Celestia Forums

I have, from a previous project, a compilation of the known globular clusters
in the Milky Way. There are approximately 200 of these. I modelled them
simply with a power law distribution of about 1000 stars each and viewed
it using VRML.

A frame grab from the VRML can be found at:

http://terpsichore.stsci.edu/~summers/files/celestia/

Note that the VRML has no dimming with distance for the stars. By the
way, the 2D texture of the Milky Way is an illustration based on the
current best mathematical model of the spiral arms and knowledge
about the bar/bulge.

Globular clusters would be a nice addition to Celestia. Since I generated the
VRML with perl scripts, it would be simple to generate a text file of the globular
clusters (i.e., like galaxies.dat). It would be a little bit more to create a binary
file (i.e., like the stars data).

What would be the best route for providing such an addition?

The best way to do this right now is probably to add the globular clusters to galaxies.dat. Use a type E0 galaxy to get a spherical distribution of blobs . . .

I know this is a hackish way to do globular clusters . . . For one thing, they'll never resolve into individual stars. Real support for globulars will have to wait for a future version of Celestia.

--Chris

Check out this topic for a info on how someone reproduced the objects in the Messier catalog my making additions to galaxies.dat:

http://ennui.shatters.net/forum/viewtopic.php?t=317

--Chris

Here are some ideas for star beams and volumetric light as well:
http://developer.nvidia.com/view.asp?IO=cw_demos

Yes, I saw the Messier add-ons and they already include a number of
globular clusters. However, they don't look very good.

I think I'll take a look at the stars database and the messages here about
adding to it. I can hack together an addition to the stars database that
would put a number of statistical stars at the position of each globular
cluster.

The software should not be bothered by the additional stars, as they
will generally not be visible unless you are close to the cluster. And
othes have shown that Celestia can handle the data (apart from
the start-up time).

What you would really want is to have a billboard image of the globular
cluster that is shown from far away. As the user gets closer, it would
change over to stars.

This technique is also the way the whole Milky Way galaxy (and other
galaxies) should be done. The thing is that stars are point objects when
nearby, but a star field is an area light source when viewed from far
away. Although the individual stars are not discernable, the combined
diffuse glow is. Of course, for the galaxy, one then has to deal with the
dark nebulae - but one problem at a time.

Would it be possible to add a billboard class of objects? A class where
each object is represented by a billboard image, and that image can
fade away at a defined distance from it?

Actually to save time and effort...why dont you just add stars to stars.dat and leave out the billboard effect...Currently in realism the rendered galaxies look correct as viewed from a distance...They do not have a very high illumination...With that in mind all you really need are the stars...

Up to you as its your project, that was my 2 cents Good luck!

fjsummers wrote:Would it be possible to add a billboard class of objects? A class where
each object is represented by a billboard image, and that image can
fade away at a defined distance from it?

Yes. This is exactly what I want to do eventually for galaxies and globular clusters. It's in principle similar to what's already done for stars and planets. When they're distant, a billboarded proxy is rendered instead of an actual model. Once an object occupies one full pixel on screen, the proxy is fades out (until it's completely gone when the object is 4 pixels across) and replaced by a triangle mesh.

--Chris

One problem with using eliptical galaxy types is the type refers to the
ellipictiy has seen from earth. An E0 galaxy is not really a sphere,
it is a galaxy that looks circular from earth and it might be a
sausage or a pancake.

One limitation is that we really don't know what a given elliptical
galaxy looks like in three dimensions since we can only see
the two dimensional projection.