New spiral galaxy types rendering

[t00fri]

A few further considerations about galaxy color:
-------------------------------------------------------

As a reminder: galaxy catalogs include the following parameters per galaxy:

V_mag (visual magnitude)
B_mag (blue [photographic] magnitude)
SB (surface brightness) which is [magnitude/area]

There are now very convincing arguments why the
[V,H,S]=[Value,Saturation,Hue] parametrization of color is much preferrable for our purposes, to the common [R,G,B] color characterization. There are of course well-known formulae to express [R,G,B] - colors in terms of V(alue), H(ue) and S(aturation) .

The basic reasons for why the [V,S,H] basis is so suitable for our purposes are as follows:

-- V is fixed directly in terms of V_mag from the catalog. So we may forget about that parameter right away. Thus, only two are left.

-- Next: according to the above philosophy of a distance-dependent galaxy coloration, we may keep the Saturation fixed and low (~grayscale image!), as long as we are far away and at roughly constant distance from the galaxy. Note that S varies only the coloration-intensity NOT the underlying profile.

In these cases, therefore, the required color profile across a galaxy is uniquely dertermined by only ONE variable, HUE!
The range of the Hue variation will be determined by the difference of
V_mag - B_mag from the catalog!

-- When we approach or zoom in, we merely need to rise the Saturation parameter to intensify a certain color profile from largely gray to intensively colored. Very handy indeed!

Finally, all elliptic galaxies are known to be ~equally red since they are very old. So they are easy to take care of as to color...


Bye Fridger

[t00fri]

Instead of polishing around switches or sliders, I think it would be more useful to have the new galaxy rendering code in CVS, even if it is not 100% finished and perfect.
But I guess this depends upon Chris to approve the change...

steffens

Chris has promised to me (REALLY ) yesterday that he will be looking into the code, today...

I don't see any need for uploading at this stage to CVS. Toti's patch is available in the forum, so everyone is invited to experiment with it.

Moreover, Toti's patch is only part of the story: much of the recent progress sits in my Perl code, notably also the sequence of non-trivial quaternionic transformations that get the galaxy orientations right.
So far we never uploaded my Perl code to CVS, but everyone is welcome to get it from me.

In order to optimize the C++ code, we need to start from realistic catalog inputs about many different types of galaxies that may be directly compared with the corresponding DSS-photographs. Once we are certain about these crucial issues, coding is by far the less challenging part here.

We normally tend to upload only code to CVS that has a good chance of being almost final or at least appears /correct/. That is not yet the case for the new galaxy rendering. There are many leftover flaws that I know already. Tonight I shall try to improve on these significantly. Lets hope that Chris will also report back soon...

Bye Fridger

[steffens]

Toti's patch is available in the forum, so everyone is invited to experiment with it.

That's true, but that patch is very old (May 14), considering that the code is under heavy development. The most uptodate patch mentioned in the forum was from dirkpitt on Fri May 27, and this patch does not seem to be available anymore.

What concerns me is that anyone who wanted to have a look at the code would have no way to get the most current version. There is most probably little similarity of the code in Toti's patch with what toofri has today, so working on that is not very appealing.

If you want to keep experimental code from CVS, maybe a diff against current CVS, released every now and when and published in a not-so volatile place would be great.

Thanks to Toti, toofri, dirkpitt, ... for your hard work on this exciting subject! Regards,

steffens

[t00fri]

Toti's patch is available in the forum, so everyone is invited to experiment with it.

That's true, but that patch is very old (May 14), considering that the code is under heavy development. The most uptodate patch mentioned in the forum was from dirkpitt on Fri May 27, and this patch does not seem to be available anymore.

What concerns me is that anyone who wanted to have a look at the code would have no way to get the most current version. There is most probably little similarity of the code in Toti's patch with what toofri has today, so working on that is not very appealing.

If you want to keep experimental code from CVS, maybe a diff against current CVS, released every now and when and published in a not-so volatile place would be great.

Thanks to Toti, toofri, dirkpitt, ... for your hard work on this exciting subject! Regards,

steffens

OK, right now the situation with the galaxy project is unfortunately not very transparent: everything is highly experimental still, by far not ready for CVS. Also Chris is working on it. I know from his emails about what he is doing, but I have not seen any code from him yet.

All people who are interested and know how to code have grabbed Toti's code as a start and went on from there with own experiments. For example, in my new thread on users->Galaxy Colors, I have described the great advantages to use V,S,H color "coordinates" instead of R,G,B. I have also made public a corresponding C++ method that transforms HSV to RGB and illustrated the benefits in case of an example.

Other people might currently try other things. Chris is busy writing a new shader code for beautifully rendering elliptical galaxies that we generate in the code (instead of reading in templates like in the spiral galaxy case).

That's a fair account of the present status with this project I think. So I hope you understand that noone is very keen right now to perform CVS administative tasks . clearly, if you plan to contribute, you would be most welcome.

Actually the base code around Toti's patch is still not much different.
Of course I wrote extensive Perl scripts with lots of research entered in between, in order to get Steinickes NGC|IC catalog done for Celestia.

Bye Fridger