planetary rings aren't being blocked

[safemode]

It seems that planetary rings aren't correctly rendered.

The planet doesn't block out the part of the ring that should be on the other side of it. They appear to be drawn on the planet...that is to say, the planet isn't transparent or anything, but i can see the entire ring.

[pyramid]

I have noticed that too and it has been reported previously. Maybe something is mixed up in z-buffer.

[safemode]

I really hope it has something to do with why it takes much more gpu power to view the center of planets and large units than edges of said units or empty space. That would be awesome to kill two birds with one stone, but i doubt i'm that lucky.

[Coragem]

I have noticed this too, its not random. every planet with rings, shows up like this.
Rings behind planet are visible.

[pyramid]

New rings have been committed (see 4 screen shots in gallery > User Screenshots > page 33+).

I have noticed that the bug does not always occur. Please start VS with

Code: Select all

./vegastrike test/planets.mission

after checking out svn 11975+. I have put 2 gas giants in the planets test system. They are almost exactly equal but differ in their orbit and ring parameters. On "Gassy" you can not find this error at all, no matter from what angle you are looking. However, on "Hestia", try the following:
* right after starting up, switch to target camera ("8" key)
* zoom out slightly so that planet covers full screen heigth ("0" key)
* pan camera ("z" key)
> The rings should be correctly hidden behind the planet

Now:
* zoom out a lot and pan the camera
> the rings will not be totally hidden behind the planet

This leads me to believe that camera is not handling z-buffer correctly.

[safemode]

not handling the z-buffer may cause things to get drawn that should be blocked....hence lots of gpu usage when viewing objects close up.... sounds like another problem....can't quite put my finger on it... hrmm

[hellcatv]

the problem is that far objects are sorted rather than rendered using the zbuffer

our solution has been to break up the rings into many sub-rings

however the problem probably is that we broke them up in to too few subrings

the number of subrings broken into needs to be on the order of the delta between the planet and the rings

I forget where the var is to divide into subrings, but if we can up that for high detail it should be better