Celestia Forums

I'm sure it's been done, but for my own uses I am working on a program to generate Celestia data; a picture is worth a thousand words: http://w.tarazedi.com/image1095

The idea is to generate large batches of fully-usable star systems within a defined region of space for the benefit of sci-fi worldbuilders. I have functional single through quaternary star systems, with (reasonably) accurate population distributions. Planets, which will be rather more complicated, are coming soon. I will be making this available when I am done.

I have a 0.9 version of the planet textures (a sample of which may be found here: http://w.tarazedi.com/image1094 ) available at this location: http://w.tarazedi.com/Atlas The texture pack has about a hundred files and is about 30MB. All textures are 1024x768. There are various layers including maps, bumps, specs, nights, and clouds. When the planet building part of the program is complete, planets will be randomly generated and classified according to the following scheme to choose the textures which might apply to them:



Enjoy.

Can't download anything from that link. I go to http://www.tarazedi.com/files/StargenTextures.exe and I end up getting redirected to http://w.tarazedi.com/About .

--- edit ---

Sorry about that. My .htaccess files are a bit weird because of how my host disables hotlinking. Check the top of this page; I added a link to the textures that WILL WORK. http://w.tarazedi.com/Atlas I have also made it a .zip file as suggested.

Later I'll have to figure out why the forums can't link a file. *sigh*

--- edit ---

Hotlinks do not work, but it does work if you download it from the page.

So far the star generator makes complete and more aesthetically-pleasing star systems, and orbital elements and temperatures for planets... so I'm about 75% done. If you want to play with the previous version (star systems only): http://www.tarazedi.com/files/backup%20 ... Sept10.php

Right now my database says my 15 hottest planets are:



The Va Gushthyuak system is a binary with an M2Ia red supergiant and a DB9 white dwarf, separated by 13.7 AU and orbiting in 10.9 years.

Mneme wrote:So far the star generator makes complete and more aesthetically-pleasing star systems, and orbital elements and temperatures for planets... so I'm about 75% done. If you want to play with the previous version (star systems only): http://www.tarazedi.com/files/backup%20 ... Sept10.php

Hi Mneme, I'm playing with your Star Generator and I find it very interesting.
Just curious: why the "cube" area?
Why not a "sphere" area or even a "random" one?
Thank you for your addon.
Bye

Goofy

Laziness, to put it bluntly. CBA to come up with the math.

Mneme wrote:Laziness, to put it bluntly. CBA to come up with the math.

Mine was not a critic, obviously.
But if you'll go on with this project (I strongly hope so!), could you think about the possibility of a Gaussian star concentration?
I would like to try it for "fictitious" globular clusters.
If you'll try this solution, I suggest you to split your addon, i.e. with and without planets generation.
Bye and thank you once again, and take your time.
Bye

Goofy

Goofy wrote:I would like to try it for "fictitious" globular clusters.
..
Goofy

Goofy,

how about having some fun with my "crazy" CM thread: "Imagine the Solar System was part of a Globular Cluster"?

http://forum.celestialmatters.org/viewtopic.php?t=422

My shader stars satisfy all astrophysical constraints including a realistic luminosity (or mass) distribution and a correct Hertzsprung-Russell relation with adequate distributions of star colors. Moreover the normal star rendering and the rendering of the globular clusters is done with the SAME shader code.

Then that's how stars should look like:

[Click on image by all means!]

Soon users will be able to land on any one of the cluster stars which will provide an amazing sky experience ...

Fridger

t00fri wrote:

Goofy wrote:I would like to try it for "fictitious" globular clusters.
Goofy

Goofy,
how about having some fun with my "crazy" CM thread: "Imagine the Solar System was part of a Globular Cluster"?
http://forum.celestialmatters.org/viewtopic.php?t=422
Soon users will be able to land on any one of the cluster stars which will provide an amazing sky experience ...Fridger

Hi Fridger.
Yes, I've already seen your incredible results on Celestia.Sci globular clusters rendering (and even more on galaxies rendering, wow! ) and cannot wait to put my hands on it.
But, meanwhile, I can use Mneme's fictitious star generator, in order to give my students an idea on how the sky could be as seen from an inner planet.
I found somewhere Rassilon's globular cluster generator, that gives a strangely flat globular shape, at least to me, and this new one may give more real results.
Moreover it's still a work on progress, so can be issued to obtain very nice results, provided that we'll accept them as unreal.
Just to give an example, I tried this combination of queries:
Star number 10000 10000 10000 10000
Cube size 100 50 30 15
Star code 666 777 888 999
Below you can see the results.
Despite the cubic shape (but not so evident), I like it.
Waiting for Celestia-Sci wonders, bye.

Goofy

OK, your present results look pretty good, except for the star sizes in the core regions, perhaps. They are looking too similar to me.

Much success,
Fridger

t00fri wrote:OK, your present results look pretty good.
Much success, Fridger

Thank you, Fridger.
I'm sure that Mneme will make a good tuning of his Star generator, and this will give much better results.
Let's wait for it.
Bye

Goofy

Damn that's a lot of stars. I was wondering how clusters would look; not too bad it seems. Unfortunately the new version for performance limits to 999 per query because it does a lot more work than the version I have up there, but you can run it as many times as necessary.

Currently my position data calculates like this:

$edge = angular width of $cube;
$ra = rand( $ra - $edge, $ra + $cube2 );
$dec = rand( $dec - $edge, $dec + $cube2 );
$dist = rand( $dist - $edge, $dist + $cube );

I'm not clear on how to make a Gaussian distribution, and while I am clear on at least one way to make a spherical distribution, I had some trouble getting PHP to play nice with it. I was converting the input to Cartesians, generating a position, and converting it back, but I wasn't having any luck on that last step. I figure that I can use my current method but weight it toward the center by using two rands per coordinate as a 0 to 1 and multiplying them together, then multiplying that by the cube size. That would fuzz out the edges a lot and remove the cubic appearance. If anyone can show me the math for a better way I'd consider using it; I'm not an astronomer, programmer, or mathematician; I just play with things.

If you were using 666, 777, etc. you are overlapping star numbers on that version because it didn't populate a database. Try 60000, 70000, etc. Also if you want more blue stars and/or giants within the core, you can set those, if you didn't already. The default probabilities for star types are set to approximate estimates I found for the solar neighborhood, which for main sequence stars is:

if($starclassrand > 5000000){ $starclass = "L"; $starcoeff = 0; }
if($starclassrand <= 5000000){ $starclass = "M"; $starcoeff = 10; }
if($starclassrand <= 605000){ $starclass = "K"; $starcoeff = 20; }
if($starclassrand <= 380000){ $starclass = "G"; $starcoeff = 30; }
if($starclassrand <= 150000){ $starclass = "F"; $starcoeff = 40; }
if($starclassrand <= 30000){ $starclass = "A"; $starcoeff = 50; }
if($starclassrand <= 6500){ $starclass = "B"; $starcoeff = 60; }
if($starclassrand <= 30){ $starclass = "O"; $starcoeff = 70; }
if($starclassrand <= 2){ $starclass = "WR"; $starcoeff = 80; }

There is a 10% probability of a white dwarf, and logarithmic probability of each successive luminosity class, since I didn't really find any good estimates to work with.

Today I'm gonna see if I can finish off planet calculation... this is gonna be interesting... *grumble*

t00fri wrote:... except for the star sizes in the core regions, perhaps. They are looking too similar to me. Much success, Fridger

Hi Fridger.
You are absolutely right, and moreover I think there are a bit too many external blu supergiants.
Perhaps Mneme should give another kind od adjustment in the options, e.g. giving as "0" really NO blu (or red) stars.
Anyway he is working on it, so let's see what happens.
Bye

Goofy

0 blue skew means normal color distribution, which is to say getting blue stars is like winning the lottery.

Mneme wrote:If you were using 666, 777, etc. you are overlapping star numbers on that version because it didn't populate a database. Try 60000, 70000, etc.

Thank you, but I was already using 6660000 etc, here I wrote three digits only just to show what I did.
The 40000 stars are all there, believe me, and it's a pleasure to fly within!

Mneme wrote:Also if you want more blue stars and/or giants within the core, you can set those, if you didn't already.

Yes. I did. But I see that if I put "0" at blu or red stars, there are some blu or red supergiants in the external cloud.
I think it would be better to have the real possibility to ask for "0" blu or red supergiants.

Mneme wrote:The default probabilities for star types are set to approximate estimates I found for the solar neighborhood....

Mneme, this is Fridger's business, but I'm sure that globular clusters have an H-R Diagram very different compared with our neighborhood's: in globulars most stars are population 2, with a small percentage of red and blu giants and supergiants, if I'm right (Fridger?).

Mneme wrote:Today I'm gonna see if I can finish off planet calculation... this is gonna be interesting... *grumble*

Nice, waiting for.
And please, if possible, give the possibility to chose "no planets", too.
Bye

Goofy

Just finished, so here the new globular: 107000 stars, in tel concentric cubes, with decreasing size and star number going to the center, and increasing red stars value going to the center.
The stc file is 54MB, and my machine gasps a bit starting Celestia, but when loaded the cluster can be navigated without problems.
Hope you'll like it, if you don't mind the cubic shape, but the middle image is very nice, in my opinion.
Bye

Goofy

Goofy wrote:Mneme, this is Fridger's business, but I'm sure that globular clusters have an H-R Diagram very different compared with our neighborhood's: in globulars most stars are population 2, with a small percentage of red and blu giants and supergiants, if I'm right (Fridger?).

Yes, this is true. If anyone can show me some good numbers on spectral class and luminosity class distributions in clusters of age X, I would love to implement that.

I may include a skip planets option, good idea.