New sprites based nebulae

[W0RLDBUILDER]

I've been thinking. Back when I used to mod Far Gate, I replaced the explosion texture with a cloud of sparks, bearing a resemblance to a globular cluster. This made exploding objects look as if they were surrounded by glowing debris. Using the same kind of texture on a sprite could produce clouds of millions of stars, perfect for galaxies or globulars. Of course these wouldn't be real stars or fake-yet-still-3d stars like in Celestia's internal globulars, they'd be several thousand pictures of several hundred stars.

[Goofy]

Hi Cham. I'm following Celestia from more than one year,
but this is my first post, that I decided to write due to
my interest in the work you are doing with CMOD nebulae.
Please excuse me for my bad English.

I checked all of your given examples, and tried to modify
one of them as follows:

I decided to try to obtain the M 57 planetary, using your
shell5_sprites.cmod (because it has very nice small blobs).
I duplicated it 6 times giving the same space position and
new names, from AA to FF, and giving each of them the
correct color (or at least I tried to do it) and transparency ,
as measured with PhotoShop from a professional M 57 image,
this one:



I used the 6 cmods as PhoShop layers, one over the other, varying
their radius accordingly.
I'm sufficiently satisfied with the results, and flying around and
within the nebula is a beautiful experience:



Obviously this is only a first attempt, the overall shape is less than
correct, color saturation must be modifyed, so, if someone wishes
to try and/or modify it, I can publish here the .ssc file and the
changes that must be done to the cmod files headers.

Thank you very much, Cham, all this looks very stimulating for me.
I'm sure that there are a lot of nebulae that can be modeled in a realistic
way, at least aesthetically, if not scientifically.
Now I'm trying with the Eskimo nebula, I'll show results in a next message,
when finished.
Bye soon.

Goofy

[Fenerit]

Hi Goofy, your work is absolutely stunning. Because your deal with overlays, I was wonder about how the blend premultiplied propriety were to be of help in such a case, or if you have test it on nebulae's construction.

[Fenerit]

Cham, if you add some structures which resembles legs to your last morphology, the Ant nebula is ready.

[Fenerit]

A great feature of the Perlin noise function is that one may control the "granularity" and other features very easily by varying just a couple of parameters. That is certainly handy for rendering PN as well.

It's interesting. I'll check if the Perlin noise is included in Mathematica. However, I can control and adjust the granulity too with the Beta distributions in my models, by adjusting several parameters (actually, the parameter space of my model is very large, and the results are hard to expect since there are so many possibilities).

Maybe Texturing and Modelling. Third Edition. A Procedural Approach could be for you. All articles are written by Ebert, Musgrave, Peachey, Perlin and Worley and deal with both history of computer-generated procedural approach and their methods. Their math functions are presented as OpenGl GLSL shader language (c/c++ code) but it can be easily reckoned.

[Cham]

Another lobes variation from this morning. I'm eating nebulae at breakfast, now

lobes1.jpg

lobes2.jpg

I think I like this one pretty much. Need to make more experiments with the 18 parameters !

[Goofy]

Hello, here once again.
This time I wish to show you my work on another planetary, the Eskimo Nebula (NGC 2392), also known as the Clownface Nebula or Caldwell 39, this one:



My first experiment, using no blending:



Then I tried some suggestions by Fenerit, the first one was the use of "blend add":



Then I tried using "blend premultiplied", that gave me this very strange result



that I wish to investigate later.

IMHO the "blend add" #3 version looks more realistic.
What do you think?
Surely I have to make variations in order to find a better final result, but I'm sufficiently satisfied of the way in which things are evolving.
Moreover I'm waiting for further wonders from Cham....
Bye

Goofy

[Cham]

Cool, Goofy ! Welcome to the board, by the way

I'm also wondering about the blend premultiplied option. I don't understand what it's doing.

Are there any other options to the "blend" declaration ? Chris ?

1. blend add
2. blend premultiplied
3. blend substract ? (didn't tried)
4. ...?...

I'm also wondering if there's a way to define the add option to each sprite individually, instead of a global declaration in the material.
That would allow more rendering effects.

[Fenerit]

Gulp! No, that is not what I supposed were the multiply behaviour. Maybe the blob image doesn't need the transparent background, thus isn't useful. I'm unaware about other layer's propriety.

[Fenerit]

Hi Goofy; your work on the Eskimo nebula is really fascinating; it seem a real nebula. My perplexity concerns only its shape. I think Eskimo nebula is seen from side respect to the Fridger's plot below:



in which the white central lobe is one of the two red "shots" and the coronal brown is the blue torus. Do you have tried to shape it with the "butterfly" morphology?

[Goofy]

Hi Goofy; your work on the Eskimo nebula is really fascinating; it
seem a real nebula. My perplexity concerns only its shape. I think Eskimo nebula is
seen from side respect to the Fridger's plot below:

in which the white central lobe is one of the two red "shots" and the coronal brown
is the blue torus. Do you have tried to shape it with the "butterfly" morphology?

Hi Fenerit.
Thank you for your kind words, but if it looks nice is due to your precious "blend add" addition.
Anyway it's still a work in progress. so I didn't try t00fri's suggested shape yet, but
I think we could draw a full shape using three different Cham cmods :
the ring, the butterfly and the sphere ones.
BTW, many thanks to t00fri for the link to that paper, it opened me new horizons, I was sure that almost all planetary nebulae were approximately spherical, and now I see that I was completely wrong!

I hope we'll find a good compromise between real and fictional.
Anyway I'll have some problem to choose among the many "butterfly" shapes issued
up to now, while I have seen only two "rings" shapes (Cham, did I miss someone?).

I'll be a bit busy for a couple days, so probably no news from my side up to next
Monday, sorry.
Bye soon and happy week-end everybody.

Goofy

[t00fri]

Here are several further comments reflecting some of my principal considerations both as a professional (astro-particle) physicist and a longstanding Celestia developer:

Celestia strives for a reliable database of deep space objects (DSOs), as extracted from the best scientific catalogs (CDS Strasbourg). Being a 3D visualization, Celestia always requires trustworthy distance values and orientations in 3D space of the included DSOs. From Earth we only can see the 2D projection thereof by looking at the skyplane! An extraction of the required 3D orientation parameters from the 2D measurements is usually challenging and may invoke additional model assumptions. Successful examples of this task are the orbital parameters of multiple stars or also the 3D orientations and distances of galaxies. Globular orientations are trivial as long as the distribution of stars is assumed spherically symmetric. The distances of the 157 galactic globulars are pretty well known.

An important further constraint in Celestia is the spectral window for the imaging of DSOs. Celestia is limited by design to the visual range and hence imported images should strictly be of so-called true color. In Celestia.Sci a much wider approach to color is taken. There I use the concept of color profiles that allow a rather precise matching to the filter characteristics of scientific surveys (e.g SDSS, HST,..., infrared, ultraviolet,...). Multi-frequency astronomy ranges among the most powerful modern observational tools!

++++++++++++++++++++++++++++++++++++++++++
Let me specialize next to problems with a systematic rendering of planetary nebulae (PN) catalogs under these headings.
++++++++++++++++++++++++++++++++++++++++++

PN orientation parameters:
=====================

At CDS Strasbourg you find many smaller catalogs about PN. They list basic parameters (names, coordinates, sizes, spectral info) and distance estimates for about 170 of about 600 known galactic PN. Meanwhile there are also catalogs about PN in other galaxies (LMC, M31,...).

But unfortunately, all catalogs only list a single PN size parameter ("radius"), NOT accounting for elongated or more intricate PN morphology (like 2 lobes etc). For that reason there is also NO information about orientations of the many non-circular PN appearances!

Since we talk about rendering hundreds of PN from leading catalogs, an individual fiddling of the model orientation is impractical at the Celestia distribution level. Yet, this fact is not to demotivate PN affectionados from submitting add-ons with individual PN renderings.

Another critical comment concerns PN color.
========================================

If you google e.g. for the ring nebula M 57, using the "images" option, you get a large listing of M 57 photographs with entirely different colors and apparent shapes!

Hence it does not correspond to Celestia's scientific standards to just pick e.g. the yellow-green image above (Goofy) without making sure that you want to be as close as possible to the visual color rendering. Some knowledge about the physics that makes PN glow is certainly helpful here

Here are three examples of hires M 57 photos that look very much different from each other:

• Wikipedia image (see also Goofy above):
  
  
  
  
• This is an LRGB composite CCD image taken with an SBIG STL-11000M CCD and a RCOS 24-inch telescope. The image was taken by Adam Block from Mount Lemmon SkyCenter (Univ. Arizona).
  
  
  m57.jpg
  
  
• Finally this one (AP 12" F12.5 Mak-Cass, by Roland Christen),
  exposing a rich additional structure far outside the familiar ring structure. The additional structure becomes visible in infrared light / h-alpha filter!
  (from http://geogdata.csun.edu/~voltaire/roland/m57.html)

M57-5.jpg

I hope this illustrates that also a PN add-on needs some additional work and documentation when it comes to color! It's certainly NOT a good idea to just refer to images from Hubble/NASA because the source is "respectable".

Fridger

[MiR]

....I hope this illustrates that also a PN add-on needs some additional work and documentation when it comes to color! It's certainly NOT a good idea to just refer to images from Hubble/NASA because the source is "respectable".

Fridger

Just exactly what I meant. But what wonderful pictures ... from both (Hi, and welcome here, Goofy)!

Michael

[Goofy]

Here are several further comments reflecting some of my principal considerations both as a professional (astro-particle) physicist and a longstanding Celestia developer:
"...snip..."
I hope this illustrates that also a PN add-on needs some additional work and documentation when it comes to color! It's certainly NOT a good idea to just refer to images from Hubble/NASA because the source is "respectable".
Fridger

T00fri, have you got seen my signature?

Goofy

[t00fri]

Here are several further comments reflecting some of my principal considerations both as a professional (astro-particle) physicist and a longstanding Celestia developer:
"...snip..."
I hope this illustrates that also a PN add-on needs some additional work and documentation when it comes to color! It's certainly NOT a good idea to just refer to images from Hubble/NASA because the source is "respectable".
Fridger

T00fri, have you got seen my signature?

Goofy

Sorry, anything wrong on my part? I am not yet with you...

Fridger

[Cham]

On the previous messages, here's my point of view :

About the nebulae colors (and also the shapes), what I'm currently doing is **pure fictious** and "generic" nebulae. I don't pretend to do a reconstruction (or simulation) of any given real nebula : it's too hard to do and it would be very unsatisfying since we simply can't compete with nature (there are too much details and shapes in real nebulae). I still rely on real photographic pictures on simple flat billboards for the true stuff.

Since Celestia is only showing a **very small subset** of what is really "out there", there's plently of space in Celestia to add "semi-realistic" or even totally "fake" nebulae anywhere else, around false white dwarfs, pulsars and black holes, for example. Especially in other galaxies (Andromeda is a nice place to add lots of fictious addons). In all cases, there is no confusion possible (to my students, at least) when the addons are using names which shows that it's about a "fake" representation. For example, I'm calling all of my fake white dwarfs like this : "Naine blanche fictive" (in French, or "Fictious white dwarf"), and nebulae are simply called "N?buleuse", without any catalog number.

Celestia is calling for artistic expressions, even if there is no real scientific motivation ! To me, this is very important.

As Einstein once said : "Imagination is more important than knowledege" !

[Cham]

Lets stay on topic.

Here's an interesting experiment. I'm not sure I really like this model, but...

blobs1.jpg

blobs2.jpg

Here's the model :

shell29_sprites.cmod.zip

It's actually two models fusioned into one : one part is using the blend add option, while the other part doesn't. The rendering in Celestia is really interesting. I just don't like much the view from the inside because of a lack of details.

[Cham]

Strange. My Mathematica generator gives so much variations in results, just by some slight changes in parameters, that I don't know how to choose anymore ! It's like playing with an AlienSkin plug-in for Photoshop. You get so much variations in output by changing slightly the various sliders, it's getting boring after a while !

nebs.jpg

[Cham]

Yet one more parameter for the lobes models : a translation of lobes...

translation.jpg

[Fenerit]

Maybe that could be the main parameter for several correlations. Something that rules: for translation = x then color = y; or: for translation = x + 1 then color = z.