seen around the track completely. The pitlane seems to use the texture ok

Granted, the texture I'm using for a base (1024x1024) was really designed to be used as-is to map the entire width of the track and tiled in 1 direction not chopped up and used in pieces.. so it may just be down to a different style of track mapping.

I've never looked at any of the tracks in this detail before, quite interesting what can be found anyway :)



Regards,

Ian

[micha1980de]

the attempt to crop any road-Extras and resize the JA_roadxtras.dds to 1024 pixels² doesn't sovle the problem as you can see in the pic, the sewing is dark now... geez

[AndroidXP]

It looks like this is actually black pixels "bleeding" between the seams there. When you modified the textures, did you replace any whiteish stuff with black pixels? Try to paint them gray instead.

[Zion]

As someone said beforeme; Sorry for digging up this old thread, but i have the same problem with my setup:

Asus A8N-SLI
3700+ @ 2,9GHz
2GB Ram
Gecube X1900XTX

Catalyst 6.9


Is there anyone that ca confirm if this issue is driver-version related?
Or if you can solve inte in any toher way without having to reinstall the drivers.

Also I got big problems with shimmering, My Mipmap setting/LOD is not set to -numbers, although i hate the blocky apperance with the trees on 0.0 setting.
So if anyone can help me with that, I would be most grateful.


I treid all possible combinaton i catalyst controlcenter, and the "seeming -problem" is not related to AA or AF as it is in Richard Burns Rally.
I doesnt matter which settings I'm using, its still there and I'm still really pissed that I'm not able to solve this problem.

Anyone got any ideas?

[Roadie]

I have the same problem now with patch V it seems...

The texture seems show up even with 16xAA, and 16XAF.