New Celestia-1.4.0pre-FT1 (galaxies) for Testing

[BlindedByTheLight]

Well, that didn't take too long to crash and burn... anyone having any luck compiling this puppy on OS X? On 10.4.2, using Xcode 2.1, I got these 4 errors almost immediately...

[fsgregs]

Fridger:

you might take our FT1 code "to warm up" until our FT1.1 update will be out that contains a number of fixes and further enhancements.

Cheers,
Bye Fridger

Any idea when FT1.1 will be ready? Your last post suggested it would be at the end of this week just past.

Obviously we are ANXIOUSLY awaiting its release.


Frank

[buggs_moran]

Folks,

Speaking of possible extensions of the galaxy rendering, let me share that I have been thinking of adding a pseudo realistic dust disk to the current F&T disk+bulk rendering of spirals. Perhaps it could give a realistic touch, especially when they are seen sideways. I have been reading the code, but again, with little C++ knowledge and less time, it has been uphill... Perhaps someone else...

Cheers,

Guillermo

Was just thinking on this and came up with a possibility (possibly ). The idea would utilize Fridger and Toti's galaxy code and the .pts files. Would it be possible to introduce another smaller or flatter "galaxy" superimposed over/within the originals but with the luminosity off, thus making the .pts files act as opaque dust clumps within the galactic plane?

[abramson]

I am not sure whether I understood precisely where you want to place the dust disk in case of spirals. Do you have some way of illustration with a graphics program, perhaps? (GIMP)

My idea was to obtain something like this:

instead of this:


I acknowledge: (1) that the current rendering is fantastic; (2) that each galaxy is different in dustiness; (3) that dust seems to be particularly visible when exactly edge-on; (4) that there are notable exceptions to (3), for example the Sombrero galaxy (and the beautiful Centaurus A -not even a spiral-, which we have in the Southern hemisphere ). Still, I believe that a little -random, realistic- obscuring dust clouds or filaments would be a nice detail.

What do you folks think? The idea of buggs_moran seems reasonable, at least to see if the addition is worthly reflected in the appearance of the galaxies.

Guillermo

[ElChristou]

...What do you folks think? The idea of buggs_moran seems reasonable, at least to see if the addition is worthly reflected in the appearance of the galaxies...

Agree, after this first work on massive galaxies rendering, it would be cool to have the possiblility of "customizing" some of the best known... the point of dark clouds is not new, but apparently it's not so easy to make them work correctly...

[maxim]

I acknowledge: (1) that the current rendering is fantastic; (2) that each galaxy is different in dustiness; (3) that dust seems to be particularly visible when exactly edge-on; (4) that there are notable exceptions to (3), for example the Sombrero galaxy (and the beautiful Centaurus A -not even a spiral-, which we have in the Southern hemisphere ). Still, I believe that a little -random, realistic- obscuring dust clouds or filaments would be a nice detail.

That's what I suggested some posts above. There are however some serious constraints. How would you draw if someone rotates the galaxy with your dust ring superimposed? You'll soon get serious trouble here.

maxim

[abramson]

How would you draw if someone rotates the galaxy with your dust ring superimposed? You'll soon get serious trouble here.

As far as I understand bugs_moran's suggestion, the dust clouds would be a disk of "clumps" like the bright one, same spiral type, etc., only that it would be opaque instead of bright, and flatter, "inside" the bright one (or perhaps sparser instead of flatter?). The dark clumps would obscure part of the bright disk from some prespective more than from others. Typically, more from a side than from the top. From the top (or, symmetrically, from the bottom), one would mainly see the "upper" part of the bright disk, unobscured. Sideways, the dark clumps would obscure more and more bright ones, perhaps giving an appearance similar to the one produced by real dust clouds in real galaxies. Ah, if I only knew how to make this with the current code!

Guillermo

[Toti]

This is a very low-quality example implementing "bin-sorting" of particles. See the discussion with maxim on this thread.
Note the dark dust ring in the edge-on view (the galaxy model template is not very good but it's enough for a quick test)
There are numerical stability problems still to be resolved.



Bye

[PlutonianEmpire]

Wow! that is sweet!

[maxim]

As far as I understand bugs_moran's suggestion, the dust clouds would be a disk of "clumps" like the bright one, same spiral type, etc., only that it would be opaque instead of bright, and flatter, "inside" the bright one (or perhaps sparser instead of flatter?). The dark clumps would obscure part of the bright disk from some prespective more than from others. Typically, more from a side than from the top. From the top (or, symmetrically, from the bottom), one would mainly see the "upper" part of the bright disk, unobscured. Sideways, the dark clumps would obscure more and more bright ones, perhaps giving an appearance similar to the one produced by real dust clouds in real galaxies. Ah, if I only knew how to make this with the current code!

Well the problem is, that things aren't sorted. The effect of that is, that it will perhaps look ok on a still pic. But if you start rotating, the clumps will rotate /through/ each other or even start flickering heavily back and forth, because the sequence order in which the single clumps are drawn may change from one moment to the next in an unpredictable way.

maxim.

[buggs_moran]

How would you draw if someone rotates the galaxy with your dust ring superimposed? You'll soon get serious trouble here.

As far as I understand bugs_moran's suggestion, the dust clouds would be a disk of "clumps" like the bright one, same spiral type, etc., only that it would be opaque instead of bright, and flatter, "inside" the bright one (or perhaps sparser instead of flatter?). The dark clumps would obscure part of the bright disk from some prespective more than from others. Typically, more from a side than from the top. From the top (or, symmetrically, from the bottom), one would mainly see the "upper" part of the bright disk, unobscured. Sideways, the dark clumps would obscure more and more bright ones, perhaps giving an appearance similar to the one produced by real dust clouds in real galaxies. Ah, if I only knew how to make this with the current code!

Guillermo

That is exactly what I meant. You would actually have a separate set of pts files governing the dust structure, maybe random, ring (think Sombrero), etc... I understand Maxims sorting questions, as I experience them in my cmod designs. However, I overlayed a couple of different galaxies through the milky way and got no flickering, it just increases the brightness on overlapped parts. That is not to say flickering wouldn't happen with opaque clumps vs. the existing luminous transparent ones.... This is an example of an added core to the Milky Way in which overlying points in the pts file showed more luminous. I do not think the galaxies work in the same fashion as overlapping png/jpg files in cmods. However, I do not claim to know anything about OpenGL layering rules...

WITHOUT


WITH

[fsgregs]

I have been enjoying the view of galaxies in 1.4.0 FT 1, but I've been quite surprised to see the shape of the Milky Way as a prominent barred spiral with only two arms. In particular, I just read an article in Scientific American which pointed out that the latest conclusion on the Milky Way is that it is an SBbc spiral with at least four spiral arms and a weak bar. NASA agrees and has released a new sketch showing it to be a weak bar.

Here is a website that shows the sketch.
http://en.wikipedia.org/wiki/Milky_Way

I checked and have found at least 10 other sites all saying the same thing.

Just curious! Is there new data showing it to not be an SBbc?

Frank

[t00fri]

I have been enjoying the view of galaxies in 1.4.0 FT 1, but I've been quite surprised to see the shape of the Milky Way as a prominent barred spiral with only two arms. In particular, I just read an article in Scientific American which pointed out that the latest conclusion on the Milky Way is that it is an SBbc spiral with at least four spiral arms and a weak bar. NASA agrees and has released a new sketch showing it to be a weak bar.

Here is a website that shows the sketch.
http://en.wikipedia.org/wiki/Milky_Way

I checked and have found at least 10 other sites all saying the same thing.

Just curious! Is there new data showing it to not be an SBbc?

Frank

Frank,

the explanation is quite near at hand:

Yes you are right, the best estimate is an SBbc morphology with more than 2
arms. The present galaxy rendering algorithm has the strongly idealized task
to render 10000+ galaxies in terms of only ten Hubble type templates!
The type SBbc is a subtype between SBb and SBc. Yet our present, standard
Hubble galaxy classification (still) ignores subtypes for various reasons. The
present reduction algorithm always converts to the next "higher" Hubble type
in that case, i.e. SBbc -> SBc. The corresponding template has only 2 arms.

If you look closely, you will find many more such "idealizations" among our
10000+ galaxies. But this is a compromise that will always remain in one form
or another.

In the next major release, we shall incoporate the possibility of attributing
a custom template file to each galaxy. The Milky Way would then be a typical
application for that new feature. Another extension could be (as I posted
earlier somewhere) to generally enlarge the number of Hubble class templates
to include subclasses. Catalogs use a much finer morphological subdivision
meanwhile. But since each templatate "costs" 250-500 KB of storage one can
easily double the size of the Celestia distribution this way...

I hope this solved the present "discrepancy" .

Bye Fridger

[Malenfant]

The artist's impression of the Milky Way shown on the wiki site (and in the SciAm article) looks really unusual to me, I can't think of another galaxy off the top of my head that I've seen a picture of that looks anything like that.

Do we have any images of other galaxies that look similar to what we believe the Milky Way looks like? Or is our galaxy somehow unique in all the universe (which I really can't believe)?

[t00fri]

The artist's impression of the Milky Way shown on
the wiki site (and in the SciAm article) looks really unusual to me, I
can't think of another galaxy off the top of my head that I've seen a
picture of that looks anything like that.

Do we have any images of other galaxies that look similar to what we
believe the Milky Way looks like? Or is our galaxy somehow unique in
all the universe (which I really can't believe)?

Right, that "artist impression" image indeed does NOT look very realistic. But
the subclass type SBbc makes a lot of sense (remember : S=spiral,
B=barred) : the c type typically has 2 very pronounced arms while b
as more somewhat less pronounced arms. So 4 arms could be just
right.

Here is a 2-armed SBbc galaxy(NGC 2442) that looks pretty much like our MilkyWay template:



Here is another SBbc type (NGC 7479) that looks quite different from
the first one as to the arms. The strong bar is identical, though.



Bye Fridger

[ogg]

I've always liked the plots/impressions that Richard Powell does, as part of his "An Atlas of the Universe" pages.

See: http://www.anzwers.org/free/universe/milkyway.html

This 'artist impression' is more a schematic than anything, and probably more useful as a guide for 'template' developers than other more artistic efforts. I also like how he lists his sources. Seems fairly scientifically respectable.

[Malenfant]

Interesting, thanks Fridger. Have we not seen a four-armed barred spiral then? So our Milky Way would probably look like the ones you showed, but with a couple of extra arms?

EDIT: Oh wait, there's some in the link that ogg posted.


NGC 3953 (above) looks most similar to the artist impression I think.

There's a good article about galactic structure in the latest Scientific American too, BTW. It presents an interesting explanation for how bars are formed.

[buggs_moran]

In the next major release, we shall incoporate the possibility of attributing
a custom template file to each galaxy. The Milky Way would then be a typical
application for that new feature. Another extension could be (as I posted
earlier somewhere) to generally enlarge the number of Hubble class templates to include subclasses.

Bye Fridger

Fridger,

I was wondering if you could tell me how you are creating the pts template files? Sorry if this was addressed in an earlier post...

[t00fri]

In the next major release, we shall incoporate the possibility of attributing
a custom template file to each galaxy. The Milky Way would then be a typical
application for that new feature. Another extension could be (as I posted
earlier somewhere) to generally enlarge the number of Hubble class templates to include subclasses.

Bye Fridger

Fridger,

I was wondering if you could tell me how you are creating the pts
template files? Sorry if this was addressed in an earlier
post...

Well, I used some "vectorizing code" written by Toti , (after
eliminating some nasty buggies). This simple command-line tool
has it's "own will" somehow, but after practicing enough, the
results are quite neat.

The point is to start from some representative galaxy photograph,
edit it considerably, make it grayscale and 128x128 in size. Then
save it as a .bmp format. This then acts as input for Toti's
command-line tool 'bmp2pts'. There are 2 crucial parameters to
play with :threshold and randomness.


Soon or later I am sure we shall release that tool for people to do
their own .pts experiments...


Bye Fridger

[Malenfant]

Just FYI, there's a bazillion images of galaxies here:
http://cosmo.nyu.edu/hogg/rc3/
http://www.astro.princeton.edu/~rhl/PrettyPictures/NGC/

NGC 151 looks like another contender for a galaxy similar to our own: