Well I thought that a collection of pictures obtained with celestia.Sci could help to bridge the waiting time until I feel ready to start with a public version. Anyhow, I am working hard to make progress...

Enjoy,
Fridger

(sorry, I jumped too fast...)
Guillermo

Here is a starter:
++++++++++++++++++++++++++++++
celestia.Sci uses the measured and tabulated B-V color magnitude of galaxies for rendering their color appearance. Unlike Celestia, there is now an extra entry of B-V in galaxies.dsc
++++++++++++++++++++++++++++++

A galaxy with B-V <~ 0.5 is blue while galaxies with B-V >~ 0.7 are getting increasingly orange in case of the SDSS color profile.

Here is a nice spiral with B-V = 0.51 (M 74) and Hubble class Sc:

[Click on image by all means]


Please note the bright blue starforming clouds (H II) along the arms! While in the SDSS color profile (g', r', i') => (b, g, r) they render bright blue, in visual colors they are pink as discussed here:

http://forum.celestialmatters.org/viewtopic.php?t=437

The Sc galaxy NGC 99 has B-V much smaller than 0,5, namely B-V = 0.28. Hence it looks really blue:

[Click on image by all means]


Finally, the two galaxies NGC 3226 and NGC 3227 on the official celestia.Sci splash show the typical SDSS orange, since they both have B-V >= 0.75

More precisely, the E2 elliptical galaxy NGC 3226 has B-V = 0.86 while the SBa spiral NGC 3227 has also a high value of B-V = 0.76



Enjoy,
Fridger

No Guillermo

when you visited, I was just busy writing and uploading the next post that includes three neat galaxies in representative colors (depending on B-V)... There are more to come

Fridger

Here is a vast field with elliptical galaxies around NGC 4874. Since almost all ellipticals habe B-V >= 0.85, they are orange throughout.

Compared to SDSS imaging, this picture may be called "photorealistic"

[click on images by all means!]


and another orange fellow: Sa galaxy NGC 4698 with B-V = 0.83
Being of Sa morphology, its arms are tightly wound...


Enjoy,
Fridger

Excellent! Much better than the current Celestia rendering.

By the way, what star database are you using, Fridger? There seems to be a lot more than in the 2 millon stars of old. And Gaia has not yet flown.

G

_________________
Guillermo Abramson
Bariloche, Argentina

For now I am still cheating a bit, since I am "only" using the 2M star data as well. The stars you see around the galaxies above are just a separate layer of normal 2M stars..

However, I am working on a statistical implementation of faint stars with magnitudes beyond the limiting magnitudes of the star data set used. In the end, the result will look rather similar to the above. Yet the star patterns will NOT be repetitive as they are above, if you look carefully. The scheme uses my Automag mechanism: Just like I generate the stars for globulars randomly according to the relevant Hertzsprung-Russel diagrams, I do that as well with field stars upon zooming in beyond the rendering limit of the genuine star data set.

Galaxies without stars do not look nice at all...and we are talking here about faint stars beyond ~ 13m and galaxies of 14 - 18m! For the fewer bright galaxies, there is no need for this, of course.

The extra random stars will remain nameless for now. Their main "raison d'être" is to populate the galaxy views naturally.

Fridger

Fridger,

I seem to be seeing an interesting visual effect in your screen-grabs of galaxy images above. They look to me as if there were a bright light source in the very center of each of the galaxies, illuminating the arms from that center. The arms which are nearest the galaxies' centers seem brighter than those which are farther away from the centers.

_________________
Selden

Yes Selden,

that's the magic of OpenGL, creating much of the improved galaxy illumination. I coded the central glow in form of a set of (six) overlaid spherical sprites whose main purpose is to illuminate the neighborhood with light of B-V dependent color. The effect you described uses a few more little tricks, though. Like a procedural texture shader etc.

This setup corresponds to the fact that the central glow extends above/below the levels of the galaxy arms. Here is a reminder:



Edge-on spiral NGC 4565.
Credit: EUROPEAN SOUTHERN OBSERVATORY/SCIENCE PHOTO LIBRARY

Or e.g. here, for ESO 243-49 (Hubble):


Moreover, the central part always represents the brightest part of a galaxy (except Irr's)

Fridger

Here is another great face-on Sc spiral: M100 along with its two ~14.5m dim S0 satellite galaxies NGC 4323 and NGC 4328

Since the B-V color of M100 is measured to be 0.65 it is expected to range in between blue and orange within the SDSS color profile. As expected, here is the blue-orange rendering of celestia.Sci

[Click on image!]


And here is the SDSS photo for comparison. Don't overlook the two dim satellite galaxies!
The resolution of the celestia.Sci rendering is (of course) better than the SDSS one.



Enjoy,
Fridger

Here is a couple of amazing views with multicolored multi galaxies

1) The famous Stephan's Quintet (NGC 7320,...)
===========================================



There are many photos of the rather dim Stephan's Quintet, but often the colors are bad. Here is one of the best taken by Hubble



2) The region around NGC 70 (B-V=0.55 => blue spiral) with orange elliptics and spirals
=====================================================================
[Click on image by all means!]


Here is the best photo of these 14.5 - 16m dim galaxies (inverted BW) that I could find:

VERY similar I'd say...
Enjoy,
Fridger

Stunning!

One question: how is defined the base sprite (blob)? An alpha attribute for the diffuseness and a single tone?

I believe that Selden is referring to a slightly different thing. To some extent, I also see it, even though I appreciate the quality of this new rendering.

In some of the spiral galaxies the central bulge seems to be illuminating the arm. That is, they seem to be *reflecting* light. This does not happen in real galaxies, of course: the arms glow by themselves.

I am not sure what exactly produces this illusion, but in some galaxies is very powerful. M100, for example. 4698 also.

Is that what you meant, Selden?

G

_________________
Guillermo Abramson
Bariloche, Argentina

Sorry I wasn't clearer, but yes, that's what I meant. In the picture of M100, it looks to me like the sides of some of the arms are darker on the sides which face away from the center, creating the illusion that those arms are solid, not glowing, and being illuminated by the galactic center. Maybe there are multiple overlapping light sources? Perhaps it could be explained by a gaussian(?) light curve being applied to the galaxy's image to simulate the Population II stars?

_________________
Selden

Thanks, Christophe, Guillermo and Selden,

for bringing your observations about the galaxy lighting to my attention.

Let me try to explain in words (not code) what I do.

In the rendering code every point of the galaxy templates stores an associated brightness value ranging from 0 to 255. The brightness value of the arms is visible in the gray-level of the template: e.g. for an 'Sa' spiral exhibiting tightly wound arms,


To each such point I associate a centrally peaked small 2D texture (-> a sprite) that smears out the brightness of each base point over a domain of varying size, depending on the size of that sprite. The 3D galaxy shape is modelled by tangentially orienting each sprite to the true 3D shape in each base point. Hence each galaxy element acts itself as a light source. The central region, however, represents a much stronger source of light than the arms, since there I overlaid 6 circular sprites each with the same high brightness.

OpenGL takes additively into account all these light sources. So the region adjacent to the strong central illumination will get additively illuminated both by its own light and that radiated in from the strong sources in the central blob. Shadowing effects are also taken into account automatically in such an OGL approach.

To my opinion similar things happen also with real galaxies, where the spherical central area emitts most light that is also shed onto the nearby parts of the arms and adds to the light directly emitted from the arms (see my above edge-on photos). Towards the periphery, the central light becomes negligible and the emitted light from the arms also gets weaker.

So what is so different in principle between the OpenGL way of illumination and real galaxies?

Fridger

PS:

Perhaps as a further clarification, I have recompiled my code after switching off entirely all light sources of the central blob in m100. The result is attached. You see the weaker light sources in the arms remain and provide an illumination of the galaxy.



and here is again the link to the full image for a direct comparison:


Not a very desirable rendering