Celestia Forums

My W Ursae Majoris stars add-on is now up on Celestial Matters for all those who want to add 43 contact binary systems to their Celestia experience.

Enjoy

Thanks Andrew.

The model for i Boo has a small singularity :


Also, can you raise a bit the mesh resolution around the main symetry axis ? I easily see the polygonal shape in some orientations, and I find it a bit annoying.

Ok I've found where the bug that causes the "singularity" is located - I accidentally output 0 instead of $x - this bug will not affect any of the other models because only 44 Boo has a zero fillout factor (i.e. the components just touch at that single point). This seems to have been caused by my finding out halfway through the developm

I could increase the mesh resolution however I don't want to increase the filesize too much. I'm not on my Windows partition right now so I can't use CMODFIX, but here's models for 44 Boo and RW Dor with higher resolution, in ASCII format, is this better?

If so I'll reconvert them all to binary format and upload a new version of the add-on. The increase in filesize will probably be about 60%.

(attachment removed, updates now incorporated into main download)

Andrew,

yes, it's better with this slight increase in resolution. You should update all the other models too (in binary format).

Ok, I've made an updated version. New link

Thanks Andrew. Much better now.

Unfortunately, there's a huge depth sorting bug with Celestia (OS X only ?) which gives a bad rendering on stars with a non-ellipsoid shape :


Chris, please ?

Wow that looks bad. I'm definitely not seeing that on my machine (either using Windows XP 32 bit or Xubuntu 64 bit), which has an nVidia 8600GT card.

ajtribick wrote:Wow that looks bad. I'm definitely not seeing that on my machine (either using Windows XP 32 bit or Xubuntu 64 bit), which has an nVidia 8600GT card.

This rather looks like "egg art" Never mind, Easter is near...

Seriously, I don't see anything like Cham's screen shots.

Fridger

On further investigation, I can reproduce the effect. I see the problem in the Basic and Multitexture render paths, however the OpenGL vertex program, OpenGL vertex program/NVIDIA combiners and OpenGL 2.0 paths are fine.

I'm experiencing the problem under OGL2. I think I also saw the problem under all the rendering paths (I'm not sure, since I usually don't use the other rendering paths, and I'm not at home right now).

This serious depth sorting glitch is killing many addons.

The problem is with the positioning of the glare sprite in situations where the hardware point sprites are unavailable (in legacy render paths and on Macs with ATI cards and their broken GL drivers.) It's likely that fixing the problem will only require making the legacy render paths use the same depth sprite position as in the OpenGL 2 render path.

--Chris

chris wrote:It's likely that fixing the problem will only require making the legacy render paths use the same depth sprite position as in the OpenGL 2 render path.

Does that mean that you'll try to fix it in the near future ? And what about the bug I reported to you, few days ago, about the model with sprites used as a star ? Is there any relation to the problem shown above ?

Cham wrote:

chris wrote:It's likely that fixing the problem will only require making the legacy render paths use the same depth sprite position as in the OpenGL 2 render path.

Does that mean that you'll try to fix it in the near future ? And what about the bug I reported to you, few days ago, about the model with sprites used as a star ? Is there any relation to the problem shown above ?

The attached patch for render.h and render.cpp should fix the problem. It modifies the code to use the same glare sprite calculation regardless of the render path.

I haven't looked at the sprite models problem that you submitted a few days ago. I should get to it today.

--Chris

Well the recent updates to the SVN have fixed this issue for me. Don't know if it works on Macs.

ajtribick wrote:Well the recent updates to the SVN have fixed this issue for me. Don't know if it works on Macs.

Yes, the problem is completely solved on my Mac.