Pseudo Globular Clusters

[abramson]

Hi all,

I have updated the globular clusters' .dsc catalog that Selden keeps in his page. I have added absolute magnitude values, so that they are reasonably well represented in Celestia with the new galaxy rendering (the old catalog became useless since globulars turned out over-exposed). They are still rendered as galaxies, but they look nice from a distance. Flying around the Milky Way is much more lively with globulars around. I have also added a few obscure clusters that were mising, and alternative names.

The absolute magnitude data are from a catalog suggested in the Wikibook as a possible source for an eventual offcial catalog of globulars.

Find it here, at Selden's page, or here, at mine.

By the way, I believe it is time to have a new object class for globular clusters (and eventually a rendering of their own, but at least a class that can be turned on and off independently of galaxies).

Enjoy.

Guillermo

[danielj]

It isn?t working.I don?t know what format it have to be able to use,but simply saving in .txt format and putting in extras don?t works.Maybe there is a .ssc lacking.I have a M4.ssc and M4.dsc from an old addon of Selden,but Celestia didn?t recognizes it.

[revent]

It isn?t working.I don?t know what format it have to be able to use,but simply saving in .txt format and putting in extras don?t works.Maybe there is a .ssc lacking.I have a M4.ssc and M4.dsc from an old addon of Selden,but Celestia didn?t recognizes it.

Daniel, the filename has to end with '.dsc' to be recognized and loaded by Celestia. Just downloading the file and putting it in your extras directory WITHOUT RENAMING IT will make it work. When you read the file and then tell your browser to save it in text format, the browser will add a .txt extension. What you want to do it right-click on the /link/, and choose 'Save as'.

BTW, .ssc files define objects within solar systems, i.o.w things orbiting a star (hence Solar System Catalog = ssc). This addon has nothing to do with ssc files.

[danielj]

I rename to .dsc and still didn??t work.
First of all,the file is saved in a HTML format.
I copy all the content,paste in notebook and change the termination to .dsc.Either in Celestia 1.5.0 SVN R4273 or Celestia 1.5.1,it didn?t work.

[selden]

I've put it in a Zip file.

http://www.lepp.cornell.edu/~seb/celest ... usters.zip
(6KB, expands to 30KB)

[danielj]

Now,It works.Thanks.

[abramson]

There. I zipped mine too, to avoid confusion with the format. Really, Daniel, I didn't expect problems with a .dsc file, especially from old Celestians...

If anyone finds erros in the catalog, please let me know so I correct them...

Regards,

Guillermo

[chris]

I think it should be quite easy to add a globular cluster type to Celestia. There was some discussion a while ago about this, and I think I was arguing then that we should just use galaxies to represent globulars, because the line between dwarf galaxy and globular cluster seems rather fuzzy. I'm of a different mind now: even if there are some intermediate cases (Omega Centauri?), there are many more objects that are clearly distinct from galaxies. Globular clusters could be adequately represented in the same way that spherical galaxies are now. Perhaps there's some additional data we could use to add further realism, say spectral profiles?

--Chris

[selden]

It seems to me that it might be reasonable to try to represent their shapes (radius, density profile, etc) algorithmically. There are certainly enough parameters specified in Part III of William Harris' CATALOG OF PARAMETERS FOR MILKY WAY GLOBULAR CLUSTERS at http://physwww.physics.mcmaster.ca/~harris/mwgc.dat
Unfortunately, I don't know what most of them actually mean.

[abramson]

Perhaps there's some additional data we could use to add further realism, say spectral profiles?

The mentioned catalog provides "Integrated color indices (uncorrected for reddening)", U-B, B-V, V-R, and V-I, which could be used for the hue. Represented as elliptical galaxies, I believe, they are too blue for their mostly Population II stars (metallicity is also given in the catalog, so is an integrated spectral type for most of them).

Concerning their shape, there are a central concentration, c = log(r_t/r_c), an ellipticity e = 1 - (b/a), and three radii: core, half-mass and tidal ("border"). (r_t and r_c and tidal and core radii, resp.)

Guillermo

[chris]

Perhaps there's some additional data we could use to add further realism, say spectral profiles?

The mentioned catalog provides "Integrated color indices (uncorrected for reddening)", U-B, B-V, V-R, and V-I, which could be used for the hue. Represented as elliptical galaxies, I believe, they are too blue for their mostly Population II stars (metallicity is also given in the catalog, so is an integrated spectral type for most of them).

Concerning their shape, there are a central concentration, c = log(r_t/r_c), an ellipticity e = 1 - (b/a), and three radii: core, half-mass and tidal ("border"). (r_t and r_c and tidal and core radii, resp.)

Given all this available data, it does indeed seem that an algorithmic approach to rendering would be best. A sprite technique might be easiest to implement, with dynamically generated sprite locations, sizes, and colors instead of the template derived sprite properties used for galaxy rendering. But, my intuition is that a geometric approach would give better appearance. The ellipticity of some of globulars would make the implementation a bit complicated--it'd be much easier if we could assume sphericity.

--Chris

[t00fri]

Perhaps there's some additional data we could use to add further realism, say spectral profiles?

The mentioned catalog provides "Integrated color indices (uncorrected for reddening)", U-B, B-V, V-R, and V-I, which could be used for the hue. Represented as elliptical galaxies, I believe, they are too blue for their mostly Population II stars (metallicity is also given in the catalog, so is an integrated spectral type for most of them).

Concerning their shape, there are a central concentration, c = log(r_t/r_c), an ellipticity e = 1 - (b/a), and three radii: core, half-mass and tidal ("border"). (r_t and r_c and tidal and core radii, resp.)

Given all this available data, it does indeed seem that an algorithmic approach to rendering would be best. A sprite technique might be easiest to implement, with dynamically generated sprite locations, sizes, and colors instead of the template derived sprite properties used for galaxy rendering. But, my intuition is that a geometric approach would give better appearance. The ellipticity of some of globulars would make the implementation a bit complicated--it'd be much easier if we could assume sphericity.

--Chris

That's what I did in galaxy.cpp for irregular galaxies which didn't fit to a particular template. There are well-known statistical algorithms [acceptance-rejection method (Von Neumann)] that I used, which allow to generate easily arbitrary non-uniform distributions of points in 3d.

That way I have also generated the dark lanes in galactic planes etc.

F.

[abramson]

Once an appropriate class is defined, with its own parameters, one could write a perl script to read Harris's catalog and convert its data to Celestia format.

I imagine that just U-B and B-V indices should suffice to determine the RGB color. But I'm not sure. Simpler: just the integrated spectral class. Rendering with "blobs", like galaxies, should introduce some randomness. Metallicity and integrated spectral class could be given as info on the upper left corner, like in other cases, together with distance, etc.

For the shape, revolution ellipsoids with a smooth density profile given by too parameters, representing the core and the tidal radius. I presume the ellipticity refers to the shape as seen from Earth. There is no indication of other orientation of the ellipsoids in the catalog.

Guillermo

[t00fri]

I might add that I have already since years a Perl script that reads out ALL other DSO's besides my galaxies from Steinecke's revised NGC catalog and converts them to Celestia format. It's just the complement to my galaxy data. This gives another ~10000 DSO's consisting of globulars, open clusters and nebulae. All I need to change is an object flag...

Steinecke's coordinates are all fit to the photographic DSS I and II data...
and hence VERY accurate.

F.

[abramson]

I might add that I have already since years a Perl script that reads out ALL other DSO's besides my galaxies from Steinecke's revised NGC catalog and converts them to Celestia format. It's just the complement to my galaxy data. This gives another ~10000 DSO's consisting of globulars, open clusters and nebulae. All I need to change is an object flag...

Ohhhh!!!! While reading galaxies.pl I was just thinking of adapting it for that use!!!

BTW, sorry for some repetition in previous post, concurrent with others'...

Guillermo

[danielj]

I know maybe it?s not adequate to represent hunthousand of globular clusters,but the old model of M4 by Selden works perfectly even in Celestia 1.5.1.I think it?s very realistic...

[chris]

For the shape, revolution ellipsoids with a smooth density profile given by too parameters, representing the core and the tidal radius. I presume the ellipticity refers to the shape as seen from Earth. There is no indication of other orientation of the ellipsoids in the catalog.

Is there an orientation parameter? The ellipticity isn't too useful without knowing the long axis.

--Chris

[abramson]

No orientation. Just the ellipticity. I presume it represents the ellipticity of the apparent ellipse (the ellipsoid as seen from Earth). We could ask Dr. Harris. I'll do that.

Guillermo

[abramson]

Guys, I've got an answer from Bill Harris, who's an expert in the field of globular clusters, regarding their spatial orientation.

As we presumed, the ellipticity reported in the catalogs is the projected one as seen from Earth. Only for a couple of clusters, there may be enough information about the true spatial orientation of the spheroids.

Here's his answer to my message:

Dear Guillermo,

The published ellipticities are the ones projected onto the sky,
that is, they are 2-dimensional projections of the true 3D spheroid.
So all we can say for sure is that the true 3D shape must be AT LEAST
as elliptical as the published value.

They are usually assumed to be oblate spheroids, because the ellipticity is thought to be due to overall rotation of the whole
cluster. This is a different
situation from E galaxies, where the ellipticity is often that of a
prolate spheroid and is due to internal anisotropy of the stars' orbit
distributions.

I think there may be only one or two clusters for which enough detailed
information about the internal dynamics is known to decide on its true
3D shape. I'm thinking of 47 Tuc and Omega Centauri. This doesn't really
help your rendering problem, I guess. Probably the best thing to do is
to assume they are nearly-spherical oblate spheroids,
take the published e-values, and multiply them by some average projection correction factor, assuming that the true orientations of
the rotation axes in space are
all random. This is actually a solved problem -- it was done years ago
for E galaxies where the e values are larger. I can't recall the factor
offhand but it can be found.

All of this may not help very much!

Bill Harris

[fsgregs]

Hi.

I was wondering if the globular cluster dsc file could be edited so that clusters are classed as "open clusters" instead of "galaxy". that way, we could turn on their labels separate from galaxy labels. I tried it and it didn't work. Nothing showed up. However, since 1.5.1 has an open cluster class in the render menu, what does it do exactly? Do we have any open cluster examples available to view in 1.5.1? Would one use a dsc file to define an open cluster? Could not globular clusters be called "open clusters" for the sake of the add-on?

Frank