Courtesy of @LilSKi
*******************
Another topic that can be discussed a lot. But before that, it is always best to have a good start point. So Lilski laid down some steps of his best practice on how to go about building actually the most important part of your track: THE ROAD [insert holy music]
Well here is another best practice post. In this we will discuss the difference between the visual road and the physical. There has been some confusion and misinformation going around I would like to clear up and set the record straight.
First off years ago there was no difference between the visual and physical road. What you saw is what you were driving on. This was fine but with the advancement of the physics engines we needed more road detail to feel all the little bumps that every track has. The problem was if you added too much detail to the visual road the framerate suffered. So what developers came up with was to separate the physical road from the visual. This allowed a very high detail, and invisible, physical road to hide under the visual.
Just to show the difference this is what they look like.
Visual. The key here is the poly count needs to be just high enough so your turns looks smooth and not a bunch of straight lines. Notice the turn is of a higher detail than the straight.
Physical
Be reasonable with your numbers. For example Kunos Monza is exactly 400k triangles and the Nordschleife is around 730k triangles. The key is to make sure you set the renderable setting in the editor to false. Also obviously the the physical model should be called 01ROAD and the visual can be anything else like VISROAD
The reason this is important is if the physical layer is low count like the visual you get this "notchy" feel to your track. If you have a hill instead of driving smoothly over it you are essentially driving over one flat plane after another.
You want this
Not this. This is exaggerated for show.
There are many different ways to accomplish this so I will simply show you a few things not to do.
First is a simple subdivide. You start off with something like this.
A simple subdivide looks like this. All you did was add to the poly count but didn't change the shape.
Then there is the Catmull-Clark subdivide. Notice it smoothed it out as well as adding to the poly count. This obviously also changes the shape somewhat and there are different ways around it. This is meant to show the basic concept.
One other thing to watch out for is the spline you build the track from. You have to make sure the spline resolution is properly set. Here are some examples.
By default the resolution for the spline in blender is 10 or 12. Shown here it is set to 2 to better show what is happening.
You can see no matter how many sections you add to the road it will always follow the resolution of the spline. And remember it isn't just doing this on a turn. It is also doing this up and down hills causing the constant bumps we all know and hate from many conversions from older sims. What you want is this. Here the spline resolution is set to 64 and the road is now smooth with the same poly count as before.
The next part will show how to apply different displacement modifiers assigned to different vertex groups to create bump details in your now higher poly road surface. After that we will apply decimate and subdivide modifiers to get the triangles shown above in the second shot. If I have time later I will post it but not likely until Monday or Tuesday.
PART 2, in next post.
Last edited: