Thanks Mark, I was hoping to get a situation where the clouds were easily accessible and you’ve found one straight away.
As I indicated before, there seem to be 30 vertices in the file cloud1m1.3 based on knowledge gained from the tent files.
If these are laid out in a 6×5 grid we have 20 rectangles or 40 triangles. There are however 62 polygon descriptions which in some ways resemble the rectangle descriptions from the tent file although they are a couple of bytes longer here.
I’m assuming that not all the polygons are used at the same time, and that there must be some selection mechanism which we can worry about later.
Anyway, in each of the 62 polygon descriptors I set the value (xx) to zero:
00 2E 00 00 00 03 00 xx 00…
I then ran the take off
scenario as Mark describes and as expected I see what look like grids of 20 squares in the clouds (the cloud1m1 file relates to a middle file according to Polak) … but after flying up for a closer look each of these squares seem to be made up of 2 triangles, which was a bit of a surprise.
It may be though that the common game 3D engine uses rectangles but these are converted to triangles by direct3d for display.
It also looks to me like that the cloud.3 files just desribe flat tiles at 15000' for viewing at a distance and that some other process produces the haze effect of actually flying through the clouds.
I apologise for the rambling nature of this post, but I don’t really have any concrete information to give.