Hmmm, interesting. Something slightly strange about this though, in that trilinear filtering is a technique for using mipmaps.

A. Upper left screen: EnableMipMaps=0
B. Upper right screen: EnableMipMaps=1
C. Lower left screen: EnableMipMaps=1 + 8x Anisotropic via drivers
D. Lower right screen: EnableMipMaps=0 + 8x Anisotropic via drivers

I vote "D" 😀

What am I missing? I really can't see a difference.

Check out the windows on the building on the far left (under the '3' in '3:00' and left of the '3' in the center of the screen).

You can see that the textures on the sides of the buildings that are parallel to the road look better on the lower pictures. Also, there are abrupt changes in the blurriness of the same textures on the upper-right picture due to mipmap filtering.

I think trilinear filtering + 8x anisotropic + FSAA would be even better :p

Not quite. The extra interpolation between bi and tri is between mip maps. So if there are no mip maps you can't do trilinear. The purpose of trilinear interpolation is to smooth out the abrupt changes from one mipmap to the next.

There are two possibilities

1) When you have mipmaps disabled in OpenGlide, trilinear makes no difference.

2) When you ser trilinear, it overrides the mipmap setting in OpenGlide.ini, so that setting makes no difference.

In fact, disabling mipmaps (trilinear filtering) leaves the a simple bilinear interpolation on, while anisotropic takes care of the ugly transitions. I'm however, quite curious about the results in the lower left screen, by the geneal rule, trilinear+anisotropic assures the best possible results, but in Glidos you can clearly see that anisotropic has no effect at long distances, exactly where it should. Perhaps it's the LOD set to close?.

Spædbarn Dakin wrote:

In fact, disabling mipmaps (trilinear filtering) leaves the a simple bilinear interpolation on,

But you always get bilinear, so unless your turning off of mipmaps in OpenGLid.ini is being overridden by the trilinear setting, trilinear can have no effect.

while anisotropic takes care of the ugly transitions.

Avoiding transitions between mipmaps is trilinear inerpolation. Its not so much that that is an effect of trilinear interpolation; that is what trilinear interpolation IS.

I'm however, quite curious about the results in the lower left screen, by the geneal rule, trilinear+anisotropic assures the best possible results, but in Glidos you can clearly see that anisotropic has no effect at long distances, exactly where it should. Perhaps it's the LOD set to close?.

Why should anisotropic have effect particularly at long distance?

If you mean LOD bias, that's set by the game, although there is some chance that Glidos ignores the games setting.