WIN32 - Low cloud layer ground collision.

[stratochief66]

I don't imagine this is how the cloud height is intended, but it appears that if I launch from places somewhat above sea level, or near hills/mountains the clouds very brutally collide with the ground and cause graphical glitching. For example, Vanderburg.

I would assume the cloud height would ideally be above the local ground level, and not above sea level, but it looks to me to take that as sea level, which looks pretty rough at a number of launch locations.

[stratochief66]

I took some videos of the effect, in OpenGL and DX9 so they are easy to see.

https://www.youtube.com/watch?v=IMVLERF-ozw

[Pingopete]

Cloud height is fixed at whatever altitude the global 2D cloud sphere is at, this is a fundamental limitation of EVE unfortunately, It might be possible to match the cloud mask shape and position with land masses but this wouldn't really solve the issue as you'd need another entire layer for land masses, on top of that cloud movement speed is no longer relative to earth rotation (i.e. speed 0 would no longer lock clouds to a static surface position) so positional errors would accumulate over time plus it's not persistent between users. Thanks for throwing the videos together though it makes for a good comparison between setups. I guess revisit the idea of increasing cloud altitude as a temp fix for both opengl and dx9?

[Blacksuu]

one litle advice if you doo test use memory mod done by sarbian http://forum.kerbalspaceprogram.com/threads/92907 is good to track in real time ksp memory footprint :)

[stratochief66]

@Blacksuu I have that enabled for the other two videos, and in the comment for this video I mention what memory measurement reported when I later reproduced these circumstances with it on.

https://www.youtube.com/playlist?list=PLPdqukdb2-oc6QFVhnEJ3x4CKKnueIBDk