Spoiler: Realroad Nurburgring VEC server Above here you see the realroad generated on the VEC server. Here I noticed quite a few things. 1- How thin the grippy line is 2- There only being one single line 3- How close that line is on the kerb 4- The line in the very slow T1 being just as dark as the quicker left hander T2, with the small straight in between being very dark as well. So I'll explain further here and how my brain thinks this doesn't seem right (aside from it being a weird brain). 1- A car has 2 sides of tires, 4 in total (thank you capt. obvious). There seems to be only one single thin line, that doesn't seem to come from the tires. If it came from the tires. If it came from the tires it wouldn't be a single line from T1 to T2, but would look more like my beautiful illustration here: Spoiler: Drawing not to scale To dig a bit deeper on the line itself, it seems like the middle of the line has the most grip. Here a bad sketch of how much grip is where on the line: Spoiler: Realroad grip now What my brain tells me what feels more realistic, is that it comes from the tires. Of course the outer tires generates more rubber, since there is more pressure on them. So here a beautiful black lined representation of how I think it should generate it. Spoiler: For a left hand turn Pointo two - I have been testing for VEC, which is multi-class racing. I drive quite different lines to a GTE, yet still I only see one single line. That seems quite weird to me, as we are definitely driving different lines. This also goes for oval racing, where there are ovals with NASCAR cars there are 2 lines possible, a high and a low line. I have never seen 2 different lines there either. 3 - I actually tried getting my outside wheels onto the darkest part of the realroad, as that is logically speaking the most grippy part. If I do that though, I have to go 2 wheels trough the grass on the inside. This tells me that the real road is generated from the middle of the car, which is wrong. 4- Doesn't track get more rubbered in where there is more pressure from the tires on the termac? This mostly is from high speed turns (as downforce is a thing), yet the slow turns generate just as much as the Schumacher S for example. And in the straight between there is hardly any pressure on the wheels, and since the cars have TC, hardly any wheel spin too. This also ties into point 2. You would expect the line to scatter there, since GTE cars keep a far lower line towards T2, whilst the LMP2 cars have a lot more grip to take a wider entry to T2. I would be happy for you guys to prove me wrong on some stuff, but my mind tells me pretty much that there is a lot wrong with this system, which should be improved
I would search for aerial pictures of real tracks, maybe you have a too digital idea of how rubber works, while maybe is much more analogic. The fact that AI is not a good candidate to put down rubber on the track is a known fact, because AI follow more strictly their imposed lines. Humans should make wider rubbered lanes. I think you don't have to think as rubber as paint made with the tires, maybe there is a spread effect. Maybe the car lay down a single line, as you suggest but never noticed that , and to me dark area have always made sense, never felt they were offset.
One thing i want to mention about RealRoad: It depends on the underlying resolution of road polygons too. While the beginning of rF2 IIRC 4 polygons (quads) per road width was common -> like ~2 meter wide -> res of RealRoad is 2 meter wide then too. later 8 polygons -> ~1 meter wide, now laser scanned and irregular meshes, ? meter wide. But the RealRoad resolution depends on road polygon resolution. So it depends on the road polygon resolution too whether you even are able to see RealRoad for left and right tires (even if rF2 distinguishes, which I don't know) EDIT: Just like this. Every road polygon vertex gets a new parameter (rubber build up). Rubber interpolated between vertexes defines darkness of RealRoad.
pretty sure the real road tex could do with rotating 90 degrees at nurburgring . . . this is real road build up as it is . . . and with tex rotated . . .
Drawing #1 is what should happen irl IMO. Not that instant of course, but there should be a gradient flow from light to dark or less/more rubbered-in tarmac. Just physics, where the load is heavier more rubber wears off. When a track is in constant use and the rubber build-up is huge you'd get a more washed out black line caused by all variations of driving lines. https://lh3.googleusercontent.com/p...W3MLfPYBwHeqFlaU_G7aTncZVNjf0eD16S2rKg3xH1gMg
That looks a lot more like it should yea. They are working on PBR for this track, so they might fix it in that update
@Remco Majoor is the RR in the pictures after a race with humans? Or is this a saved/preset line possibly done by AI?
Interesting observations. I was always assuming that here in rF2 realroad will read tires load and amount of rubber being sheared, but also knew that it might be that it is just tracking trajectory of a car and thats it. Your observations show that it might be more likely just reading a trajectory. It is also more likely thing to expect, because it definitely must save on calculations... 2400hz stuff you know.... Individual real road buildup should be a thing in rF2, unless something got wrong with that lately ? One thing worth to mention about Nordschleife/Nurburgring. It is very much behind standard ever since it was released (it was released after Le Mans, and LeMans is a tier higher graphically). So we wait for Nordschleife update very much, we do right ? Ohhh whatever... The point I want to make is that groove is known to be rushed out and wrongly made for whole Nurburgring and Green Paradise. Devs has informed that they used AI to generate grooves for this track, and they should change it for the better for the update. I also wonder if AI has different influence for groove build up since they use a bit different physics... By the way original ISI groove texturing makes no sense, in PBR tier S397 use much better texturing and shading for groove. So thats a bonus. Some day in simracing:
The top picture really shows what I mean with the 2 lines the realroad should create. Also shows a little bit of that shattering at the exit of that turn I was talking about, although it would not be as noticeable there as on T1/2 of Nurb of course. With the new CPU's coming up, I guess there is more possibility to calculate that stuff better too, but then again I also still read people using a 4th gen intel quad core for rF2. The problem is, is that the real road affects the line you have to take a lot. And if the line is narrow and expects you to drive over the grass with 2 wheels because the real road is too simplistic, then I see it as a big opportunity for improvement. Although S397 already has a long list to improve, with the biggest one taking quite a while already, I totally see this being as important as tire compounds working and tires being affected by the track temp for example.
Well we had splendid interaction right here. We are evolving fast as human beings online. Now have that delicious pizza.
