I've noticed pretty big change of RealRoad graphic appearance with build 85. Particularly rubber laydown and marbles have changed. While I was tracking down the reason of this, I've noticed weird difference between "old" marbles texture and the "new" one. New marbles alpha channel do NOT cover whole texture area, just a quarter of it. In my understanding this will lead incomplete spread of marbles on track surface. Am I missing something..? Cheers!
Good catch ! That must have happened when I resized the original image. Ultimately, it makes no difference though, as I don't think the alpha channel is used by RealRoad right now. It's not a transparency map as you might expect. The RGB channels are multiplied with the rest of the road, and the alpha isn't used (according to a quick black/white alpha test -- and that's also why I didn't catch he dodgy alpha channel when I tested the texture). This is similar to the RaceGroove: RGB are multiplied with the road; A controls the specularity. I could in fact re-export the marbles texture as DXT1 in the future and save a little bit of memory .
Ah, now this makes sense. Thank you for clarifying this up. I was wondering how actually groove was calculated on track surface. RGB multiplying is very clever way to do it. So, marbles and racegroove RGB channels are multiplied with diffuse map and spec map controls overall specularity (groove and rest of the road outside of the groove)? Is multiply map affected (with RGB multiplying) also or is it just "detail" for diffuse map? Cheers!
It's a bit more complicated than that. Well, complicated to explain at least . In fact, your questions confused me . The shader description really does a good job of explaining how every texture stage is used, especially if you have played with the textures for a while. It's much easier to try a few texture changes and seeing the result change than me trying to explain everything in a very clear way .
Yeah, I noticed my question was bit confusing... And yes, few texture tests last night gave me more accurate impression how each stage affects final result. Thank you, Luc. Cheers!
