Problem Spacecraft illumination near terminator

[lcs]

I consider this a problem unless it can be changed. For example, the ISS starts dimming into a yellowish color way before the sun starts to sink below the horizon. This simply does not happen as anyone who watches NASA TV can verify. The onset of dimming should be rapid and over with in about 10-15 seconds. This was done almost correctly in Orbiter 2006. Now in 2010 inexplicably dimming starts 10 minutes before sunset! How to change this?

 

[River Crab]

I'm pretty sure you can fix it with a replacement of the visual parameters in Earth.cfg. I'm not sure of this but I think you can just replace the visuals with those from 2006.
These are the parts I mean:

; === Atmospheric Parameters ===
;AtmPressure0 = 101.4e3         ; pressure at zero altitude [Pa] (defined via module)
AtmDensity0 = 1.293            ; density at zero altitude [kg/m^3]
AtmGasConstant = 286.91        ; specific gas constant [J/(K kg)]
AtmGamma = 1.4                 ; specific heat ratio c_p/c_v
;AtmAltLimit = 200e3            ; cutoff altitude [m]
AtmHorizonAlt = 64e3           ; horizon rendering altitude [m]
AtmHazeExtent = 0.14           ; horizon haze extent
AtmColor0 = 0.55 0.75 1.04
AtmHazeColor = 0.6 0.8 1.0
AtmFogParam = 5e-5 3e-5 4e3
AtmFogColor = 0.55 0.85 1.10

; === Cloud parameters ===
CloudAlt = 7e3                 ; altitude of cloud layer
CloudRotPeriod = 1e6
CloudShadowDepth = 0.3
CloudMicrotextureAlt = 35e3 300e3

; === Visualisation Parameters ===
MaxPatchResolution = 14        ; surface texture resolution limit (1-12)
MinCloudResolution = 1         ; cloud layer from this resolution
MaxCloudResolution = 8         ; highest cloud resolution level
SpecularRipple = TRUE          ; enable specular water microtexture

Note that if you replace this whole thing, the pressure model will be different and so you'll just have the atmosphere model from 2006, which is less accurate.

 

[lcs]

Thanks. When I uncomment AtmAltLimit the reduction in ISS illumination occurs much closer to when the sun touches the horizon, as one would expect. The default is 200 km. Does anyone know what this parameter means? If it is literally the upper cutoff in the atmosphere model then it seems there is a flaw in the optical model for light attentuation when it is commented out.

Orbiter IMO does not properly simulate the appearance of the backlit Earth horizon at sunset. The colors are too warm and the physics of light scattering more strongly at short wavelengths (Rayleigh scattering) is not simulated. There should be a bright blue band on the horizon as the sun approaches it and after it sets. To me this is the biggest weakness in an otherwise excellent product.

 

[statickid]

you should make an add-on that mends this problem ics :thumbup:

 

[lcs]

An add-on cannot fix the physics model, and the code is closed-source. So I can only appeal to Martin to at least look at Celestia and see how Earth's horizon should look as the sun approaches and sinks below it.

 

[Moach]

the implementation of this particular effect has always puzzled me... specially back when i was developing my planetary navigation and effects test for a game concept of mine...

is that done through some shader? or is there something else going on?

 

[lcs]

My impression is that Celestia has some kind of atmosphere light scattering model (or fudges it) separate from their earth shader, while Orbiter uses ONLY a shader. But I'll defer to the experts.

Here is an attachment showing Celestia horizon effects. While not perfect, and still lacking the brilliant glow of the horizon immediately after sunset, to me it is more realistic than what Orbiter produces:

 

[Moach]

the OGLA client seems to have that effect going more in the ways you describe....
it uses raycasting to achieve that, which makes for some very cool effects

i'd still like to know how it was done in Orbiter stock graphics tho

 

[martins]

I consider this a problem unless it can be changed. For example, the ISS starts dimming into a yellowish color way before the sun starts to sink below the horizon. This simply does not happen as anyone who watches NASA TV can verify. The onset of dimming should be rapid and over with in about 10-15 seconds. This was done almost correctly in Orbiter 2006. Now in 2010 inexplicably dimming starts 10 minutes before sunset! How to change this?

The problem of early dimming will be fixed in the next beta. That was caused by light attenuation through the atmosphere being linked to the AtmAltLimit parameter. In the next beta, there will be a separate parameter AtmAttenuationAlt for this, and it will be set to 100km for Earth by default.

The visual appearance of the atmosphere during sunset/sunrise I'll leave for somebody to fix who cares more about it than I do. What about yourself? Code for the graphics subsystem is at OVP.