Celestia Forums

I discovered something very wrong with the new atmosphere code (latest CVS version). I defined a planet around HD 98231. There are two other stars pretty close (Alula Australis Ba and Bb), as seen on the picture below. Look at the vivid atmosphere. This isn't normal. I used exactly the same code for a planet around another star, and the atmosphere works as it should there.



And here's exactly the same planet around another star (Sig Col). The atmosphere rendering is okay in this case (with only one source of light) :



Here's the planet's definition which revealed the problem (I don't think the textures matter much here) :

Cham,


I think perhaps you've forgotten that bright red = vertex shader program error. In other words, it isn't simply a problem with the atmosphere code. It's usually caused by Celestia being unable to cope with a large number of simultaneous surface effects. (I suspect it'll go away if you turn off "show cloud shadows.")

Please provide a copy of shaders.log

selden wrote:Cham,


I think perhaps you've forgotten that bright red = vertex shader program error. In other words, it isn't simply a problem with the atmosphere code. It's usually caused by Celestia being unable to cope with a large number of simultaneous surface effects. (I suspect it'll go away if you turn off "show cloud shadows.")

Please provide a copy of shaders.log

It's probably a bug in how Celestia deals with atmospheres when there are multiple light sources. The surface itself looks fine, it's the atmosphere that looks wrong. Unlike surfaces, atmospheres don't require a lot of interpolants (since there are no textures), so the usual problem where the shaders exceed the hardware limit is probably not the issue here.

--Chris

Selden,

The surface itself is prefectly fine. All textures are working properly. And cloud shadows is already OFF (it simply doesn't work at all anyway!). In this case, as Chris said, it's only the atmosphere itself which has a problem.