Celestia Forums

See the updated attached source.
I've developed and tested that with Linux and don't expect problems on the other supported platforms, maybe there are.
I'm pleased to get further reports and feedback.

Note: The source will work with the current trunc version 4741 (1.6.0).
Let me know if you want a (simple) version for 1.5.1 (or an x86_64 Linux binary)

Thus, if you are interested in real 3d impressions with celestia, just try to compile with celestiacore.cpp
from the zip archive instead of the trunc version

Short description of what the patch will do:
* SHIFT F2 or F3 switches between anaglyph modes and normal mode

There are 8 anaglyph modes available now. They are using different anaglyph building techniques.
Modes 1 (the previous) up to 5 (except 4) work pretty good for flying around the stars and milkyways with high speed
I prefer modes 4 6 7 8 for single individual scenes and for screenshots.
Just check them out and have fun.

You could always play around with
* SHIFT F5 increases apparent "Depth Level"
* SHIFT F6 decreases apparent "Depth Level"

*(Splitting views will disable anaglyph mode because it's not reasonable yet)

Tips & Hints for anaglyph mode:
The aiming on object I had described here before, doesn't work anymore, but I think that doesn't greatly matter until there is a screenpointer with Z value available.
If you see "ghost" objects or disturbing shadows, calibrate your monitor or tinker with its colour temperature settings for better results.

Why red-cyan and not red-blue red-green?
In the matter of colour distortion, red-cyan gives us the best capability.
You'll also get a 3D impression with the other glasses but with much more colour distortion.
Last but not least anaglyph rendering brings out the cheapest way to have a real 3D impression, and as far I'm concerned,
it's just a must have for celestia, isn't it?

Screenshot with mode 8, try to touch the parabolic reflector



--------------------------------------------------------------------------------------
A bug fix: (?)
ogg video capturing was broken (at least for Linux) by changing oggtheoracapture.cpp

< std::srand(std::time(NULL));
---
> #ifdef _WIN32
> std::srand(std::time(NULL));
> #else
> std::srand(time(NULL));
> #endif

Should'nt it be " std::srand(std::time(NULL));" for all platforms?
Update: With trunc rev 4741 ogg capturing works as expected again, with or without this determination.
I don't know the reason for the determination so I keep this information here for now.
-------------------------------------------------

regards
neTear

A. Einstein said, if we knew what ( it was ) we were doing, it wouldn't be (called) research, would it?

I can confirm that this works very well in mac os x (10.5.7 intel) - had trouble getting the shift-f2 key to work but commenting out the other f2 command in the source solved that problem.

Many thanks!

Maybe "modifier" (keys) does not work with MacOS as expected, at least any other "normal"
key could be taken for toggling and setting within a main development tree.
Is "Shift F5/F6" working for you?

BTW, there are two lines of debug code I've forgot to remove.
Just search for "fprintf(stderr" and remove these both lines.
Not sure, but I think especially windows compilers will stumble on this.

regards, and thank you for the information
neTear

I can confirm that this patch works on Windows XP, using the VC++ 2008 compiler. I've got no problems with the modifier keys.

Hi friends!
Is it possible to have a compiled version for windows (as for Celestia pre-releases!!), so I can try it on Vista?
I have not the compiler
Thanks a lot!
Ciao
Pierluigi

Could you add some screen shots please.
It will be really illustrative of the improvements you are doing.
Thanks

Take a look at viewtopic.php?f=4&t=13904

Do not forget to attach your red-cyan glasses.

regards
neTear

Hello,

It is an exciting idea but that doesn’t work as expected for me. I don’t see actual relief.
Here is an example:

I hoped to get something like that

But instead, I get this:

The red and cyan canals seem to be taken from the same point of view. The resulting image is just behind the plane of the screen but remains flat.

I watched your video (very aggressive web site) and indeed I had a sensation of relief. But if I pause it, the relief disappears! I feel a similar relief perception even with standard visualization when I move rapidly in a field of stars.

Reagards

Hi jogad,

the different "channels" aren't taken from the same point of view actually, so that flying through universe works by now.
It also works with relief for most individual single scenes (at least for me). (Have you played around with shift F5/6, Focus, Object rotation, object placement?)
What you expect is the same what I tried to describe with:

it's possible to have closer visible objects seemingly coming out of the monitor instead of let them all hanging behind it..

This is a question of the (virtual) dimensions, and is much more trickier to implement because it requires to be dynamic.
I'm having a clue how to implement this and already tried, but didn't succeed yet.
I'll keep busy with this issue (will take some time, have to do other things at the moment...).

"very aggressive web site"=better don't use it?

Edit: Hm, what about this one? Source with more anaglyphic modes will follow next weekend.

regards