Celestia Forums

Jupiter is suposed to have rings, small ones but still rings!

I kave tried tu kopy saturns rings and use a diffirent textur but thay turn wight !

If anyone kan do this pleas post it!

Be sure you are using a .png file and that you haven't lost its transparency while editing it. This crucial to keeping the rings looking right. Another point is that Jupiter's rings are not visible to the naked eye because of their faintness. They can only be seen when backlite fron the darkside of the planet. Keep this in mind.

Wat are you using to make the .png files?

Auscreely,

I'm sure most sophisticated paint programs can be used to create and save png files with transparency (also known as an alpha channel). However, since most of them have a fairly steep learning curve, I used the command-line based NetPBM programs to convert, split and merge appropriate images when I created a repaired ring image for Saturn. (The transparency channel is offset from where it should be in the file currently shipped with Celestia.)

NetPBM is a very old set of image translation programs which is once more under development at SourceForge. V9 of the Windows Cygwin version of NetPBM is available from a site in France. Hmm how about that: V10 Windows ibinaries are now availble via SourceForge. See
http://gnuwin32.sourceforge.net/packages/netpbm.htm

I hope this helps.

I would love to see Jupiter's rings too, but there's something I noticed in Celestia...

When viewing a ringed-planet from it's "dark side", the rings on the inside as well as the outside of the planet's shadow become invisible, instead of reflecting sunlight as they should outside of the shadow. Because of this, it wouldn't make much sense to give Jupiter rings because they should only be viewed from behind the planet since they are so faint in reality. Is this an issue with my hardware, or something that is just not implemented and/or possible in the sim yet?

The above anonymous was me...

Actually, I have to think it needs to be considered a "bug" -- or at least a limitiation in the way rings are currently implemented.

To clarify for Chris:

The brightness of the shadowed sides of Saturn's rings varies with one's viewpoint, but, as best I can tell, not in any physical way. The brightness of the ring surface which is turned away from the sun seems to depend on what fraction of the lighted face of te planet is visible from the observer's viewpoint. I don't think this is right.

On the other hand, doing it "correctly" is going to be rather difficult, i suspect.

But a Google searh did turn up one person's description of what has to be done. See http://www.mmedia.is/~bjj/data/saturn/rings.html

selden wrote:Actually, I have to think it needs to be considered a "bug" -- or at least a limitiation in the way rings are currently implemented.

To clarify for Chris:

The brightness of the shadowed sides of Saturn's rings varies with one's viewpoint, but, as best I can tell, not in any physical way. The brightness of the ring surface which is turned away from the sun seems to depend on what fraction of the lighted face of te planet is visible from the observer's viewpoint. I don't think this is right.

The shadowed side of the rings is rendered the same as the unshadowed side.

The way that I'm rendering rings most definitely is physically based, though my model could use some refinement. The rings are modeled as a collection of opaque spherical particles. For each vertex, I calculate the 'phase' of a particle; the illumination is 0.5 * (1.0 + cos t), where t is the angle between light-particle and particle-viewer rays. I could improve the model by adding a term for the backscattering of light. Also, no shadowing between the ring particles is taken into account. The amount of shadowing should be based on the ring density, which is a per-pixel value.

On the other hand, doing it "correctly" is going to be rather difficult, i suspect.

With a programmable pixel pipeline, it's actually shouldn't be too bad . . . I think I could use a more sophisticated model without noticeably impacting performance. It's mostly just a bunch of dot products . . .

But a Google searh did turn up one person's description of what has to be done. See http://www.mmedia.is/~bjj/data/saturn/rings.html

Bj?rn J?nsson's technique illuminates the entire ring system uniformly; mine does not, so I'd argue that mine's in a way more accurate. However, his scheme does make some effort to model back scattering and self-shadowing.

--Chris

Chris,

Thanks for taking the time to explain how you're modelling the ring particles. It does sound like the results should be reasonable.

As a point of curiosity, however, how do you determine the size of the particles? Is it some function of the brightness and transparency images?

Looking closely at Bjorn's ring reflection and transparency images, it's obvious that they were the original images used for the current Saturn ring PNG file. Unfortunately, as he implies, the two images are distorted differently along their radial dimensions. When both of his images are scaled linearly to the same size (e.g. both scaled to 1024x64) the pattern in the transparency does not align with the reflective pattern. It's not even close. The feature offsets also seem to change along the length of the images. The correlations for some of the features are obvious, but not all of them.

Also, when I've scaled his reflective image to 1Kx64 and use it in Celestia, it also seems obvious that the ring image does not include all of the rings. The inner edge of Enke's division overlaps with the orbit of Pan, meaning that the image has been stretched too far. I suspect some of the outer rings are missing from the reflective image while the transparency image goes past the outer radius defined in Celestia. At least this seems to me to explain the location of the two Enke gaps relative to Pan in the "atandard" Celestia saturn-rings image. Hmm. explaining it this way suggests how I might edit the two images to some reasonable alignment...