Joint Galaxy Template Design Effort

[t00fri]

I openend this new thread on demand from several people. Its purpose will be to communicate and test further galaxy templates.

Please, if you think you got some improvement, specify the URL to the respective .PNG template file. So we can all try it out.

For more experienced designers, PLEASE send e.g. psd (Photoshop, GIMP) or .xcf (GIMP native format) multilayer designs.

Here is ElChristou's scheme that has proven very effective:

- Layer1: background filled with black
- Layer2: grid for general mid tone sprites 25% gray
- Layer3: global shape with a gray 60%
- Layer4: enforcing the shape with a gray 70% (almost optional)
- Layer5: some touch of pure white for triggering bright sprites

Here is an example:

layer1 + cross layer(for centering): layer1 = black background

layer1 + layer2+cross layer(for centering): layer2 = 25% grid + black feathered circular mask

layer1+layer2+layer3+cross layer (for centering): layer3 = basic galaxy shape 60%, 70%

layer1+layer2+layer3+layer4+cross layer(for centering): layer4 = white accentuation


Except you have a real point with 256x256 or even 512x512 sizes, the "official" 128x128 size is much preferrable.

Good luck,
Bye Fridger

[Cham]

Is it posible to link some Photoshop file (with those layers) as a template for building a template ?

[t00fri]

Here is a generic layer setup in .psd format

1) black background
2) 25% grid (triggers diffuse background sprite activity)
3) feathered circular black mask (smoothly merges galaxy with black Universe)
4) red centering cross

http://www.celestiaproject.net/~t00fri/images/ ... layers.psd

Now one introduces further layers between 3) and 4):

1) transparent layer, drawing in 60% gray the base galaxy shape
2) transparent layer, drawing in 70% gray further accentuation
3) transparent layer, drawing in white for highlighting central arms etc.

Bye using these normalized intensities, the resulting galaxies are becoming automatically correct in brightness...

+++++++++++++++++
If you design things, never visualize what you draw, but rather what sprites your drawings will trigger off .
+++++++++++++++++
That's the NON-trivial aspect of this whole game

The point is that it's the sprites not that drawing that provide the galaxy rendering!

Good luck,
Bye Fridger

[t00fri]

Some further generic rules:

The Hubble morphology from Sa => Sc is characterized by

1) decreasing size of central spot
2) increasing opening of arms

In case of barred spirals, draw the bar at ~45 degrees inclination. From SBa => SBc, the same qualitative criteria 1), 2) apply.

What I usually do is to /vectorize/ the galaxy shape from a photo using vector graphics programs like 'inkscape'. Then you start working on that result which gives a far more natural impression.

Before definite conclusions about the quality of your new "creation" may be drawn, you have to start Celestia several times consecutively with your new template. The reason is that important parts of the rendering depend on randomized galaxy parameters in the code.

Bye Fridger

[Vincent]

OK, here's my first try for the MilkyWay template.
I only made a few changes from ElChristou's template :
- Arms a little bit thicker
- 25% Grid attenuated
- Little blur effect on all the layers

Here's one single shot :


And here's a zip file including the MilkyWay.png template and a .psd file with all the layers :
http://vincent.gian.club.fr/celestia/MilkyWay_V.zip

If the visual aspect of the MilkyWay template is satisfactory and if it goes through Fridger's accuracy test ( ), I may be able to give a try on some other templates...

[t00fri]

OK, here's my first try for the MilkyWay template.
I only made a few changes from ElChristou's template :
- Arms a little bit thicker
- 25% Grid attenuated
- Little blur effect on all the layers

Here's one single shot :


And here's a zip file including the MilkyWay.png template and a .psd file with all the layers :
http://vincent.gian.club.fr/celestia/MilkyWay_V.zip

If the visual aspect of the MilkyWay template is satisfactory and if it goes through Fridger's accuracy test ( ), I may be able to give a try on some other templates...

Vincent,

again, I have a little problem. You made some minor changes here and there. But you did NOT state what improves from an OBJECTIVE point of view.

Do you intend that we make a poll at the end, voting who get's most votes?

I think you should really try to demonstrate which features specifically your design improves. Personally, I don't find your image better than the one by ElChristou.

So what basis are we going to apply?

Bye Fridger

[Vincent]

Vincent,

again, I have a little problem. You made some minor changes here and there. But you did NOT state what improves from an OBJECTIVE point of view.

Do you intend that we make a poll at the end, voting who get's most votes?

I think you should really try to demonstrate which features specifically your design improves. Personally, I don't find your image better than the one by ElChristou.

So what basis are we going to apply?

Bye Fridger

Fridger,

To tell you the truth, I was asking myself the same questions, particularly about the Milky Way...
On which OBJECTIVE standards can we base our judgment, since we have no photo ? Here is my point of view :

- First of all, the shape of the template must provide a scientificaly accurate rendering : known arms, angle of the barcentral, position of the sun in Orion arm, etc... I think that my template is satisfactory as for that.
To help us, maybe you could tell on which image/document you based your template design.

- Then, the luminosity of the galaxy represents an important criterion to me. The luminosity of the galaxy must be perfectly balanced between the outside view and the view from the Earth. I know thet "perfectly balanced" is subjective, but I'm afraid we can't be more precise here...
However we might apply some rules : the inside view must have a realistic luminosity with AutoMag ON and an AutoMag level set to 7.0 at 45?°. The same base can be applied to the outside view.

- Finally, the last criterion is the main visual aspect. As for this, I'm afraid that no objective judgement can be given, since we're talking about the "artistic" aspect...
I'm not really for a poll. But I think there are some interested users here who can give their argued opinion...
Here's mine :
IMHO, ElChristou's template is a bit confused because there is too much material between the arms, relative to all the galaxy photos that are available in "true color". That's why I lowered a bit the luminosity of the grid. And even if this is not our galaxy, maybe we can base our "artistic" judgment on the photo of the M109 galaxy I posted in the other thread, since this is the most precise photo we have about a SBbc galaxy...

[Vincent]

As for the main "visual" aspect and if we accept to make a judgment from the M109 photo, maybe the solution would be to make a compromise between the 2 versions :



It would be also very interesting to get the filamentous aspect of the inter-arm material...

[Cham]

Actually, I prefer ElChristou's version, but this is just a matter of personal preference.

I would prefer to see a solution to the unatural dark straight line (as seen on the side), before doing any custom templates.

[t00fri]

Actually, I prefer ElChristou's version, but this is just a matter of personal preference.

I would prefer to see a solution to the unatural dark straight line (as seen on the side), before doing any custom templates.

Cham that's very easy to get rid of! But I think it's a good first approximation to the dark lanes we see from Earth and also in may sideways views of galaxies.

Remember that photo comparison I showed.
Look at the bottom image!



Bye Fridger

[Cham]

Fridger,

Your pictures are too small to see the line I'm talking about. Here's your template as seen from Earth. It's OBVIOUS this isn't right , there is no such line in the real sky:

[t00fri]

Cham,

on my system there is NO line whatsoever. Are you really talking about a line? Red? Please try to be more precise.

From when is this build. Did you do it?

Bye Fridger

[Cham]

Of course, I wasn't talking about the red line ! This red line is there to put into evidence the wrong line in the galaxy rendering.

Yes, this is seen from the latest CVS version (made this morning), and it was also visible in all the CVS builts I made since a week (or two?).

It's not a clear line. It's a bit fuzzy and dark, but can easily be noticed on the picture, above the red line. The Moon is even "sitting" on it.

[t00fri]

Of course, I wasn't talking about the red line ! This red line is there to put into evidence the wrong line in the galaxy rendering.

Yes, this is seen from the latest CVS version (made this morning), and it was also visible in all the CVS builts I made since a week (or two?).

It's not a clear line. It's a bit fuzzy and dark, but can easily be noticed on the picture, above the red line. The Moon is even "sitting" on it.

You should be able to amplify that mysterious line of yours in a clearcut image (cel://url for example!). I DO NOT see a line, except what I told you: the emulation of the dark lane in the galactic plane.

If you are not able to show me this line in reality, I cannot do anything about it, obviously.

Bye Fridger

[Cham]

Then I guess your CRT monitor is very dark, or not enough contrasted ? It's so visible on the picture above. I tried to make it more visible, on this same picture, edited in Photoshop (higher contrast and brightness) :



Just follow the green lines ...

Here's an URL to see it in the middle of the screen,. Horizontal dark fuzzy line.

cel://Follow/Earth/2002-05-24T11:36:00. ... 38295&lm=2

[t00fri]

Sorry, Cham,
I definitely do not see that strange line!

I have most carefully checked all the way along the MW, using your cel://url configuration. NOTHING whatsoever.

To prove this to you: here is your cel://url center part, with the brightness key turned up i.e. ")".



That line that you have on your image, I would have definitely seen on my monitor long ago.

So we must wait if other people (with experience in compiling and NON MACs!) see the line as well.

Bye Fridger

[Cham]

Fridger,

could you please show the entire image from the URL I gave, and without changing the luminosity ?

And do you agree there's a line on MY pictures ?

[t00fri]

Of course there is a line on your image.

Here I turned the light up: it's the same region than my image above:



Since I am using 1600x1200, my whole image is too large.

Bye Fridger

[Cham]

While I agree that the line doesn't seems to be visible on YOUR picture, here it is again, on mine, pushing the galaxy brigthness to the max. Follow the red lines :



EDIT : When moving around in Celestia, using a normal brightness, it's so clear, and so annoying, so unatural.

[t00fri]

Note that you are NOT allowed to use a FoV that is too big, since it may generate such artefacts.

Please always show images with FoV 30-40 degrees!

What FoV did you take on your previous image?
With FoV ~ 120 I can show you artefacts all over Celestia

Bye Fridger