I wasn't entirely sure where to post this, so I thought it best just to create a new discussion... I had a thought for adding more 'resolution' to the nature of the driving surface. Why not create a 2D grayscale 'bump map' (for the sake of argument, I'm going to refer to the entire surface of both bumps and cracks as just 'bumps') to place over the track? This way one could take laser-scanned track data, create a displacement map from the rendered surface to get nice details that are too small to model into the track such as cracks in concrete, grooves (i.e. the 2nd apex to Big Bend at Lime Rock). Plus, for those of us creating fantasy tracks, or even if we don't have access to laser scanning equipment, we could 'model' the surface to actually match what you're seeing on the road. This would come in particularly useful with dirt tracks. All you'd need is a single 8-bit grayscale image to get more than enough height detail, multiplied by some scalar value to determine the max deviation from the surface (with a value of 127 being neither a crack or a bump). As far as I know from my limited experience with .tdf's, about all you can define for different sections of road is a bump wavelength and amplitude, and the sim creates a somewhat generic sine 'bump' (or is it a random function?). There's really no way to actually get the bumps to follow what you see on track. As far as I can tell there's no way to create a bump or crack that runs along the track (thus affecting you on turn in or track out for example). If my math is correct, an uncompressed single-channel 8bit 1024 map would be 1MB. Assuming that map covered an area of 10meters^2, you'd get a resolution of about 1cm^2 per pixel. Of course, to save space, you might not need that high of a resolution everywhere, so you could apply all sorts of texturing tricks such as repeating, flipping, different resolutions, etc. As a disclaimer in case this idea is actually good enough to get implemented...I'm releasing this idea completely freely in the public domain. If anyone wants me to sign a release, I'd do it! If anyone wants to take this idea and run with it, I'd be happy just to be able to use it in a mod! Anyone know if it's already been tried and to what degree of success? Or why it hasn't been done already? Any thoughts?
Is this idea a result of how rFactor currently works? I know nothing about modding but I assume that rFactor 2 tracks will be quite a bit different based on the fact that RealRoad will exist.
Yeah...I guess it comes from a mixture of the desire to be able to have bumps follow along with the textures, along with being able to visualize how the bumps are laid out during track editing, and a dissatisfaction with the way it feels like someone is just shaking the car instead of the car actually riding over bumps and cracks while driving in the sim. Also, people are always saying how they love "iBrand's" laser-scanned tracks. I think I read/saw somewhere, though, that the fine detail of the laser scanning doesn't make it into the rendered mesh, it's only used as a template for building the rendered geometry and for simulation purposes (i.e. a HAT/collision mesh). So, I figured instead of actually creating the equivalent geometry (which would be over 1 million polygons for the same 10m^2 patch with a resolution of 1cm) a texture would work much better. Not to mention, people who create fantasy tracks could SUPER easily just paint that kind of detail in Photoshop, or even create some sort of cool texture in something like zBrush or Mudbox and render it to a heightmap; or even take a photo of a road surface and use some of the really cool normal/heightmap extraction tools out there (such as nDo--a FREE plugin for Photoshop which creates a normal map which can be converted with the nVidia texture tools to a heightmap--or Crazy Bump). So, ultimately, even if ISI didn't want to deal with laser scanned tracks and data, it'd be nice if they at least gave the modding community the tools to be able to utilize that information if someone were to actually acquire it. Even still, I would think that if properly implemented, it would bring the quality of tracks up even closer to that of "iBrand's" and make the argument in favor of "iBrand" that much tougher, possibly resulting in more sales for ISI. So far I haven't read anything about rFactor2 doing anything differently to rF1, but a lot of information is yet to be released with regard to modding. I'm sure by now it's far too late for my suggestion to actually make it in even if it's sound and they liked it and thought it would be a better system.
