Amazing colors

[jogad]

Hello




Amazing!… yes. But witch is correct?

Theses pictures are taken from the great “Celestia Educational activities Vol 4E” by Frank Gregorio. We are landed on Titans’ surface and our space ship has blasted the clouds out.

- The first image is seen with the render path set to Open GL and looks like other render path option (but one!)
- The second image is shown with the render path set to “Open GL 2.0” and the color of the sky is very different.

------

Another example even more confusing can be found by clicking HERE . We are on Mars and the Sun is rising. Cycling by Ctrl-V through the render modes show us (with my Nvidia graphic card) three different colored versions of the scene.

So the sky of Mars may be pink or bluish while the ground color varies too.

Where is the truth?

[Hungry4info]

You're looking for scientifically accurate visualizations in Celestia?

stop...

[chris]

You're looking for scientifically accurate visualizations in Celestia?

stop...

Please elaborate on what you mean by that. While Celestia's rendering of atmospheres isn't perfect, there many situations where Celestia's rendering is quite accurate.

--Chris

[Hungry4info]

I know it's limitations with software, but

-Comet tails
-Stars (still rather dim)
-Lack of ring shadows onto moons.
-Both sides of Saturn's rings are equally lit.
-Lack ring-shine onto Saturn.
-Atmospheres lacking some opacity without a cloudlayer. (In Celestia, descend through Titan's/Venus' atmosphere and the entire surface suddenly becomes visible once you're past the cloud layer. In real life, the surface fades into view.
-Brightness being independent of the distance to the star. (a 0.99 light year-away object will be rendered just as bright as a 0.99 AU away object).

I'm not trying to bash Celestia, I love the program, it's just not what I would consider a good representation of what the human eye would see.

I want my monitor screen to go pure white from saturation when I look at a nearby star. Lol.

[jogad]

I am not a specialist. Neither in computer science nor in astronomy.
So I imagine that the different options bring both a visual improvement AND accuracy in colors.

I am surprised that the Open GL 2.0 gives a result so different in comparison with the others.
Is it really the programmer’s will that the same object appears to be brown or blue depending on the user’s choice?
That is perfect if the user knows the real color. It’s up to him. Or if he just want to get a pretty picture.

But as I use Celestia to familiarize with astronomy, it is a bit confusing for me.

So the newcomer’s real question is: when should I use (or not use) the OpenGL 2.0 render path ?

[cartrite]

Most of the code development done lately has been for the opengl 2.0 render path. The atmospheres especially have been upgraded in this path. This is why there is such a difference in color, quality, etc. between the different paths The open gl path also has a better model for rendering atmospheres so that would be the most realistic. It is still under development though and is improving with time.
cartrite

[ajtribick]

First thing to point out here is that turning the clouds off on a world known for its 100% cloud cover isn't exactly the way to go for producing scientifically accurate rendering! If you view Titan with clouds from above the atmosphere, only the OpenGL 2.0 path renders correctly the blue layers above the orange clouds. This is not to say there aren't problems and improvements to be made, but choosing a fantasy scenario such as cloudless Titan is not really a good test of reality! With the clouds on, I get a nice orange sky in the OpenGL 2.0 render path.

Also, Hungry4info, I agree you are not trying to bash Celestia. You are bashing Celestia. Way to minimise the amount of effort that is being put in to get scientifically reasonable results in Celestia, some of which involves coming up with rendering strategies to handle thousands of objects. It is totally unreasonable to say that there is no realistic rendering in Celestia just because there is still work to do in some areas.

[rthorvald]

That looks like my old Titan scenario. If Frank used all of it in his Activity (i don?t know), Titan will have a second atmosphere layer implemented to simulate the blue bands. Maybe that is the culprit here?

- rthorvald

[jogad]

I Agree that Titan without clouds is not a perfect example. I choose it because it is a beautiful scene.

But Titan is NOT the problem. We have a similar behaviour. On Mars, with the standard distribution of celestia 1.5.1 we have a similar behaviour.

It's more demonstrative in celestia itself by clicking HERE

These images are shot with respectively
- “Open GL vertex / Nvidia combiners” and
- “Open GL 2.0”

Why a so big difference ?

Don’t see any attack against Celestia. It is by far the best program ever seen anywhere and I love it. That is to be said. It is not only a cold scientific program but also a very creative one. Thanks a lot for that.

[ajtribick]

The OpenGL 2.0 render path supports a more physically-based simulation of atmospheres than the other render paths, so you will obviously get different results with it.

From what I can tell, the OpenGL 2.0 path uses the various Mie, Rayleigh, etc. parameters while the other render paths use the older system based on the parameters Lower, Upper and Sky. Since we don't want to drop atmosphere rendering in older render paths the old system is still supported, but I don't think anyone's edited the parameters for the old system for quite a while, in contrast to work that's been done trying to get the new system working correctly, hence the discrepancy. If you want to tweak the Lower/Upper/Sky values to try and approximate the OpenGL 2.0 view on the other render paths, go ahead.

In general, you should be using the OpenGL 2.0 path if your system supports it.

[cartrite]

Why a so big difference ?

I'll try to answer this as best I can.
This is a portion of solarsys.ssc from celestia 1.5.1. It describes Mars to the Celestia program. I don't think I altered it.

Code: Select all

   Color   [ 1 0.75 0.7 ]
   HazeColor [ 1 1 1 ]
   HazeDensity 0.45
   Radius 3396 # equatorial
   # Oblateness 0.0069

   Atmosphere {
      Height 30
      Lower [ 0.8 0.6 0.6 ]
      Upper [ 0.7 0.3 0.3 ]
      Sky [ 0.83 0.75 0.65 ]
                Sunset [ 0.7 0.7 0.8 ]
                # Slightly bluish sunset, as seen in true color pictures
                # from Pathfinder

      Mie 0.0024
      MieAsymmetry -0.15
      Rayleigh [ 0.0010 0.0006 0.0003 ]
      Absorption [ 0 0 0 ]
      MieScaleHeight 20
   }


This is the part that controls “Open GL vertex / Nvidia combiners” render path.

Code: Select all

   HazeColor [ 1 1 1 ]
   HazeDensity 0.45
   Atmosphere {
      Height 30
      Lower [ 0.8 0.6 0.6 ]
      Upper [ 0.7 0.3 0.3 ]
      Sky [ 0.83 0.75 0.65 ]
                Sunset [ 0.7 0.7 0.8 ]


The atmosphere that you see here is mostly because of the 2 lines for HazeColor and HazeDensity. I think the values for Lower, Upper, and Sky are rgb values. This is where the color comes from.


This is the part of that file that controls the “Open GL 2.0” render path.

Code: Select all

      Mie 0.0024
      MieAsymmetry -0.15
      Rayleigh [ 0.0010 0.0006 0.0003 ]
      Absorption [ 0 0 0 ]
      MieScaleHeight 20


These are parameters that try to model the light scattering and light absorption. There is a place to see a discussion about these settings and the theory behind them.
http://forum.celestialmatters.org/viewt ... =2048#2048
Hope that helped.
cartrite

[Hungry4info]

Also, Hungry4info, I agree you are not trying to bash Celestia. You are bashing Celestia.

Lol aww

Way to minimise the amount of effort that is being put in to get scientifically reasonable results in Celestia, some of which involves coming up with rendering strategies to handle thousands of objects.

I meant not to do that, but rather to point out simply that Celestia does not (yet?) realistically display everything. I realise that a huge amount of effort has been put into Celestia, effort for which users of Celestia like myself are certainly grateful for, but I try to keep a realistic mindset about Celestia. Sure, there are areas where Celestia shows a fantastic degree of realism, The moons of Mars, and Saturn's moon Phoebe, look very realistic to me (at least not when extremely close).

It is totally unreasonable to say that there is no realistic rendering in Celestia just because there is still work to do in some areas.

I agree 100%, and nobody here said that Celestia has no realistic rendering.

[t00fri]

I am not a specialist. Neither in computer science nor in astronomy.
So I imagine that the different options bring both a visual improvement AND accuracy in colors.

I am surprised that the Open GL 2.0 gives a result so different in comparison with the others.
Is it really the programmer’s will that the same object appears to be brown or blue depending on the user’s choice?
That is perfect if the user knows the real color. It’s up to him. Or if he just want to get a pretty picture.

But as I use Celestia to familiarize with astronomy, it is a bit confusing for me.

So the newcomer’s real question is: when should I use (or not use) the OpenGL 2.0 render path ?

YES to your last question!

After following this discussion for a while, I really don't see why it should not be intuitive that with different rendering scenarious the differences can be substantial:

Firstly the main rendering path is OpenGL 2.0, the others will be DROPPED soon or later. For now they only survive as "workarounds" for people with VERY old graphics cards...

++++++++++++++++
Obviously, these last resort graphics paths will not get the same amount of attention as OpenGL 2.0.
++++++++++++++++

As cartrite mentioned already, ONLY in the latter, the sky rendering is based on the general Mie theory for the scattering of light on atmospheric droplets. Hence, there is simply NO comparison in (potential) realism as compared to non-OpenGL 2.0 rendering paths.

This was written several times more or less clearly...

However, while serving as a general atmospheric framework, the Mie atmospheres depend on a number of parameters. If these are badly set, you will get crazy atmospheres. So anybody without solid understanding of Mie theory, who produces add-ons, can make a real mess...

Tha't about the "bandwidth" of this issue

Fridger