Celestia 1.4.0 prerelease FAQ

[chris]

Celestia 1.4.0 has two major new features: real support for multiple star systems and improved rendering using GLSL shaders. There's more work to be done on both features, and there's been a lot of confusion in distinguishing between bugs, unimplemented functionality, and intended behavior. I hope this FAQ helps clear things up somewhat.

Can my graphics hardware use the new OpenGL 2.0 render path?
If you have a GeForce FX, a GeForce 6xxx series, a Radeon 9500 or higher, or a Radeon X series, then your hardware is capable of using the OpenGL 2.0 render path. You must also have a very recent driver--older ones don't have the necessary OpenGL extensions.

My hardware only reports supporting OpenGL 1.5. Does this mean I can't use the OpenGL 2.0 path?
No. All the graphics boards listed in the previous question can use the OpenGL 2.0 path if you've got a recent driver. The critical OpenGL feature that the drivers must support is the GLSL high level shading language. Although GLSL is part of the OpenGL 2.0 specification, it is also exposed via an extension in older versions of OpenGL.

Will I need to buy a new graphics card to run Celestia 1.4.0?
No. If Celestia 1.3.2 runs with your current graphics card, so will Celestia 1.4.0 and subsequent versions. The new OpenGL 2.0 features won't be available unless you have recent hardware, but the old render paths will work just as before.

I've turned on the OpenGL 2.0 render path, but planets are all rendered red. What's going on?
The red pixels indicate that an invalid GLSL shader was loaded; to help debug shader problems, I wanted it to be obvious that there was a problem. You probably have an ATI graphics board. I develop on NVIDIA hardware, so I miss bugs that only appear with ATI drivers and hardware. If you want to help me debug the problem, post the contents of the file shaders.log (it should be in your Celestia folder) to the bugs forum.

What binary star systems can I visit in Celestia?
No binary or other multiple systems are included in Celestia yet. We're still working on converting catalogs of star orbits to Celestia's format. For now, you'll have to make your own system or download one of the several add-ons created by a forum member.

I've placed a planet in orbit around a star in a binary system, but I only see light it lit by one of the stars. What's wrong?
It's likely that this is correct behavior. The distance between the planet and the star it orbits is probably much less than the distance from the planet to the second star. If the stars are approximately equal absolute brightness, this means that the light from the closer star will be so much more intense than the light from the more distant star that the secondary illumination will be imperceptible. Remember that the brightness of light falls off in proportion to one divided by distance squared. If the second light source is a lot further away than the primary, it must be intrinsically much brighter in order to visibly illuminate a planet.

How many simultaneous light sources can Celestia handle?
Up to two in the OpenGL vertex shader paths, and up four in the basic, multitexture, and OpenGL 2.0 paths. The light sources with the greatest apparent brightness are given priority.

Ring and eclipse shadows don't look right when there are multiple light sources.
Only the OpenGL 2.0 render path handles shadows correctly when there is more than one light source. In the other render paths, only the shadow from the brightest light source is considered. The OpenGL 2.0 path can render ring shadows and up to three eclipse shadows per light source.

What are some known problems with the current prerelease?
No orbits are shown for stars. Overlay textures don't yet work in the OpenGL 2.0 path.

[Startyger]

When will the Mac/Linux versions be available?

kindly awaiting your reply....

[Harry]

When will the Mac/Linux versions be available?

There is a Mac prerelease available here http://www.celestiamotherlode.net/creators/dirkpitt/ (there is also a thread about it).

And for Linux you can always try the compiling the current CVS code. Of course that's a bit more work than just downloading a ready-made installer, and some people have been reporting problems with some versions of gcc/libraries/etc.

Harald

[no_fear]

Hello!

Im a new user of Celestia. Ive been trying to find the link, in this forum, that would lead me to the Celestia 1.4.0 for Windows download page. Could you please tell me how can i get there?

[selden]

http://www.celestiaproject.net/~claurel/celest ... .0pre6.exe

If you have experience building software under Windows, then you might consider building the most recent version from sourcecode. For details, see the "Stickies" at the top of the Celestia Development Forum, especially Free Windows-based compiling for Celestia.

[no_fear]

Cheers!!!

[Giorgio]

I hope I'm not posting a stupid question, but: does Celestia 1.4.0 prereleases support haze for ATI cards?

[XRaiderV1X]

im assuming it would be safe to modify v 1.4.0 to run in low res by renaming the hi res folder etc ?

[selden]

Yes, renaming the medres texture folder would work just the same as with other versions of Celestia. If it can't find the medres textures, it'll show the lores ones.

[Startyger]

LOL will it be processor independent? 64bit? just asking.... i'm glad to see it is there a possibililty that Cheetah3D and Celestia can talk about the format?

[ch0kl0]

hello!

I??m a new user of Celestia.
I want to know what is the difference between the older version and this new.
thanks!

[Hunter Parasite]

when will the full release for v1.4.0 be? the prereleases dont work on my pc.

[selden]

If the prereleases don't work, then the final release probably won't work, either. You need to describe in detail exactly what problems you're having, otherwise they won't get fixed.

[Hunter Parasite]

when i use the pre realeses they dont respond when downloaded, it say the files dont exist

[selden]

Some of the prereleases are only Celestia.exe itself and not the entire package.

v1.4.0pre6 is the one you should be using. It's a complete package. Is that the one you installed?

If so, what directory did you install it in?
Did you check the boxes to create a new desktop icon and to associate URLs? They're required except in very special circumstances.

What files does it say it can't find?

[Hunter Parasite]

It turns out, I had some errors with my PC so I did some dignostics and it works.

[Malenfant]

Might be useful to mention here that pre7 has been available for a few weeks and can be downloaded at:

http://www.celestiaproject.net/~claurel/celest ... .0pre7.exe

[selden]

Celestia v1.4.0pre8 for Windows is now available at
http://www.celestiaproject.net/~claurel/celest ... .0pre8.exe

[Fightspit]

Celestia v1.4.0pre8 for Windows is now available at
http://www.celestiaproject.net/~claurel/celest ... .0pre8.exe

What are there more than Celestia pre7 and FT1.2 ?

[selden]

It incorporates all of FT1.2, an updated GTK interface, an updated Windows Render/ViewOptions menu, plus quite a few bug fixes.

If no new bugs are found, it probably will become the final release, except that the changelog (below) also will describe the updates that have been made to the exoplanets and solar system moons catalogs.

Code: Select all

1.4.0
* New stars.dat format with HD catalog numbers and parallax errors omitted
* Added catalog cross indexes for HD and SAO catalogs
* Fixed bugs in parsing catalog numbers
* Added white dwarf temperatures and  spectral types DA, DB, DC, DO, DQ, and DZ
* Added handling for partial spectral types where one or both of the subclass
  and luminosity class are unknown
* Permit extended star attributes in .stc files: mesh, texture, rotation
  elements, semiaxes
* Support multiple star systems of any complexity with orbits and barycenters
* Changed .stc loader so that a star definition with a duplicate catalog
  number replaces the previous definition
* Made catalog numbers optional in .stc files; stars can be defined without
  having to create fake catalog numbers
* Star colors now chosen based on temperature; either 'classic' and black body
  colors may be selected
* Bound % key to toggle between star color tables
* Fixed black holes: infinite radius bug gone, rendered just as a plain black
  sphere
* Added comprehensive catalog of all knows stars within 25 light years of the
  Sun, including orbits for multiple star systems
* Created new command line tools for building binary star databases
* Removed limitation that two or more solar systems couldn't be shown
  simultaneously
* Show light from multiple nearby stars illuminating planets
* Implemented new GLSL render path; NVIDIA combiners and GeForceFX paths
  deprecated.
* Display UTF-8 superscript digits in some star names
* Updated the Titan texture. It accomodates all published high resolution
  imaging from the Cassini mission until and including the flyby of 03/31/05.
* Updated the Iapetus texture. It accomodates all published high resolution
  imaging from the Cassini mission, including also a unique hires photo taken
  in "Saturn shine".
* Added catalogs of 163 visual and 39 spectroscopic binary orbits,
  respectively, (S?¶derhjelm 1999, Pourbaix 2000) with known primary/secondary
  mass ratios.
* Added an extended and precise data base of 942 galaxies
  (Steinicke's rev. NGC/IC, 2005) with a magnitude cutoff Bmag < 12.
* Included the commented PERL scripts used to extract and adapt the binary
  orbit and  galaxy data from published professional catalogs.
* Added --extrasdir command line option for specifying additional directories
  to search for add-ons
* Allowed time scale to be modified while simulation is paused
* Replaced Windows date picker with a custom one that can be used to set dates
  prior to 1752.
* Fixed a bug in the Windows time picker that was causing the set time to be
  off by however many seconds the dialog was active.
* Added parent field for stars in celx scripting's getinfo command
* GTK+ front-end overhaul
   - source split into multiple files
   - performance work to minimize number of preference resyncs
   - redesigned interface to use GtkActions and GtkUIManager (XML)
   - added preference saving to file
   - commented heavily
   - new about dialog
   - made non-modal: GoTo, View Options, Tour Guide, Browsers, EclipseFinder
   - changes to allow for compilation on Windows
   - greek letters in Star Browser names
   - added tab-completion support in overlay entry
   - replaced deprecated widgets in several dialogs
   - new GTK save/open dialog
   - added "copy" and "open" url functions
   - updated GConf schema with new options
   - fixed automake bug where GConf schema would always install
   - removed linking against glut for no reason
* Implemented complete precision catalog (Steinicke's revised NGC/IC, 2005) of
  10610 galaxies with
   - distance information from  four catalogs ( 6 methods ),
   - <= 4 alternate names,
   - info-URLs,
   - absolute magnitudes and
   - correct sizes & orientations in space, as calculated from catalog
     parameters.
* Included well commented Perl script (deepsky.pl) as documentation. The used distance determination method is indicated in catalog for each galaxy.
* Updated binary star data base (visualbins.stc, spectbins.dsc) and respective
  PERL catalog extraction scripts (visualbins.pl, spectbins.pl), such as to eliminate double occurences wrto nearstars.stc (, which remained unaffected).
* Added much improved galaxy templates.
* Implemented galaxy coloration based on hue profiles from true color
  photographic imaging. Being older than spirals, elliptical galaxies have a somewhat redder coloration.
* Implemented a realistic relative apparent brightness display scheme for
  galaxies.
* Added apparent magnitude display as function of distance on canvas.
* Added display of DSO Type (Galaxy,...) and Hubble subtype in case of
  galaxies.
* Changed the galaxy label color from yellow to blue-green for better
  visibility.
* Did extensive octree rebuilding. Culling of deep sky objects is logarithmic
  now, enabling support for very large catalogs with little performance expenses.
* Performed optimized tuning of octree parameters that practically recovers
  the "pre-galactic" performance, despite a large data base of 10610 galaxies.
* Added support for multiple names of deep sky objects and implemented
  non linear keyboard selection of names with corresponding tab completion, etc., like with the rest of the celestial bodies.
* Replaced deep sky catalog class with a more complex database, following the
  original star database layout. This allows for more sophisticated and fast treatment of massive catalogs.
* Enabled very fast search of close-to-viewer deep sky objects.
* Implemented brightness corrections involving the angle between 1-2 surface
  normals and the viewing angle, such that the brightness distribution
  gets ~ uniform  for all galaxy orientations and Hubble types. This also
  avoids excessive, unnatural brightness effects in edge-on views.
* Galaxies get brighter with decreasing observer distance according to
  realistic human eye sensitivity profile.
* Brightness display optionally coupled with Automag scheme.
* Enabled very fast picking of deep sky objects.
* Placed galaxy templates in 'models/' directory.
* Added various crucial fixes in galaxy.cpp,
   - factor 2 was missing in the 'size' definition via the radius such that
     distances among DSO's were grossly incorrect;
   - fixed correct ordering of (non-commuting) operations in matrix m.
     This gave incorrect galaxy orientations;
   - To save space and to improve display: generate neat spherical E0
     template from S0 disk template via rescaling by (1.0f, 3.8f, 1.0f);
   - fixed wrong x,y alignment of elliptical rescaling.
* Added code to ease compilation with newer Xcode versions (Macintosh).
* Updated src/celengine/Makefile.am for Linux building.
* Introduced a new cel function => renderflags {set "nebulae"}.
* Reduced the default value of 'faintestAutoMag45deg' from 8.5 to 7.0.
* Fixed the visibility of the Milky Way during day-time and the abrupt
  appearance or disappearance of galaxies near the day-night border.
* Enabled "geosynchronous Follow" (key 'Y') also for selected DSO's
  (yet it does not seem to be very useful...).
* Moved render flag check in the code such that labels can be drawn even if
  e.g. renderGalaxy flag is turned off.
* Suppressed apparent magnitude display for galaxy add-ons, lacking
  specification of an absolute magnitude (absMag) input value.
* Deleted various source code files that became superfluous.
* Improved comet display. Introduced a neat scheme implementing comet tail
  fading (between 4 and 6 AU for Sol). Systems with several suns and luminosities different from the solar one are accounted for.
* Implemented a new, simple scheme avoiding overcrowded galaxy labels. 
  Their "importance"  is sorted according to apparent magnitude! Thus the labels of the brightest galaxies pop up first upon zooming in...