Page 1 of 8

New spiral galaxy types rendering

Posted: 25.03.2005, 17:28
by Toti
Hello all,

This is a preliminary rendering of the spiral types that I'd like to add to Celestia.

Image

They follow this scheme:
Image

A few notes:
-- All types are procedurally built using the same function, with varying parameters. The general layout of this function is very close to the original one (written by Chris)
-- Brightness has been increased for better visibility.
-- The rendering code won't be modified at all.

Happy Easter to everyone

Posted: 25.03.2005, 17:39
by ElChristou
Are these rendering inside Celestia?

Posted: 25.03.2005, 17:40
by Cham
Very good. Did you made the C++ code ?

Posted: 25.03.2005, 17:44
by t00fri
Toti,

Very good! I think we are getting on the way....I would make the structures somewhat more fuzzy.

Yes what about C++ code?

Bye Fridger

Posted: 26.03.2005, 00:39
by Toti
Thanks. This is done in Python (it's faster for this kind of essay)
I'll optimize and rewrite it in C++ soon.
There is a parameter to control the noise amplitude, so it is easy to make the models more fuzzy.

A question:
I made the radial distribution of points gaussian. Also, a transversal cross section of any arm shows a gaussian distribution pattern. Are these acceptable assumptions?

Posted: 26.03.2005, 01:05
by Cham
It appears to be a good assumption to me, but I'm not a specialist of galactic dynamics.

Can you add a color parameter ? It could be very usefull to give a weak tint of color to any galaxy.

Posted: 26.03.2005, 17:03
by maxim
Looks very good!

How does it show from side and ~45deg?

All types are procedurally built using the same function, with varying parameters.

How many parameters?

So it is easily possible to show/render smooth transitions between the depicted states of evolvement - both for educational reasons, and to adress galaxy types that are not clearly part of one defined class?

Is it an idea to make the fuzzyness parameter part of the render options - more natural vs more symbolic depicting?

maxim

Posted: 26.03.2005, 17:55
by t00fri
Toti,

some further thoughts...

1) I would be curious to see your math parametrizations that
you used.

2) Besides making the structures more fuzzy, one might also
consider for the Celestia display to inject randomized, bright
"nots"
into the spiral arms, like in reality...

Image
Image
Notice also the distinct colors

I am really excited since such an approach will allow for fast
mass rendering of ~ 10000 galaxies according to strict
scientific data (spiral classes & colors & visual angles as
seen from Earth)!

Bye Fridger

Posted: 26.03.2005, 18:55
by t00fri
Toti,

I accidentally replaced my original galaxy classification scheme and did not manage to find it anymore ;-)

So I replaced 'galaxies.jpg' with another nice illustration from ESA, where besides the schematic layouts some real photos are shown. This might give you some further clues...

Here are 2 other "Hubble Tuning Forks" that I found:

Image
Image

Bye Fridger

Posted: 26.03.2005, 20:26
by maxim
2) Besides making the structures more fuzzy, one might also
consider for the Celestia display to inject randomized, bright
"nots" into the spiral arms, like in reality...


My impression is, that it's more the spidery dark nebula structure that gives every galaxy that voluminous 3D look.
You think we could work out a smart procedural approach for this - as some kind of a second layer?

---> Should be oriented along the arms, but condense areas to (bubbly?) lines somehow.

Any idea if there are some PS/GIMP plugin filters that do similar things? Just as a starting point?

maxim

Posted: 27.03.2005, 02:22
by Toti
Maxim,

There are five parameters so far: torsion, noiseAmplitude, logRatio, bulgeRadius, barRadius. From the side the models follow the classical edge-on profile quite well.

maxim wrote:So it is easily possible to show/render smooth transitions between the depicted states of evolvement - both for educational reasons, and to adress galaxy types that are not clearly part of one defined class?
Is it an idea to make the fuzzyness parameter part of the render options - more natural vs more symbolic depicting?
The last should be possible if a view filter feature is added. One could assign a brightness value to each blob and render only those that are greater than certain threshold, etc. Given the normal point distribution, only the principal structures would remain.

Fridger wrote:
1) I would be curious to see your math parametrizations that
you used.

2) Besides making the structures more fuzzy, one might also
consider for the Celestia display to inject randomized, bright
"nots" into the spiral arms, like in reality...

Celestia already does this. (The rendering that I posted at the top of the page only shows the points distribution: there isn't any additional process made to the image)

Fridger wrote:
Notice also the distinct colors

I am really excited since such an approach will allow for fast
mass rendering of ~ 10000 galaxies according to strict
scientific data (spiral classes & colors & visual angles as
seen from Earth)!

As long as we could massively extract color information from some catalog, tinting is possible. Notice the radial-varying hue, too. Perhaps if we could relate a given galaxy type with its star population (ie. more common classes) as a function of radius we could also simulate that. Under normal view conditions it would be just the usual grayish galaxy image, but with an increased light sensibility the effect would show up.

Thanks for the new images.

Posted: 28.03.2005, 04:10
by Toti
Hello,

This is how Celestia renders the new models.
The image has a very strong brightness boost, just to show the details (otherwise galaxies are just the usual faint grayish patches)

Image

After making some adjustments I'll upload a patch for revision.

Posted: 28.03.2005, 16:37
by Cham
I can't wait to see this live in Celestia ! 8*)

Will it be available for the OS X version too ?

And what about the actual galaxies database ? It will need some correction, I guess.

Posted: 28.03.2005, 17:11
by t00fri
Cham wrote:I can't wait to see this live in Celestia ! 8*)

Will it be available for the OS X version too ?

And what about the actual galaxies database ? It will need some correction, I guess.


I can easily write a Perl script that extracts the full (corrected) NGC catalog of galaxies, including the respective new information needed to render the various galaxy classes. Of course, before starting that (straightforward) venture, I need to see Toti's C++ code...

Bye Fridger

Posted: 29.03.2005, 18:29
by ElChristou
Toti wrote:This is how Celestia renders the new models.


Very nice... Bravo!
Like Cham, I hope this will be part of the next update on osX...

Posted: 29.03.2005, 23:49
by chris
I'll check this change into the Celestia CVS tree when your patch is ready Toti. PM me if you need any help with the C++ conversion.

--Chris

Posted: 30.03.2005, 08:12
by maxim
Could we seed the new algorithm with the ability for galaxy evolution?
I mean - not really implement it, but lay out parameters and eventually a virtual interface so that it's easy, in a later stage, to see galaxies evolving by only providing a time parameter?

I would like to see galaxies changing through the millenias.
Not that Celestia is actually designed to handle such large timescales - but later maybe.

So just another question:
Does actuall catalog data provide rotational information, or information about proper motion for ANY galaxy - I mean beside redshift movement, which only applies to very distant galaxies. Any such known data for our local group members? Or for the Virgo cluster members?

maxim

Posted: 30.03.2005, 09:14
by Evil Dr Ganymede
maxim wrote:Could we seed the new algorithm with the ability for galaxy evolution?
I mean - not really implement it, but lay out parameters and eventually a virtual interface so that it's easy, in a later stage, to see galaxies evolving by only providing a time parameter?

I would like to see galaxies changing through the millenias.
Not that Celestia is actually designed to handle such large timescales - but later maybe.


They wouldn't evolve really over the scale of millennia (1000s of years) - it'd take millions of years at least to notice anything changing. And I'm not sure anyone really knows how galactic structure evolves anyway do they (it's not like we've been watching galaxies evolve for billions of years is it ;)).

Steinicke's Catalog!

Posted: 30.03.2005, 09:20
by t00fri
Hi all,

The most recent, largest and most accurate work on extended deep sky objects is
the

"Revised New General Catalogue and Index Catalogue"
by Wolfgang Steinicke from April 2004.

He is working on this since 1981!

The catalog comprises 14000 objects!

http://www.ngcic.com/steinicke/2004/explan.htm#3

It unifies Dreyer's "New General Catalogue
(NGC)", published in 1888 and its two supplements, the Index
Catalogue, published in 1895 (IC I), and the Second Index Catalogue,
published in 1908 (IC II). The NGC contains 7840 nonstellar objects,
the IC I and IC II add another 1520 and 3866 objects, respectively.

++++++++++++++++++++++++++
The main point is that besides containing lots of "criminalistic" work to
eliminate errors, Steinicke's catalog has the positions and size
parameters corrected from the Digital (Real) Sky surveys!
++++++++++++++++++++++++++

This huge and still ongoing work is done within the so-called

The NGC/IC Project
--------------------------------------------
The aim of the NGC/IC project, headed by Dr. Harold G. Corwin
jr. (Pasadena/USA), is to clean the historical NGC and IC. All
investigations start from the historical data, e.g. the observer's
publifications. Dreyer's data are often not consistent.

For code planning purposes, I list below the catalog entries and
notably further below an explanation of the object TYPEs!

Code: Select all

     FIELD                            Explanation
-------------------------------------------------------------------------------

NGC, IC             NGC- or IC-number. Some objects have an extension letter (A,B,C,...).
C                    Components. If two or more different objects use the same NGC- or IC-number, these are numbered as components (1,2,3,...). This is also applied for objects associated with the main entry.
D                   Dreyer Object. A * marks, that this is the object which is ment in Dreyer's catalogue (many objects with extension letters are not in the original NGC/IC).
X                   Second Line. If more then one line is used for an object (if the space for identifications is not sufficient) the lines are counted (1,2) here.
S                   Status. Status of the identification (see Tab. 2)
P                   Precision. Flag for high precision position
CON               Constellation
alpha, delta     Right ascension, declination. Equinox J2000.0
Bmag             Photographic (blue) magnitude
Vmag             Visual magnitude
SB                 Surface brightness (mag/arcmin2)
a, b                Larger/smaller diameter ('). If only one value (a) is given, this refers to the maximum size.
PA                 Position angle (?°). The position angle is only relevant for galaxies, and given if possible.
Type              Type of object
PGC               PGC-number. Object listed in the Catalog of Principle Galaxies [22]
ID1, ID2, ID3  Identifications
Remarks         Additional information


So you can see that besides the RA & Dec coordinates we have the visual & photographic (blue) magnitude, surface brightness, larger/smaller diameters, the position angle and a detailed TYPE information.

The latter is richer, however than Hubble's "Tuning fork"
classification, since the catalogue contains lots of other objects besides galaxies. So we should devise the corresponding class in
Celestia flexible enough and also contemplate about "synthesizing" the
other types automatically!

The most complex type information is given
for galaxies (e.g. the Hubble type), mainly adapted from UGC, ESO and
PGC or detemined by own estimation. Generally types differ between
various sources. There is a special type for ring galaxies, following
Thompson. Types for open clusters are included according to
Tr??mpler's classification. For globular clusters the concentration
class (I...XII) is given.

Code: Select all

   TYPE                              Explanation
--------------------------------------------------------------------

*, *2,..,*Grp      Star, double star,..,star group (asterism)
C, D, E, I, P, S    Galaxy: compact, dwarf, elliptical, irregular, peculiar, spiral (d=dwarf, B=bar, R=ring , M=mixed); other letters: from Hubble type or its extensions
R.., PRG      Ring galaxy (see [30]), Polar ring galaxy
GxyP         Part of galaxy (e.g. bright HII region)
OCL         Open cluster (if no Tr??mpler class is available)
GCL         Globular cluster (if no concentration class is available)
DN, EN, RN, PN      Nebula: dark, emission, reflection, planetary
SNR         Supernova remnant
NF         Not found




Finally, the Status field gives some useful info about the actual
contents of the catalog

    S Explanation NGC % IC %
    ---------------------------------------------------------------
    1 Galaxy 6494 76.9 4124 74.3
    2 Galactic nebula,
    Supernova remnant 141 1.7 102 1.8
    3 Planetary nebula 96 1.1 36 0.6
    4 Open cluster 686 8.1 39 0.7
    5 Globular cluster 118 1.4 8 0.1
    6 Part of galaxy
    (e.g. bright HII regions) 27 0.3 22 0.4
    7 Object already in
    the NGC- or IC 362 4.3 87 1.6
    8 IC-object already
    in the NGC 22 0.3 253 4.6
    9
    Star(s) 386 4.6 634 11.4
    10 Not found 115 1.4 248 4.5
    ---------------------------------------------------------------
    Sum 8447 5553



Bye Fridger

Posted: 30.03.2005, 09:43
by Paolo
I would like to add my two cents to the discussion.

IMHO There are at least a few points that remain to discuss in the implementation of this issue.

Sometime ago I made a post about a Deplhi code portion about a similar kind of problem.
http://www.celestiaproject.net/forum/viewtopic.php?t=3875

The Toti?€™s screenshots examples shows the type of galaxies as seen from the top view. Nothing is shown about the side view. The Globular and lenticular galaxies need a specific vertical distribution in order to fit properly the shape of the side view. Moreover the nucleus of the galaxies is thicker than the arms and has a shape similar to a very oblate sphere.

Another thing about the shape is related to the Halo of the galaxy. Usually the galaxies are surrounded by a globular halo of sparse stars and globular clusters of very old stars. I think that this point has to be considered too.

Another main issue is that some kind of LOD (level of detail) management is required to display properly thousands of galaxies in Celestia. This LOD must cover two aspects: the number of billboards and their transparency. So if a galaxy is very far from observer or the FOV is large a small number of billboards is used with a low transparency. On the other hand to display properly the milky way from inside the galaxy a large number of billboards can to be used but with a high transparency. This last issue can have a very bad impact on the performance of the program. So the LOD management should be different if the observer is inside the galaxy and a large number of stars are displayed or if the observer is outside the galaxy. Keeping somewhat visually smooth the transition from the two LOD modes when the observer is outside and travels toward the sun will be very complex.