Ultimate space simulation software
To me the reason why it looks wrong is because of the lighting effect itself. It looks more of a blurry, pixelated cloud that does not contrast well into the surroundings the closer you get.When reality looks wrong, change reality?
The temperature of an accretion disk with respect to radius generally follows [tex]T(r) \propto r^{-\frac{3}{4}}[/tex]
Which as a plot, looks like this:
If the disk is hot enough that it emits x-rays (wavelengths between 10-8 and 10-11 meters), then the peak temperature must be hundreds of thousands to millions of Kelvin.
Suppose as in the screenshot that the peak temperature of the disk near the white dwarf surface (about 10,000km out) is about 5 million Kelvin. Then at twice that radius, the temperature is still about 3 million K. At 10 times that radius it's about 900,000K. At 100 times that radius? 150,000K. You get the picture. Even at large distances it's still so hot as to be glowing in ultraviolet! And since luminosity follows temperature to the fourth power, that's a huge luminosity even compared to a typical star, and it would be absolutely blinding.
How far out would we have to extend the disc such that the temperature is merely 5000K and thus peaking in visible light? That would be ten thousand times the white dwarf's radius! That's 100 million kilometers -- greater than the distance from the white dwarf to the star it's accreting from!
So that's reality. White dwarf accretion disks are stupidly bright all the way out! Trying to extend the disk so that the outer edge isn't stupidly bright would result in a disk that is stupidly unrealistically big. The only reasonable correction to make it "look right" would be a massive exposure compensation, where everything else including nearby stars are black, as is shown in An'shur's screenshots.
Exposure is honestly the best method to look at details of stars. I think this game should strive for realism, and sadly the lighting suffers due to compensation for the viewer. But we have exposure, so there's no reason to compromise reality over presentation when there are tools to compensate.I agree the visual appearance of the disk is off-putting in the sense of "what the heck am I looking at?" and "I'm pretty sure if I were actually looking at such a thing in nature, it wouldn't look like that." Of course, if you did really look at one, it would look like *you are now blind*. But the presentation should be different. The way it is right now is immersion breaking.
I don't agree with changing the physics -- if nature makes the whole disk blindingly bright, then it should be portrayed in such a way that the viewer understands that that's how they really are. I agree with making the disk bigger to compensate even less -- it would actually make the problem worse, with the white dwarf and its partner star both engulfed in a huge blinding sea of whiteness. Besides being unrealistic, it would be even more confusing.
I think the best option is an exposure compensation, where everything else smoothly fades out when you get close enough to resolve the disk, and that you can see structure in the disk instead of a solid pure-white blob. You can already do this manually, but having it done automatically would look nice and hopefully not be so weird and immersion breaking.
Yeh, but there has to be some engines S.E can use in order to make this easier.I know SpaceEngineer has already given a lot of thought to this, so we might imagine the solutions cannot be very easy to implement or else we'd already have them. Mainly the problem is how to merge auto-exposure with how the rendering system handles color-temperature-luminosity relationships of objects so that it looks good in all situations.
He has, he discussed it on the previous forum as well and he and i have discussed it quite a bit.
This is a small part of a much larger problem.
Accretion disk is indeed overwhelmingly bright. I made some computations using analythic model (alpha-disk model) for accretion disk around Sagittarius A* in the Milky Way center. Using it's mass of 4.2 million solar masses and realistic accretion rate (3 Earth masses per year), I derived this temperature profile:
max temperature at 4.3 Schwarzschild radii - 96,000K
temperautre at 1000 Schwarzschild radii, where SE rendering stops for now - 2600K (still too bright)
temperature at 3600 Schwarzschild radii - 1000K (so rendering radius in SE should be increased 4x times)
Matter at 96,000K have extremely high brightness. According to the Stefan-Bolzmann law, surface brightness is proportional to T4. So 96,000K is 80,000 times brighter than Sun's surface. It will be saturated to white in SE if exposure is more than 0.0001.
Screenshots here showing accretion disk around Sgr A* with different levels of exposure: 1, 0.01, 0.0001:
Temperature in accretion disk around stellar black hole with high accretion rate is even more extreme. For 5 solar mass black hole, eating 0.001 solar mass per year, it will be 330 million Kelvins at 4.3 Schwarzschild radii and 10 million Kelvins at 1000 Schwarzschild radii. Temperature falling down to 1000 Kelvins only at 200 million Schwarzschild radii, or 20 AU from the black hole. This means what stellar black hole's accretion disk can't be such cold, because it's outer edge lies much closer - defined by orbit of the secondary stellar companion, which feeding the black hole with its own matter.
This also means what SE cannot render accretion disks realistically, at least for now. SE can't render temperatures more than 100,000 K (all higher will have the same color and brightness) and can't render accretion disks of any radius (1000 Schwarzschild radii is limit for now). I am forced to use some cheats like smooth falloff of the disk's outer edge and limiting it's maximum temperature.
