Celestia Forums

4D2U Project
http://4d2u.nao.ac.jp/index_E.html

Mitaka : A Four-Dimensional Universe Viewer
http://4d2u.nao.ac.jp/html/program/mitaka/index_E.html

Latest Version is here
http://www.systranbox.com/systran/box?s ... am/mitaka/

Its our galaxy & cosmic structures are so beautiful.

I can't try it since I'm on a Mac, but its graphics appears to be far superior than Celestia.

That's what I was afraid of since quite a while! Competition is
closing up fast, since Celestia has not embarked in
conceptionally new visualization tasks ... . I have
advocated important ones over years.

Incidentally, while for Celestia the "ears were deaf",
Mitaka from the start also seems to have a capable
"Cosmo mode" with nice displays of the large scale
structure data! Since that's where the trends go anyway
, I expect much more to come in this sector. They have
seemlessly managed to unify "near and far" displays!

+++++++++++++++++++++++
The "ears were so deaf" around here that I had even
decided to fork off (hi Chris;-) ) as soon as my time
would allow, in order to initiate a new "Cosmo"
visualization project myself, with proper 4d general
relativity framework, conformal kinematics and
light-frequency filtering.

Also this seems now superfluous...
+++++++++++++++++++++++

Celestia's entire large-scale "space-time" kinematics is
just old-fashioned...

So let's see what the Mitaka program really can do.


Bye Fridger

Looks very impressive.

... Fridger, I think I won't work anymore on galaxies!

Wow,...it does look good! Just installed it on my core-2-duo notebook. Recognizes and exploits my 2 CPU's and makes use of SSE enhancements. Very well designed framework.
Seems stable and has virtually everything that Celestia has, AND much more...


Bye Fridger

t00fri wrote:That's what I was afraid of since quite a while! Competition is
closing up fast, since Celestia has not embarked in
conceptionally new visualization tasks ... . I have
advocated important ones over years.

Mitaka does some things better than Celestia, others not as well. The galaxy rendering is beautiful, with very nice spiral arms and dust lanes. It appears to be accomplished with a ray casting algorithm. As for things it doesn't do: extrasolar planets, binary stars, extensibility through add-ons, SPICE interface, fast rendering of complex models, ring shadows. This isn't intended as criticism of Mitaka; I'm just pointing out the some of the different focuses of the two programs.

Incidentally, while for Celestia the "ears were deaf",
Mitaka from the start also seems to have a capable
"Cosmo mode" with nice displays of the large scale
structure data! Since that's where the trends go anyway
, I expect much more to come in this sector. They have
seemlessly managed to unify "near and far" displays!

Here, there's no real magic in Mitaka . . . There are a few minor modifications one could make to Celestia to accomplish the same thing. The large scale structure data could be rendered as a point sprite cmod in Celestia. A nice thing that Mitaka does is to fade the data set in at the appropriate distance: it isn't visible when the camera is positioned within the Milky Way. The fade effect is a simple and effective tool used through Mitaka, and could be added to Celestia as well.

+++++++++++++++++++++++
The "ears were so deaf" around here that I had even
decided to fork off (hi Chris;-) ) as soon as my time
would allow, in order to initiate a new "Cosmo"
visualization project myself, with proper 4d general
relativity framework, conformal kinematics and
light-frequency filtering.

Also this seems now superfluous...
+++++++++++++++++++++++

Hardly the case . . . Mitaka doesn't do any of these things. Anyhow, someone should get busy producing a cmod of the 'Large Scale Structure of the Universe' data set.

--Chris

t00fri wrote:Wow,...it does look good! Just installed it on my core-2-duo notebook. Recognizes and exploits my 2 CPU's and makes use of SSE enhancements. Very well designed framework.
Seems stable and has virtually everything that Celestia has, AND much more...


Bye Fridger

Could you post a few more shots of galaxies?

chris wrote:...A nice thing that Mitaka does is to fade the data set in at the appropriate distance: it isn't visible when the camera is positioned within the Milky Way. The fade effect is a simple and effective tool used through Mitaka, and could be added to Celestia as well...

If it's simple And effective, is there any reason to wait more for such solution for Celestia?

ElChristou wrote:8O... Fridger, I think I won't work anymore on galaxies!

That's the least...It's the scope of this program that is much more professional and up to what is needed in the future.

The galaxies are certainly nice. They use also published catalog data like I do. Their sources are not as sophisticated/accurate as mine though. But that can easily be improved. The graphics for galaxies is way superior. But that is generally true, everywhere. The code is based on DirextX > 8.1 (not OpenGL). There are also sophisticated
atmosphere options (multiple scattering etc)

Bye Fridger

chris wrote:...Anyhow, someone should get busy producing a cmod of the 'Large Scale Structure of the Universe' data set...

Is this the best solution? I mean it would be a static model, no way to add the data to come... (or by adding new models with the new data?)

And with such a model, the trick would be to fade between this model and what we have actually at close range? Am I right?

It would be nice also if someone post a few shots of the UI to see what is their approach (seems to be all by contextual menu?)...

JetCrow, Many Tx for the links!

Are you part of the Mitaka project?

chris wrote:...
Mitaka does some things better than Celestia, others not
as well. The galaxy rendering is beautiful, with very nice
spiral arms and dust lanes. It appears to be accomplished
with a ray casting algorithm. As for things it doesn't do:
extrasolar planets, binary stars, extensibility through
add-ons, SPICE interface, fast rendering of complex
models, ring shadows. This isn't intended as criticism of
Mitaka; I'm just pointing out the some of the different
focuses of the two programs.

Mitaka is still very young compared to Celestia.


I had no chance looking into the sources (MIT licence)
yet and the manual is in Japanese. So it's hard to asses
implementation aspects in depth at this point. But there
are many signs that I could make out that clearly point
towards a VERY well done general framework.


However, I noticed that their detailed galaxy rendering
also has its price. On my core-2-duo the framerates get
quite slow near galaxies (DirectX 9.0c) . Also wait until
they got 10000 at a time! Right now there is only
essentially some rendering of the local group.


There is however already rendering of the Oort cloud and
trans-neptunian Objects in general.

Here, there's no real magic in Mitaka . . . There are a few
minor modifications one could make to Celestia to
accomplish the same thing. The large scale structure data
could be rendered as a point sprite cmod in Celestia.

I think "minor modifications" would be a loosing strategy
in this respect...

Hardly the case . . . Mitaka doesn't do any of these
things. Anyhow, someone should get busy producing a
cmod of the 'Large Scale Structure of the Universe' data
set.

Would again be a short-sighted loosing effort.

Bye Fridger

Here are a few shots of the general graphics.

Saturn with smooth orbit displays and nebulosity visible in Orion's belt and around the trapezium



Earth topography from low altitude



The topographical aspects are also realized different/ better than in Celestia. One may e.g. switch between different levels of topographic exageration during the display.

Bye Fridger

chris wrote:

+++++++++++++++++++++++
The "ears were so deaf" around here that I had even
decided to fork off (hi Chris;-) ) as soon as my time
would allow, in order to initiate a new "Cosmo"
visualization project myself, with proper 4d general
relativity framework, conformal kinematics and
light-frequency filtering.

Also this seems now superfluous...
+++++++++++++++++++++++

Hardly the case . . . Mitaka doesn't do any of these things.
...
--Chris

I wouldn't be so sure about this


+++++++++++++++++++++++
First of all, one may switch between a display of the milky
way in visual light and in infra-red light! Is
this wave-length filtering or not??
+++++++++++++++++++++++

Here is a display that illustrates first steps towards the conformal mapping display:



Bye Fridger

t00fri wrote:

chris wrote:

+++++++++++++++++++++++
The "ears were so deaf" around here that I had even
decided to fork off (hi Chris;-) ) as soon as my time
would allow, in order to initiate a new "Cosmo"
visualization project myself, with proper 4d general
relativity framework, conformal kinematics and
light-frequency filtering.

Also this seems now superfluous...
+++++++++++++++++++++++

Hardly the case . . . Mitaka doesn't do any of these things.
...
--Chris

I wouldn't be so sure about this


+++++++++++++++++++++++
First of all, one may switch between a display of the milky
way in visual light and in infra-red light! Is
this wave-length filtering or not??
+++++++++++++++++++++++

I thought that you meant something much more ambitious when you talked of wavelength filtering. Mitaka allows you to toggle different galaxy models on and off, but that's all. This is an incredibly useful feature, which I want to implement via a scheme like Cham's custom classes. But, you could just as correctly call Celestia's alternate surfaces wavelength filtering if you had, say, two textures of Venus from images taken with different filters.

Here is a display that illustrates first steps towards the conformal mapping display:

This is just scaling of planet radii. Again, a useful feature, but far from conformal mapping.

--Chris

ElChristou wrote:8O... Fridger, I think I won't work anymore on galaxies!

Christophe,

after some more exploration, the galaxy display is still hard to compare with what we do in Celestia. So far only the MilkyWay is rendered in 3d the rest are just suspended 2d images like in case of many of our add-ons. I don't think Andromeda is superior to our display. Except, I didn't find yet the right button . Certainly Celestia is WAY faster and has MANY more galaxies correctly oriented. The 3d rendering of the MilkyWay galaxy is REALLY slow in Mitaka.

There are some nice displays of elliptical galaxies though (virgo cluster) .

Bye Fridger

chris wrote:I thought that you meant something much more ambitious when you talked of wavelength filtering. Mitaka allows you to toggle different galaxy models on and off, but that's all. This is an incredibly useful feature, which I want to implement via a scheme like Cham's custom classes. But, you could just as correctly call Celestia's alternate surfaces wavelength filtering if you had, say, two textures of Venus from images taken with different filters.

Of course, what I have in mind is much more sophisticated. But from what I see in Mitaka NOW, it becomes clear to me that they are in the course of getting where I want to go since a long time! On the Celestia side ...NOTHING since years despite my efforts.

They have been proceeding pretty fast so far. That's why I think your above advocated "small modifications" to match up with Mitaka will be on the loosing side.

Chris wrote:

Here is a display that illustrates first steps towards the conformal mapping display:

This is just scaling of planet radii. Again, a useful feature, but far from conformal mapping.

--Chris

Surely. Yet locally, a conformal mapping display will look pretty similar to the above image (angular views remaining undistorted, while distances shrink).

EVERY academic/professional in this field is aware of the benefits of conformal mapping in a "large scale" Universe context! After all, it came from Princeton Univ.

The Mitaka team is OBVIOUSLY targetting at a display of the Universe to it's present "end"

Mitaka wrote:Users can seamlessly navigate across the universe from Earth to the edges of the known universe.

They are striving to display the SDSS data along with N-body simulations thereof. Soon or later they will end up with proper conformal mapping. It's for sure. The above image was just a primitive first step, as I also clearly stated.

Here comes one of the real things about Mitaka:

http://4d2u.nao.ac.jp/index_E.html

Mitaka will be/is married with

Mitaka site wrote:Zindaiji : An Astronomical Many-Body Simulation Visualizer

Zindaiji is Windows PC software for visualizing astronomical many-body simulation data. Zindaiji helps researchers visualize large many-body simulations with an intuitive GUI. It also creates high quality movies suitable for theater presentations. It's available from here.

That's where they are going and that's where the "music plays"

Remember what I wrote to you about my Cosmo plans the other day? It's all there already in their planning. Just taking a little more time to be perfect.

Bye Fridger

Here are some impressions of Mitaka in "Cosmo" mode that I would have loved to see also in Celestia.

SDSS galaxies with beautiful large-scale filiaments visible


QCO's (in blue) added around the SDSS data. The scale is now of order 1-4 billion lightyears!
(orange circle = 1 billion ly!)


Bye Fridger