Hi, what would give better performance when texturing objects in gmax:doing a small texture, say 64x64 and tiling it over the object? (if its done properly, the object can look very good) or do the texturing for the full object with bigger bitmaps.I have looked at MS default textures, it seems that they make use of tiling in the autogen and in the big terminal buildings, but not for other objects in the scenery (like the gates of KLAX)if I repeat a texuture of 64x64 10 times... I would end up with an object with a texture of 64x640, the same quality in this case could be achieved with a full texture of say 128x256.What would give better performance? it seems that for terrain the MS desicion of doing tiling was performance-wise, is it the same for objects?

Hi Gabriel.This isn't an answer to your question, although I suspect that there is no substantial advantage of whether you use a small texture many times, or one larger texture in this case. Generally speaking, it's the number of polygons ( facets ) in an object, that induces a drag on the CPU. Modern Graphics cards can handle just about whatever we throw at them, as far as textures go.Sometimes the default MS designs are just the result of one of the MS team just fooling around. If I were on the team, there would be all sorts of exotic combinations of techniques. *:-* They experiment just as we do, to see effects of different combinations of design styles. I even suspect they have fun doing it.Some object designs are legacies from older versions that may have had a need for a certain design approach for that older sim.Hopefully Arno or Gerrish will respond as they would know much better than I. The only point I make here, is the default object designs aren't necessarily the best guides as to what is the best approach.Dick

I never really tested the difference in performance, but for high buildings etc I prefer to use the tiling myself. This way I can have a good resolution for the texture and still a small texture size.

Tiling of a small texture is likely to give better performance than single use of a larger texture because[ol][li]the time to load the texture from disk is much reduced (this can have a significant bearing on performance)[/li][li]the likelihood of the texture being retained in the graphics card memory rather than being paged out to make room for other textures is increased (or, to put it another way, more room is left for other textures).[/li][/ol]And, as you say, tiling a smaller texture can also help give high quality rendering. It seems that people are sometimes under the mistaken impression that using a larger texture just for its own sake will result in better definition - not true unless the larger texture actually has better resolution of fine detail.Microsoft's use of smaller textures where appropriate is almost certainly by deliberate design decision and not merely the result of experimentation.As I'm sure you know, tiling a smaller texture does not necessarily mean that one has to use more polygons, either, because one can repeat a texture across a single polygon by using texture mapping coordinates greater than 1.0. To repeat a 64x64 texture ten times across the width of a surface, as you suggest, one uses mapping coordinates of 0.0 to 10.0 instead of 0.0 to 1.0CheersGerrish

Hi,If, from an artistic point of view, the result is the same, clearly, from an engineering point of view, a big texture containing repeated parts is not good (see Gerrish post). Built-in advanced buildings, for example, is a case where MS uses texture tiling at the limit.Regards, Luis