Milky Way

[abramson]

Very pretty!

Now we need a proper orientation! (or a better projected image).

Guillermo

[t00fri]

Very pretty!

Now we need a proper orientation! (or a better projected image).

Guillermo

Still waiting for the reply of St?phane Guisard with info about the projection/coordinates/coordinate range they used.

If he provides a download slot for the 24k MW panorama, I plan to copy various nebulosities from there and mount them into mw.png after /careful/ size reduction.

Cheers,
Fridger

[t00fri]

For completeness, here is the 3k Milkyway by Nick Risinger, http://skysurvey.org/, that was quoted by Selden above. I only applied a quick, automatic star cleaning, to get an idea how it will look.

I think it has a lot of potential. It's more 'lightweight'. I like it. There is also way more nebulosity included, notably also around Antares/M4. There are also bigger sizes but it seems one has to contact the author.

Here are two analogous, consecutive shots that might give you an idea:

[Click on images by all means ]

mw_nc_3k_fri.jpg

mw_nc_3k_fri2.jpg

Fridger

[Reiko]

Is it possible to make a 3D version of this?

[abramson]

@Reiko: No. There ARE 3D models of the Milky Way (without consensus, though). But no model will look like the real Milky Way as photographed. The ONLY purpose of this model is to provide a realistic background to scenes as observed from our neighborhood.

A reasonable 3D model of the Milky Way to fly around is the one already present in Celestia (I believe Fridger made its template and coded the rendering of galaxies in general). It is a special case of an SB galaxy.

[t00fri]

Is it possible to make a 3D version of this?

Reiko,

in case you meant a 3D stereo version, the answer is YES. You just need to build Chris Laurel's new Cosmographia,

http://code.google.com/p/cosmographia/

which now enjoys Guillermo's nice galaxy background as well AND offers TWO popular 3D stereo displays. One is the familiar red-cyan method, which requires corresponding glasses, of course. Compiling Cosmographia is equally easy as compiling Celestia-Qt.

In case you meant a true 3d display like in case of my computer-based rendering of celestia's galaxies, see Guillermo's reply. Except that in my forthcoming Celestia.Sci the quality of the galaxies is strongly improved and all are properly colored. Here is a little illustration in case of NGC 3227 and it's elliptical partner galaxy NGC 3226:

[Click on image by all means]

n3227.jpg

But as Guillermo wrote already, in the computer-based automatic rendering of 3D galaxies, one can never expect photo-realistic displays. Here, the aim is rather to render all the information about galaxies contained in scientific catalogs in a precise, reproducable way. This includes, position, size, Hubble type, orientation in space AND color. The advantage is that one may include vast amounts of galaxies with this approach.

Fridger

[PlutonianEmpire]

I just got an idea...

Maybe this addon can be a default part of Celestia in a future release, where it behaves much like the way constellation borders and labels currently do, where it slowly fades after you get a certain distance from the Solar System?

Might this work?

[t00fri]

I just got an idea...

Maybe this addon can be a default part of Celestia in a future release, where it behaves much like the way constellation borders and labels currently do, where it slowly fades after you get a certain distance from the Solar System?

Might this work?

Indeed, I am experimenting since some days with various interesting options to render the Universe inside the local transparent sphere, on its surface and outside of it, along with a clever way of "transit". Clearly, I am widely using OpenGL blending techniques in my experiments. A next step would be a number of such "local spheres", each referring to the local environment of different galaxies, along with the possibility of travelling between them...

Some crucial advantages of such an approach are near at hand: The "local spheres" approach serves first of all as a means of embedding 2D projections of photographical images and available real or generated multi-million 2D star data as local backgrounds in a true 3D Universe simulation. With regard to cosmological extensions of Celestia.Sci, local spheres will play another distinguished role by separating space-time into local regions and far ones where General Relativity effects are manifest!

Actually, ChrisL's Cosmographia has placed all its stars on the surface of such a sphere and renders the solar system inside of it. However, unlike Celestia(.Sci) there is no way of transit to the "outside Universe" .

Some more concrete progress will be reported at CelestialMatters.

Fridger

[abramson]

A gradual transparency transit sounds very nice! Without it, the Milky Way background would not be adequate for a default distribution. Besides, we would need a better projection/orientation of the panoramic view.

[t00fri]

...
Besides, we would need a better projection/orientation of the panoramic view.

Guillermo,

actually, I meanwhile succeeded with a number of significant improvements of your add-on, but don't want to interfere with your nice, pioneering work. Also, as you sure know, I am not submitting work to the Motherlode as a matter of principle.

• I replaced your "sphere" with only 384 vertices by one from Selden with 703 vertices,
  http://lepp.cornell.edu/~seb/celestia/mw-msx.cmod. I'll presumably make an even smoother one myself, once I find some spare time.
  Your "sphere" is only a pretty rough approximation of a sphere due to it large facets! Besides having only 10 degree facets, Selden's model also works with the original MilkyWay image (without shifting the central meridian and adding a vertical flip).
  
  
• After using this much better spherical approximation, I also succeeded with a way better empirical alignment over the entire sky. Critical, far-apart cross-check points are the proper location of 47 Tuc (almost bordering SMC) in the South, along with the proper position of M4 & Antares and finally M3 close to the galactic North pole! Still, I am not 100% sure whether the vertical axis of the ESO original image corresponds to galactic coordinates (galactic latitude) or to equatorial ones.
  
  
• Motivated by this empirical success, I derived a formula for the alignment (axis, angle) based on my galaxy orientation subroutine in my Perl script, deepsky.pl, in src/tools/galaxies. Required inputs are the IAU 1959 coordinates of the galactic center $ra = "17 45 37.224"; #IAU 1959, $dec = "-28 56 10.23"; #IAU 1959, along with the appropriate position angle $PA and inclination $i. So fiddling now has an end . I urgently needed a general formula to position further spheres in other galaxies in a "one-shot" procedure by computer.
  
  
• I think the best location for the sphere is to choose it exactly isomorphic to the galactic coordinate frame. This has many advantages, notably if one wants to generalize these local spheres to many other galaxies as I described above. Also the sphere size is not arbitrary, with my plans in mind.
  
  
• Last not least, I focussed all my image manipulation know-how into making yet another further improved version of the MW image...

Cheers,
Fridger

[Fenerit]

Maybe the UV maps of the spheres could help. Some UV mapping programs measures the degrees of the grid's spaces; unfortunately aren't free softwares.

[t00fri]

Maybe the UV maps of the spheres could help. Some UV mapping programs measures the degrees of the grid's spaces; unfortunately aren't free softwares.

Not sure I understand...

Fridger

[Fenerit]

Not sure neither myself... Like said before, the more the sphere is refined the more is smooth the MW map; without visible quads (bends into the texture). To that is matching its UV map "grid" like in the images below:
uv map and such a grid would have the lines more closest. My wonder about, was the fact that to the grid's sectors would matching the angular separations amongs the MW sectors, like you know, and if they are on the original photo model like often is seen on astronomical maps, they could be fitted to the UV grid to find the minimum number of vertex/triangles with which to build the relevant sphere. I knew that there were softwares that made such a calcolus.