Celestia Forums

I am running Celestia 1.6.0. I'm not sure about any more specifics other than that, other than that it's the executable with Chris's improved star rendering.

I've discovered what may be a bug pertaining to the shading of a planet by its rings. Also visible is a depth sorting bug with the rings that has already been reported here, so I will not elabourate on it.

I wanted to view Chris's lovely implementation of ring shadows upon the moons of Saturn, but didn't want to have to find the equinox and what-not, so I just went into edit mode and tipped the planet over. But...


I'm not particularly sure this is a serious issue, but didn't know if it was a known one.

Furthermore, the rings do not project the geometrically expected shadows onto the moons in this orientation.

Hi! With Celestia 1.6.0 official my situation is so:



seems correct. But it's without the new star render. How about your situation without the new star render? Do you have substituted the shader files within the shader folder? Some shaders were updated from 1.6.0.

I had to go into edit mode (see thread title) and tilt the planet (see original post). It doesn't look like you have done either of these.

The bug is visible in the version without the improved star rendering. It's only seen in OpenGL2.0, and is fine in OpenGL.

Sorry, my fault: I was in edit mode but then I did capture out of. Nonetheless, is the same: no problem in all renderers but basic.

I specifically mentioned that the planet has to be tilted over (see here and here). From the screenshots I posted (first, second, third), it is clear that I have done this. You'll note Saturn is not coplanar with it's moon system. I would have figured this would have been a clear indicator that the orientation of the planet has changed. I have repositioned it's rotation axis within Celestia using edit mode. See animation:

I really would like to not have to belabour this point further...

Yes I understand what you describe with "tilt": the rings must be not coplanar with the Sun. But how do you tilt the planet in edit mode? My controls doesn't seems to works, but am I that is brain-dead here, sorry.

In edit mode, hold [Ctrl] + [Left Shift] and move the mouse. x-axis on the mouse controls one of the planet's axes, y-axis controls another. (I don't know which axes are controlled).

(edit Mar 3, 2010: fixed spelling typo)

Following your suggestions, I get this:



in which there is no shadow, being the Sun orthogonal to the rings. And for whatever Saturn's movement I achieve (tilting) the rings' shadow follows the planet. No trasversal bands at all (I'm unable to reproduce your bad projections). Notice that the rings' locations (descriptions) remains where would be the rings.

Fenerit wrote:Notice that the rings' locations (descriptions) remains where would be the rings.

That's interesting. Might constitute a bug by itself.

Were you using the OpenGL or OpenGL 2.0 render path?

I was using OpenGL 2.0 render path. But is the same for all the remains render paths but basic.
My suggestion for your case was that of the shaders, being the .exe with the new star render a 1.6.1 version, not a 1.6.0 one, so it need that the default shaders be updated from the Celestia's SVN trunk (the planetshine's improvement did modify a shader). But, supposed you didn't update the shaders, whether they could cause problem with the new star render's .exe, nonetheless it wouldn't caused problem with the official 1.6.0, being set for such a version. In conclusion, I'm waiting for other users' tests.

I can reproduce this bug with current SVN, win 32 built with new orbits, new stars. The shadow of the rings on the planet is wrong in render path OpenGL 2.0 and right in OpenGL 1. The shadow of the planet on the rings is right. When exiting edit mode the wrong shadows persist.

Guillermo

Windows 7 SP1 x64, NVIDIA GeForce 9500 GT, Forceware 8.17.12.5896

abramson wrote:I can reproduce this bug with current SVN, win 32 built with new orbits, new stars. The shadow of the rings on the planet is wrong in render path OpenGL 2.0 and right in OpenGL 1. The shadow of the planet on the rings is right. When exiting edit mode the wrong shadows persist.

Guillermo

Windows 7 SP1 x64, NVIDIA GeForce 9500 GT, Forceware 8.17.12.5896

Hi Guillermo, do you have this bug even with the pure 1.6.0 official version? (whether you can test it...).

FYI, I cannot reproduce this bug with several Celestia SVN 1.6.1 without new orbits, and without new stars.

I do see the ring shadows bug in edit mode and after exiting edit mode (render path OGL 2.0) with:

* SVN fresh (old stars, old orbits)
* SVN fresh (old stars, new orbits)
* SVN patched (new stars, old orbits)
* SVN patched (new stars, new orbits)

I do not see the ring shadows bug (any render path) with:

* Celestia 1.6.0 official unmodified just downloaded
* Celestia 1.6.1 RC0 form Chris' files (dated 31 October 2010, latest RC)

G
Windows 7 SP1 x64, NVIDIA GeForce 9500 GT, Forceware 8.17.12.5896

In my case, I don't have the shadow on the rings bug with my version of 1.6.0 (the shadow is following the rings when I rotate the planet in edit mode). However, the location labels don't follow the rings while I rotate the planet.

FYI, I can reproduce the bug with the celestia-v1.6x-svn5135 with fade orbit and with new stars (Guillermo .exe).

Ahem... noob question: how can an edited planet (in edit mode) be restored in its original position?

Fenerit wrote:Ahem... noob question: how can an edited planet (in edit mode) be restored in its original position?

AFAIK, there's no way, except from the eyes ! Hehehee !