New Celestia-1.4.0pre-FT1.1 Version for Download

[Cham]

Fridger, can you make your templates available with some download link, so we can test them ?

[t00fri]

Fridger, can you make your templates available with some download link, so we can test them ?

OK, here they are, but they are definitely experimental. A number of little issues may still be improved.

SBb designer template
SBc designer template

Bye Fridger

[Cham]

Thanks, Fridger.

I think those models are much too simple (SBb and SBc). The template for SBb is more appropriate for some SBc representation, not SBb. The template for SBc is much too simplistic in shape. I think Boux's last representation is closer to what we need.

Also, I feel that your template are too thin on the sides. The models are too flat.

[symaski62]

Here is another approach to template making:

entirely "synthetic" designer templates, one might call them

The SBb .bmp and SBc.bmp inputs were generated with GIMP from this simple 128x128 starting "circle"

http://www.celestiaproject.net/~t00fri/images/circle.jpg

by using suitable GIMP tools. E.g. for SBb.bmp , the normalized input result then looks like this

http://www.celestiaproject.net/~t00fri/images/test4.jpg

It's hard to believe that this thing gives something usable....

Here is the corresponding SBb and SBc output of bmp2pts
with <threshold> = 0.75 and <randomness> =0.025:
http://www.celestiaproject.net/~t00fri/images/SB-designer.jpg

Bye Fridger

BMP charge file.pts

[t00fri]

Thanks, Fridger.

I think those models are much too simple (SBb and SBc). The template for SBb is more appropriate for some SBc representation, not SBb. The template for SBc is much too simplistic in shape. I think Boux's last representation is closer to what we need.

Also, I feel that your template are too thin on the sides. The models are too flat.

The templates may be too simple...

Don't forget: we need to implement another 10000 clusters and nebulae, and another 10000 asteroids and comets. We are fighting mostly with performance considerations!!

BUT please believe me that I know how to do a SBb and a SBc generic template. You are simply incorrect.


The "artistic view" galaxy that Boux was converting simply does not exist in such a form.

Here is the classical SBb shape of M91, for example.



Just to remind you about SBc here is the diagram (NASA)

[Cham]

Fridger,

please don't underestimate me. I know what is the Hubble classification, what do you think ? I even teach it in the classroom. I suspect you actually don't understand what I'm trying to tell. The templates used for the Sa, Sb and Sc for example are just perfect the way they are and IMO don't need any correction. So about the SB models, it's not a problem of resolution. It's a problem of design for the barred spirals ONLY.

Like BlindedByTheLight said :

"Sometimes a direct sampling of what is REALLY there... doesn't look like what is really there... without some additional tweaking".

If you try to get very close to reality, you may lose it. The SBb and SBc models needs more imagination, than comparison with reality. It's NOT a problem of resolution at alll. It's simply a problem of shape design. I know it's difficult to do, this isn't my point.

I think Boux designs are better, even if there are some unrealistic elements in them (which may be corrected).

[ogg]

Anyone willing to try the following? It's a simple resize of Powell's 'best guess' for the milky way. It'd probably either come out like the NASA one or just be a more-or-less blank disc, I know. But I'd be interested to see if there is a difference.

[/img]

[t00fri]

Cham,

Fridger,
please don't underestimate me. I know what is the Hubble classification, what do you think ? I even teach it in the classroom. I suspect you actually don't understand what I'm trying to tell. The templates used for the Sa, Sb and Sc for example are just perfect the way they are and IMO don't need any correction. So about the SB models, it's not a problem of resolution. It's a problem of design for the barred spirals ONLY.

I underestimate nobody. Your own statements speak for themselves ...You were plainly wrong in saying that my above SBb shape was rather SBc and that my SBc was something else. Boux's template is certainly NOT SBc as advocated. I am always ready and happy to accept qualified critics...

The template for SBb is more appropriate for some SBc representation, not SBb. The template for SBc is much too simplistic in shape.

I just hope you don't make such statements in the class room...

+++++++++++++++
This issue is NOT a matter of realism, beauty and your or my personal taste.. The issue of taste is NOT at discussion here.
+++++++++++++++

SBc and SBb types have a set of characteristic features that are uniquely defined. My above examples are clearly too simple to be realistic, but they are /generically/ representing the corresponding advocated Hubble class features SBb and SBc. Nothing much else in fact

I hope you teach your pupils in the class room that there are three major criteria for barred spiral classification SBa..SBc:

-- the pitch angle of the arms relative to the bar
~90 degrees for SBa
> 90 degrees for SBb (see my M91 example below)
~180 deg: arms emanate initially in the direction of the bar

The increasing pitch angle implies an increasing opening of the principal spiral arms from SBa..SBc

-- the 'resolution' of the arms
increasing from SBa to SBc

-- the conspicuousness of the bar
decreasing from SBa..SBc

Here is again a direct confrontation of the structure of the M91 Hubble photo, a reference SBb spiral! The leftmost template you denied being of SBb type:



Clearcut SBb signatures: pitch angle (of one principal arm) > 90 degrees, well resolved, strongly wound spiral arms, conspicuous bar.

Note: the lower principal arm of M 91 having /apparently/ a pitch angle < 90 degrees is an effect of perspective due to the inclination of the galaxy! While my template (left) has correctly /both/ pitch angles > 90 degrees, also in the Celestia display (accounting for the inclination, only the "upper" arm appears with pitch angle >90 degrees in the view from Earth!

...Last NOT least:

In the many months of working on this galaxy project I have inspected hundreds of DSS photographs along with their official Hubble type designation including comparison with my galaxy templates in Celestia. So apart from my general professional background in these matters, I happen to be in /excellent/ practical training

Bye Fridger

[Boux]

OK, new try.
Here is M100 again, thanks to tell me if it is bright enough now for your displays.
Click the link below to see the screenshot.
http://jmmi.club.fr/celestia/m100_4.png

The template is there:
http://jmmi.club.fr/celestia/SBc.pts

Parameters:
- threshold 0.5
- randomness 0.02
- 9780 points generated

[t00fri]

Hi Boux,

could you perhaps reduce your above template in size or even better cut just the galaxy out of this huge display, please. This makes reading other posts VERY difficult. Try to limit the width of anything inserted smaller than 600-800 pixels, say.

Thanks!

Bye Fridger

[Boux]

Hello t00fri, done!
I am now looking at the source code of Bmp2pts.
Many parameters can be adjusted right there in the internal data and formulas, like blob's brightness, effect of randomness, etc..
I am playing around also with threshold, and other stuff.
As compilation takes only a fraction of a second, it's a no-brainer as compared to putting together the right input picture.
People who have not tried yet just can't understand

[Boux]

Anyone willing to try the following? It's a simple resize of Powell's 'best guess' for the milky way. It'd probably either come out like the NASA one or just be a more-or-less blank disc, I know. But I'd be interested to see if there is a difference.

[/img]

Sorry, ogg, but the picture you provided is unusable without complete rework.

[Cham]

Boux,

your new template is pretty, but it feels yet too unrealistic. The arms are too much circular, too well defined, and they ends too abruptly like if they were cuted. They are rolling up too much around the bar, and it's not obvious it supposed to be a barred galaxy.

I personally think that Fridger's template defined for the SBb representation is very good. What needs to be reworked are the templates for SBa and SBc.

[Cham]

Fridger,

here's my crittic of your latest template defined for SBc. Is it possible to make the arms ends a bit thinner ?



Also, the parts between the arms and the bar (the "shoulder") needs some work. I think they are too well defined, not enough "fibrous" or "stringy".

[Cham]

I tried the "old" SBc template from Celestia 1.4 with Celestia FT 1.1. I think the result isn't bad for the arms :



However, I'm not sure about the calibration and bar orientation. Only Fridger can tell if they are ok.

Is it possible to mix this template with the previous one from Fridger ? I mean take the arms from the old version and add the nucleus from the new version ?

Hmmm, I'm not convinced myself. I'm not sure about this idea...


I'm very tired right now and I think it's time to go to bed...

[Boux]

Hi all!
I have made some changes in both Celestia source code and bmp2pts and re-compiled the whole stuff (Linux).
I have added a new galaxy type "MW" for the Milky Way so that only the MW shows up with my new template (and not all SBcs as before).
In bmp2pts, I have tweaked the brightness calculation formula to pump up it a bit.
Here is a screenshot of the new Milky Way:



Those interested can grab the MW.pts template here (rename it SBc.pts if you don't change the source code):
MW.pts

[Wersadix]

Nicely done, Boux! How much did you take to make a this new template?
I hope not too long

[Boux]

Thanks for liking it
It took some time (a few hours) to figure out how bmp2pts analyses the input picture, what is really critical in the picture, what kind of algorithms it uses and then to understand what the Celestia engine is making out of the template file.
Interesting. I am on a crash C++ course now

[neo albireo]

Hi all!
I have made some changes in both Celestia source code and bmp2pts and re-compiled the whole stuff (Linux).
I have added a new galaxy type "MW" for the Milky Way so that only the MW shows up with my new template (and not all SBcs as before).
In bmp2pts, I have tweaked the brightness calculation formula to pump up it a bit.

Hi Boux!
I like your template very much when seen from above. When seen sideways though, especially from inside, e.g. from our solar system, it seems too sharp a line in my opinion.
Keep on trying out!

[Boux]

Hi all!
I have made some changes in both Celestia source code and bmp2pts and re-compiled the whole stuff (Linux).
I have added a new galaxy type "MW" for the Milky Way so that only the MW shows up with my new template (and not all SBcs as before).
In bmp2pts, I have tweaked the brightness calculation formula to pump up it a bit.

Hi Boux!
I like your template very much when seen from above. When seen sideways though, especially from inside, e.g. from our solar system, it seems too sharp a line in my opinion.
Keep on trying out!

Try this one:
http://jmmi.club.fr/celestia/milky_120.pts
Rename it either SBc.pts or whatever if you recompile Celestia after having added a special galaxy type for the Milky Way.
A lot of work on the overall thickness and on the central area, with a neat sideview.
Changed the code in Bmp2pts to get there.
Pretty realistic on that particular point, I believe.
Thanks to tell me whether it is too bright/dim.
It looks OK on my system.
Enjoy