New Celestia-1.4.0pre-FT1.2 Version for Download

[Wersadix]

Fridger, Toti great job on a new FT1.2!!!
And I was wondering can you fix one old bug in some future release of FT, that is in Celestia for some time. The bug is that stars that are farther then 16308.35 ly's from Sol. I am asking this because i want to add stars beyond this distance (primely globulars clusters) and i it is strange that some stars are to disappear. I use Rassilon Globular Cluster Generator.

[t00fri]

Fridger, Toti great job on a new FT1.2!!!
And I was wondering can you fix one old bug in some future release of FT, that is in Celestia for some time. The bug is that stars that are farther, then 16308.35 ly's from Sol. I am asking this because i want to add stars beyond this distance primely globulars clusters and i it is strang that some stars are to disappear. I use Rassilon Globular Cluster Generator.

See e.g here:

http://www.celestiaproject.net/forum/viewtopic.php?t=8385

it's not a "bug" it's related to numerical precision. Moreover, where do you want to get star distances from beyond 16k ly ??? The measured parallaxes are anywhay FAR too small.

Bye Fridger

[watersb]

Edit: I un-commented the line about ignoring OpenGL extensions in celestia.cfg, and I am able to run this version just fine on my Mac Mini.
It is FAST!

Code: Select all

  IgnoreGLExtensions [ "GL_ARB_vertex_program" ]


This build does not finish loading data on launch; it crashes before displaying the first window.

What can I do to fix this?

(/shaders.log exists but is empty)

Code: Select all

Date/Time:      2005-11-27 18:20:44.448 -0700
OS Version:     10.4.3 (Build 8F46)
Report Version: 3

Command: Celestia
Path:    /Users/bwaters/Desktop/Celestia 1.4preFT1.2ish (11-16-05)/Celestia.app/Contents/MacOS/Celestia
Parent:  WindowServer [65]

Version: Celestia v1.4preFT1.2 (11-16-05) (6)

PID:    1891
Thread: 0

Exception:  EXC_BAD_ACCESS (0x0001)
Codes:      KERN_PROTECTION_FAILURE (0x0002) at 0x00000018

Thread 0 Crashed:
0   net.shatters.Celestia       0x0000cf8c TextureFont::buildTexture() + 28 (crt.c:355)
1   net.shatters.Celestia       0x00011e88 CelestiaCore::initRenderer() + 708 (crt.c:355)
2   net.shatters.Celestia       0x0004e830 -[CelestiaController finishInitialization] + 40 (crt.c:355)
3   com.apple.Foundation        0x928e6138 __NSFireDelayedPerform + 304
4   com.apple.CoreFoundation    0x90770ae0 __CFRunLoopDoTimer + 184
5   com.apple.CoreFoundation    0x9075d458 __CFRunLoopRun + 1680
6   com.apple.CoreFoundation    0x9075ca0c CFRunLoopRunSpecific + 268
7   com.apple.HIToolbox         0x931831e0 RunCurrentEventLoopInMode + 264
8   com.apple.HIToolbox         0x93182874 ReceiveNextEventCommon + 380
9   com.apple.HIToolbox         0x931826e0 BlockUntilNextEventMatchingListInMode + 96
10  com.apple.AppKit            0x93660904 _DPSNextEvent + 384
11  com.apple.AppKit            0x936605c8 -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:] + 116
12  com.apple.AppKit            0x9365cb0c -[NSApplication run] + 472
13  com.apple.AppKit            0x9374d618 NSApplicationMain + 452
14  net.shatters.Celestia       0x000084e4 _start + 340 (crt.c:272)
15  net.shatters.Celestia       0x0000838c start + 60



[ElChristou]

Guys, some others templates in preparation (Sc, Sb), here at default illumination level... Still need some tweaking...

[Boux]

Great!
I want them

[Cham]

They look MUCH improved already. Nice work, ElChristou !

[ElChristou]

Great!
I want them

I do first the others then create an archive for download and test..

[t00fri]

What I particularly like with ElChristou's new templates is this:

--They are in a sense straight idealizations of the Hubble types, sort of perfect "textbook examples". Clearly, Nature is much more complex, much less regular etc...

+++++++++++++++++
But we should always remain consistent within the principles of the adopted approach: Foremost, the templates are supposed to remind us in nice graphical fashion of the /precise/ catalog data for many thousands of galaxies. That's also why I -- for my part-- will NOT try and produce a much more "fancy", i.e. higher resolution mass-rendering of galaxies.

In such a case, it would /increasingly/ become obvious that not the galaxy under consideration is rendered, but rather a /generic substitute/! In other words, in this mass-rendering approach, I strongly feel: "Less is More"!
+++++++++++++++++

Yet, I can see a considerable area for improvement relative to FT1.2 in form of an individual DSO coloration, by exploiting again the /catalog/ parameters, like the given B-V magnitudes etc. Note I am not talking about the coloration of a few DSO's, but of thousands of them! Certainly we may increase also the number of Hubble types to account for the extended Hubble classification and we might also allow for the possibility of individual templates for special galaxies like the Milky Way, as discussed already elsewhere.

--Despite the mentioned limitations, these "idealized" galaxies nevertheless look nice and esthetical...Because of their clean forms they also represent a high educational value!

If really /individual/, breathtaking highlight displays of some DSO's are needed/wanted, I'd rather advocate to further explore the recent amazing 3d add-on experiments being for display in another thread!

http://www.celestiaproject.net/forum/viewtopic ... sc&start=0

After flying myself right through this great new 3d nebula, here is a beautiful shot from my "little excursion"


(the 3rd dimension imaging is if course ficticious)

I guess NO mass rendering techniques will ever be able to compete with this sort of high resolution 3d views! Please imagine: in this image the actual resolution was reduced by a big factor ...

+++++++++++++++++++++
Yet the other side of the medal is that this sort of spectacular rendering is in practice only possible for a few selected objects. This clearly illustrates the complementarity of the two approaches.
++++++++++++++++++++



Bye Fridger

[ElChristou]

...--They are in a sense straight idealizations of the Hubble types, sort of perfect "textbook examples". Clearly, Nature is much more complex, much less regular etc...

Yes, for now my idea is to do some generic templates, waiting for a way to create custom ones for well known galaxies...
For now I "try" not to pass the 9000; in all case I have to do more test.
The actual templates are the result of 2 merged .pts file; unfortunatly, when merging those files, one always loose it's brightness level (why I don't know) so to do exactly what I want is not easy...

[t00fri]

...--They are in a sense straight idealizations of the Hubble types, sort of perfect "textbook examples". Clearly, Nature is much more complex, much less regular etc...

Yes, for now my idea is to do some generic templates, waiting for a way to create custom ones for well known galaxies...

In my view there is hardly a better way than doing a few custom galaxies with the above 3d mesh techniques as add-ons. That may be done any time NOW, as the above example shows.


Bye Fridger

[ElChristou]

...In my view there is hardly a better way than doing a few custom galaxies with the above 3d mesh techniques as add-ons. That may be done any time NOW, as the above example shows...

There is too much limitation for now with the 3D meshes and unfortunatly I don't think Chris will "note" the problem because it's a big work on the rendering engine ahead...

[t00fri]

...In my view there is hardly a better way than doing a few custom galaxies with the above 3d mesh techniques as add-ons. That may be done any time NOW, as the above example shows...

There is too much limitation for now with the 3D meshes and unfortunatly I don't think Chris will "note" the problem because it's a big work on the rendering engine ahead...

But really Jll's nebular is just SPECTACULAR, whatever the limitations were We never had anything comparable as far as I can remember.

Bye Fridger

[ElChristou]

But really Jll's nebular is just SPECTACULAR, whatever the limitations were We never had anything comparable as far as I can remember.

It's true, but actually those kind of addons are good in 1 or 2 directions, then you get some of those nasty problems we have already talk about... As we want the best for Celestia, well... you know what I mean ...
Later for new templates.
Bye

[Boux]

The actual templates are the result of 2 merged .pts file; unfortunatly, when merging those files, one always loose it's brightness level (why I don't know) so to do exactly what I want is not easy...

Yes, I have tried that and the result is not very convincing.
If you want, I can send to you my own modified version of Bmp2pts (source code or Linux executable) to play with.
It produces templates with a good balance between the gaussian profile at the center and the thickness in the outer parts.
Just let me know.

[ANDREA]

I guess NO mass rendering techniques will ever be able to compete with this sort of high resolution 3d views! Please imagine: in this image the actual resolution was reduced by a big factor ...Bye Fridger

Just to give another example, here is Jll's RCW_79 addon with furtherly modified texture, this time with no stars on the shell,
but a lot of "fictitious" internal stars:



The result is absolutely NOT scientific, but it is "beautiful", and this shoud not be forgotten.
IMO there are three different ways of thinking among Celestia's people:
the scientific, the fantastic and the educational one.
As I told many times, to educate people to astronomy, we need not only a technical and scientific approach (that is of primary
importance!) but at the same time an highly imaginative approach.
When on a2x3 meter wide screen I show this, and move around and in the nebula, even 80 years old people are interested,
and school boys say "wow", and the same happens when I show Millennium Falcon in close approach to Death Star.
Give a look at the number of posts existing on Sci-Fi textures, models, scripts. How many fantastic worlds!
Well, most of them follow the rules of physics, but someone doesn't.
But even if it's not technical or scientific, it's nice to look at, and even more for people whose lives are concerned with things
very far away from astronomy and science in general, like most of the world that is out of Celestia forum.
Jll's RCW_70 is scientific, mine is aesthetic only.
So, and this is my renewed proposal, why don't we split these different but coplanar points of view on the same thing, the scientific and
the aesthetic one?
I think they could live one close the other.
Bye

Andrea

[fsgregs]

EDITED!!!

HI folks:

I am going to try this one more time. I really like FT 1.2. It's galaxy views outside of the Milky Way are spectacular and some of its other internal features are wonderful. Fridger and Toti and everyone else has done a wonderful job. That said .... well .... Look for yourself!!












The above images were taken from Earth orbit with Galaxy brightness set at 0% (default). Click on the image for a high res version. Now, I'm sorry but this is not what the Milky Way really looks like from the vicinity of Earth ... is it? This is the default brightness that the Milky Way is showing on two separate CRT monitors that I use (one at home ... one at work). I have NOT adjusted the brightness of my monitors to be more bright than normal. They are right off the shelf, so to speak. FOV is set to 40, which is quite normal.

A few days ago, I got an email from another teacher who asked me some advice about using my Activities with FT 1.2 He said and I quote directly from his email:

2/ Galaxies are too bright. I remember seeing this in forums but thought
it had been resolved. I'll get a screenshot. It looks awful in my opinion.

Here is galaxy brightness in FT 1.2 set to 100%:



and



It seems to me that providing a brightness level this high is OK if you like things completely washed out, but how about in the other direction? Perhaps some of you have dimmer monitors. However, for those of us who simply think the Milky Way is far too bright at its lowest possible setting, all I am asking is for a way to set galaxy brightness at a lower default level. For example, just drop the zero setting down about 30%. Then, if people like it a bit brighter, they can simply press the [)] key a few times and Celestia will always display it brighter for them.

Surely there is more than two of us who think the default brightness level is too high!!! Now is the time to speak up ...

Regards,

Frank

[guest jo]

I can't see your pictures but generally I agree with you.

[ElChristou]

If you want, I can send to you my own modified version of Bmp2pts (source code or Linux executable) to play with.

Avec plaisir, je verrai si je peux la compiler car la version initiale avait encore quelques bugs sous osX...

...with furtherly modified texture...

Andrea, for test purpose I have done another "no star" texture of the Jll's one, without the effect of fludify a bit too visible to my taste on yours... Have a look. In the archive 2 png, one is very transparent (was for my test) but can be usefull...

To download: ARCHIVE (9mo)

********

Guys, here are some more shots; I think I'm in the limits of what can do the Toti's bmp2pts...



Here for know what I have done; some of them are two regular it's true, but the SBb and SBc began to be not so bad... (the Sa's arms are really to thin)



The big problem for now with those templates is that the brightness become really too high when playing with the FOV/distance... I hope I will find a solution...

[Cham]

ElChristou,

the barred spirals are AWESOME, on your pictures above. Looks perfect to me.

[Boux]

If you want, I can send to you my own modified version of Bmp2pts (source code or Linux executable) to play with.

Avec plaisir, je verrai si je peux la compiler car la version initiale avait encore quelques bugs sous osX...

C'est parti!