It is currently Wed, 24-05-17, 0:23 GMT

All times are UTC




Post new topic Reply to topic  [ 41 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
 Post subject:
PostPosted: Wed, 26-09-12, 21:34 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
To me the feeling of over enlightening from the center is a question of tuning, if Fridger is able to add some more "dust" all over the template this impression will disappear. But again Fridger, how is defined the base sprite (blob)? An alpha attribute for the diffuseness and a single tone (related to the brightness data from the template), right?


Top
 Profile  
 
 Post subject:
PostPosted: Wed, 26-09-12, 22:27 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
ElChristou wrote:
To me the feeling of over enlightening from the center is a question of tuning, if Fridger is able to add some more "dust" all over the template this impression will disappear.


I think I don't have overlighting from the center. Moreover, for anyone who thinks that I do, I have implemented a sophisticated key adjustment that is able to tune smoothly BOTH the central galactic lighting strength AND the brightness of the field stars in the environment. The trick is that the central galactic brightness varies periodically between its minimal and maximal value, while the star brightness increases/decreases monotonically if the familiar keys ']' and '[' are pushed. This gives a most convenient and flexible single key adjustment of all critical lighting parameters!
Same for globular clusters, actually.

Quote:
But again Fridger, how is defined the base sprite (blob)? An alpha attribute for the diffuseness and a single tone (related to the brightness data from the template), right?


I am not sure what you really want to know here. Also I don't know how to answer precisely without using complex formulae and specialized OGL terminology.

I have tried to be precise in my previous post already ;-) . First, there are the points that are read from the template. They contain the morphological info about the galaxy as well as the associated brightness values for each galaxy point. The Blob structure is like so:

struct Blob
{
EIGEN_MAKE_ALIGNED_OPERATOR_NEW

Eigen::Vector4f position;
unsigned int colorIndex;
float brightness;
};

The colorindex ranges from 0..255 and maps the radial distance of the Blob from the galaxy center to a particular color.

Next a rgb color vector is built up as a linear superposition between orange and blue tones, depending on the profile used (e.g. SDSS type). The actual mixture depends on the B-V data and the radial position in the galaxy (<=> colorindex).

Next the alpha value is composed that includes as factor the original brightness values and various other specialities. All this is transferred to the GLSL fragment shader , where the rgb color is modulated with a procedural sprite texture that is centrally peaked as I previously described.

There is actually another new pair of keys '(' and ')' that allow to adjust the rgb color saturation with everything else fixed. This allows to switch to grayscale rendering of galaxies and globular clusters everywhere or by default only at small scale (<=> visual observation!), along with various other more complex tunings.

Fridger


Top
 Profile  
 
 Post subject:
PostPosted: Thu, 27-09-12, 7:23 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
t00fri wrote:
ElChristou wrote:
To me the feeling of over enlightening from the center is a question of tuning, if Fridger is able to add some more "dust" all over the template this impression will disappear.


I think I don't have overlighting from the center. Moreover, for anyone who thinks that I do, I have implemented a sophisticated key adjustment that is able to tune smoothly BOTH the central galactic lighting strength AND the brightness of the field stars in the environment. The trick is that the central galactic brightness varies periodically between its minimal and maximal value, while the star brightness increases/decreases monotonically if the familiar keys ']' and '[' are pushed. This gives a most convenient and flexible single key adjustment of all critical lighting parameters!
Same for globular clusters, actually.

I was talking about a feeling which is subjective by definition. To me personally I'm fine with the brightness I see in your last shots but I do understand what means Selden and I think i'ts more related to the lack of global "density" (the fluffy or dusty overhaul missing layer) than a real problem with brightness intensity in the core...

t00fri wrote:
Quote:
But again Fridger, how is defined the base sprite (blob)? An alpha attribute for the diffuseness and a single tone (related to the brightness data from the template), right?


I am not sure what you really want to know here. Also I don't know how to answer precisely without using complex formulae and specialized OGL terminology...


I was thinking in another way to add realism specially for dark clouds but I realize that at this stage better not rebuilding the system from zero, the gain would probably not be enough in relation to the necessary coding AND the rendering power needed... The basic idea was to render the sprites in two passes, one for the alpha (it's radial transparency from the center) and one with a planar gradient. The planar gradient would be generated, from a 100% black to a 100% white throughout all declinations, depending on the positions of the sprite vs the observer vs the center of the galaxy. Then adding the color and brightness layer etc... Too complicated! Let's forget! :lol:


Top
 Profile  
 
 Post subject:
PostPosted: Thu, 27-09-12, 13:32 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
Christophe,

since quite a while actually, I do experiment in parallel with genuine volumetric OGL rendering techniques for galaxies that are becoming increasingly important e.g. in medical visualization. As far as I can tell, real good performance with a large number of objects to be rendered requires powerful graphics cards (GPU computing!) and fast CPUs that --for now at least-- I am not willing to assume for every potential user of celestia.Sci.

For the time being, it seems more relevant to me to implement significantly extended galaxy sets along with a much more finely subdivided morphological classification. The implementation of deeper, redshifted galaxies are a prerequisite for building a bridge to realistic cosmological visualization.

Fridger


Top
 Profile  
 
 Post subject:
PostPosted: Fri, 28-09-12, 15:43 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
t00fri wrote:
...
For the time being, it seems more relevant to me to implement significantly extended galaxy sets along with a much more finely subdivided morphological classification. The implementation of deeper, redshifted galaxies are a prerequisite for building a bridge to realistic cosmological visualization.


Yum! That sounds pretty sexy! :o


Top
 Profile  
 
 Post subject:
PostPosted: Sat, 29-09-12, 19:05 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
celestia.Sci's entirely new code for the rendering of all globular clusters is really fun!

As an impressive example, here is the dim globular NGC 2419:

Image

The faintest globular stars (appmag <= 25m!) match well with really big sized stars. The two "super bright" stars on the right are only 7.2m and 7.9m, respectively! The not-so-bright one near the bottom is 11.16m. Most visible stars in NGC 2419 are within an app.mag range of 20m - 25m. i.e extremely faint! It is apparent that celestia.Sci is able to cover a huge range of faint magnitudes compared to Celestia!

Note:
I use my new GLSL star shader BOTH for rendering normal stars AND for rendering all globular cluster stars! My globular shader stars are generated from a realistic luminosity (or mass) distribution and a correct Hertzsprung-Russell relation with adequate distributions of star colors. While my code is quite different from that of ChrisL or more precisely VESTA@ASTOS GmbH, we both start by composing the shader stars of a disk with varying size AND a Gaussian halo, in order to be able to render the huge range of brightness on normal monitors.

Due to the efficient shader code one may zoom in fast into the extremely faint 20-25m globular stars with the result indicated in the spherical region of the above display of NGC 2419!

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

Enjoy,
Fridger


Last edited by t00fri on Mon, 01-10-12, 14:06 GMT, edited 2 times in total.

Top
 Profile  
 
 Post subject: re
PostPosted: Sun, 30-09-12, 7:52 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 749
Location: N 42.38846 W 83.45456
nice

a bit late but ...

the galaxies look like the photos
and that i think is a bit off -- let me explain

in the darkroom if a see a image with a overexposed spot from a light source it is an automatic minor burn under the enlarger -- for aesthetic reasons . Enlargements just " look better" that way

And it has nothing to do with science ( astronomy that is) it is just what the eye/brain "sees" and finds to "look good"
from the example in the last page
something like this maybe
BEFORE
Image
AFTER
Image

with the center Aprox ? 25% darker and the fall off a lot longer

just my point of view so .

unfortunately looks and long exposure photos do not always work well together
And i have done many 3 to 5 to 8 hour long exposures . If the contrast is too high then it just looks " off"

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
 Post subject:
PostPosted: Sun, 30-09-12, 8:29 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
Thanks for your analysis, John.

I agree that there might be a general preference of overexposing light sources. Looks somehow more "dynamical". I confess I am also subject to this preference and just set the center brightness of the displayed galaxies according to my taste rather than a more objective measure.

Since there will always be a considerable subjectiveness involved, celestia.Sci offeres a sophisticated key adjustment for precisely this region and the background star brightness in one go. I described this novel one key mechanism already further up to Christophe:

http://forum.celestialmatters.org/viewt ... 0&start=16

Perhaps it would be fun if I'd produce an animation video of this fancy adjustment mechanism, so people can see how it works in practice. Also the amount of color saturation can be adjusted down to pure grayscale with a second pair of keys '(' and ')'.

This mechanism is to reflect the fact that within the visual range, i.e. at low galaxy resolution and brightness, the amount of available light is too low to see visually any colors of dim deep sky objects. So by default, in celestia.Sci the color saturation increases with the amount of zooming-in ("telescope mode" as part of Automag). When the objects cover ~2/3 of the OGL canvas, they appear fully colored. This effect is very pleasing. However, purists may switch this off and --in the extreme case-- only deal with grayscale imaging of galaxies, globulars etc.

Finally I should add that --unlike Celestia--in celestia.Sci the visual appearance of celestial objects will not be anymore the main guiding principle. This arises due to a much generalized visualization of the colored Universe based on an inclusion of multi-wavelength imaging beyond the purely visual range. In order for this to be a reproducable and sensible concept, I introduced color profiles to allow for a more objective presentation. These profiles administrate the different color filters and their maps to RGB as applied by specific instruments or institutions. For now I manly work with the SDSS profile.

Regards,
Fridger


Top
 Profile  
 
 Post subject:
PostPosted: Sun, 30-09-12, 13:34 GMT 
Offline
User avatar

Joined: Fri, 29-02-08, 17:58 GMT
Posts: 85
I completely agree with Fridger's approach. My impression is that the effect we were commenting is an illusion, and I believe that it arises from the templates that define the arms, which Fridger will surely improve in time. The effect I refer to is most prominent in M100, of all the previous page images. Most of its arms are bluer and brighter on the inside and more orange and darker on the outside. Moreover, most of them are sharply defined in the inside border, and softer on the outside (or: this may be itself an illusion, consequence of the colors). All this gives the illusion of an illumination from the center.

See what I mean:

Image

G


Top
 Profile  
 
 Post subject:
PostPosted: Tue, 09-10-12, 20:46 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
Let me display a selection of globular clusters, the simulation of which in celestia.Sci is very sophisticated, indeed. One day, I'll write a dedicated "white paper" to describe in detail the high degree of realism that I was able to implement.

The globular data file now includes another column, the V-I magnitude that serves as a perfect indicator for the overall color of the globular clusters. These V-I data are much less affected by reddening, compared to e.g. to B-V.

As emphasized already, normal stars and globular cluster stars are both rendered by means of my shader stars, which allows to display extremely dim stars (like 25m) along with very bright ones... Two color schemes can be selected from the Preferences dialog (like in Celestia).

I have included below also rather dim and hence less well known globulars and will contrast them with a few some very familiar ones. If you look carefully, you will note the subtle colors ranging from blue over white to orange.

celestia.Sci includes ALL known galactic globular clusters.

Here we go:


++++++++++++++++++++++++++++
EDIT
Thanks to Christophe's attentive eye ;-) (cf. following post):

Since accidentally, in the upload process, most of the delicate colors were destroyed (lossless PNG-> lossy JPG) along with a forgotten Saturation reduction switch for small DSO's, I have redone the images correctly further below. They are now kept in PNG format.
++++++++++++++++++++++++++++

Fridger


Last edited by t00fri on Wed, 10-10-12, 11:04 GMT, edited 3 times in total.

Top
 Profile  
 
 Post subject:
PostPosted: Wed, 10-10-12, 9:16 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
Over the black background and by contrast, on my screen the bluish tone of the NGC 5466 and NGC 5053 is almost inexistent, visually for me it's a grayscale, gray (for dimmer) to white. Now, in full screen without the interference of the white background of CM's forum, perhaps the bluish is visually more consistent, unfortunately unless you post a full screen shot, I could not say...


Top
 Profile  
 
 Post subject:
PostPosted: Wed, 10-10-12, 10:52 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
ElChristou wrote:
Over the black background and by contrast, on my screen the bluish tone of the NGC 5466 and NGC 5053 is almost inexistent, visually for me it's a grayscale, gray (for dimmer) to white. Now, in full screen without the interference of the white background of CM's forum, perhaps the bluish is visually more consistent, unfortunately unless you post a full screen shot, I could not say...


Christophe,

many thanks for looking so carefully! Unfortunately, when I displayed the globular images yesterday evening (in bad light), there were TWO effects that significantly added up to destroying the subtle colors of these globulars.

1) The conversion from the lossless PNG screendumps to JPG (+85% quality!) killed most of the subtle blues as you observed.

2) As I described to you above , I have an automatic reduction of the Saturation built in for small DSOs. Since I had chosen mostly tiny and dim globulars, that switch was accidentally still active ;-) and turned most of the colors into shades of gray....

What a horror.

So I have redone the images above and this time kept the PNG format.
Have a look what the difference amounts to! ;-)

Here we go again:

yellowish M22:
=============
Image

dense M 2
=========
Image

and now some dim ones, with much flatter luminosity distributions
but often subject to significant reddening!

orange Palomar 05
==========
Image

orange Palomar 14
================
Image

faint, orange Whiting 1 (14.9m!)
=========================
Image

bluish NGC 5466
===============
Image

bluish NGC 5053
==============
Image

The bright, fairly red-orange stars are so-called red Giants that occur in practically all globulars. They are located near the end of the color-magnitude diagram...

Enjoy,
Fridger


Last edited by t00fri on Wed, 10-10-12, 12:31 GMT, edited 1 time in total.

Top
 Profile  
 
 Post subject:
PostPosted: Wed, 10-10-12, 11:09 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
Nothing to say now, beautiful work! :o


Top
 Profile  
 
 Post subject:
PostPosted: Wed, 10-10-12, 12:48 GMT 
Offline
Site Admin
User avatar

Joined: Thu, 30-08-07, 22:52 GMT
Posts: 2726
Location: France, South, not far from Montpellier
Stupid question not related to the soft: in the neighborhood of the clusters, how the astrophysicists determine if a star belong or not to a cluster and so if it should enter a global data set of lonely star or not?


Top
 Profile  
 
 Post subject:
PostPosted: Wed, 10-10-12, 14:04 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4442
Location: Hamburg, Germany
ElChristou wrote:
Stupid question not related to the soft: in the neighborhood of the clusters, how the astrophysicists determine if a star belong or not to a cluster and so if it should enter a global data set of lonely star or not?


This is all BUT a stupid question... In detail, the technology of separating field stars from globular cluster stars is quite complex. It rests on some pillars though: all stars of a globular have practically the same distance from us quite unlike the field stars (foreground and background). Next the essential tool is usually so-called wide-band-multicolor-photometry.

Just some of various general considerations in this context:

1) The star density far away from the cluster center (at r> r_tidal) corresponds to the field star density that should be about constant. If extrapolated to the cluster location- it contrasts strongly with the cluster star density that decreases strongly outwards from the center and is fitted with some standard density functions (e.g. I.King that I also use).

2) If one plots two of the various measured color magnitudes (e.g. U B V, I,...) against each other one expexts to see rather conspicuous color-magnitude relations to be realized. Popular choices are V vs. B - V or V vs. V - I etc.

These relations should be the SAME, whether the cluster star sample comes from near the cluster center or from various outer regions. For field stars the color-magnitude relations are quite different, obviously.

Hence stars can easily fall outside these pretty well known smooth constraint curves and are then attributed as field stars and subtracted.

Here is a typical example (M 69): left, the raw star measurements and on the right, the field-star-subtracted sample. The plots are typically V versus B-V. Read the figure captions.

Image

For people with deeper interests in this direction, here is a recent detailed scientific paper where the all these weapons are used:

http://iopscience.iop.org/1538-3881/138 ... 5_1455.pdf

Fridger


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 41 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group