Commits on Dec 11, 2021

1. RenderOrder is in different scale in Reverse Z …

   I could not find info online about this problem, so i had to write a
   test myself.
   
   The obvious solution is to multiply the depth bias by 100 or so; but I
   was wondering what would be the side effects on other geometry.
   
   I.e. what would happen to geometry that is meant to be in front of the
   plane? We know a depth bias of 300 works at close range at reverse Z
   (where at normal Z a depth bias of 3 was enough), but would the plane
   get on top of everything as you move the camera away?
   
   So I made a test: https://www.youtube.com/watch?v=s2XdH3fYUac
   
   We can safely multiply by something like 100 or so, and we will get
   better results with reverse Z than we could with normal Z + the usual
   depth bias constant scale.
   
   With Normal Z, wven with a depth bias of 3 will eventually be on top of
   everything, but with Reverse Z + depth bias of 300, that never happens.
   
   Apparently multiplying the constant bias by 100.0f *is* the right thing
   to do.
   
   Fixes gazebosim#427
   
   Signed-off-by: Matias N. Goldberg <[email protected]>

   

   darksylinc committed Dec 11, 2021
   Configuration menu

   • View commit details
   • Copy full SHA for 5bc9e4b
   • Browse repository at this point

   Copy the full SHA

   5bc9e4b View commit details

   Browse the repository at this point in the history