How real is celestia

[Remco65]

Hello,i have celestia for a short time and it is a fantastic program,because you can travel to ervery object,planet,star,galaxy you want. But my question is how real is celestia? My impression is that it looks (very) real,it looks like how it really is ,if you are in the real space.And the illumination of the sun on the planets on celestia,is that how you approximatily it can see through a telescope or if you are in space? And for example,is asteroid Ceres shown more sunilluminated than Uranus? and (less sunilluminated than Venus? And are the positions of the stars and planets really of how it is at the moment? Can you answer this questions? Big thanks. Remco.

[Remco65]

Why no reaction? Can you give me answers please? Thank you. Remco

[GlobeMaker]

R asked:

"How real is celestia? "

98% real , 2% error

My impression is that it looks (very) real,it looks like how it really is ,if you are in the real space.And the illumination of the sun on the planets on celestia,is that how you approximatily it can see through a telescope or if you are in space?

Yes. Don't burn your eyes with too big a telescope.

"And for example,is asteroid Ceres shown more sun illuminated than Uranus?"

The illumination is modeled with consideration given to the human
eyes adapting to darkness.

"and (less sunilluminated than Venus? "

I did not compare them recently. Is one brighter? If Venus is very bright
the human eye adapts to make it less bright. Celestia also adapts
the brightness.

Your computer monitor has a limited range of black (not black).
The monitor has white that is dimmer than Venus' clouds viewed up close.
These limitations cause a narrow range of possible brightnesses.

"And are the positions of the stars and planets really of how it is at the moment?"

Yes, that is very accurate.

[Malenfant]

Hello,i have celestia for a short time and it is a fantastic program,because you can travel to ervery object,planet,star,galaxy you want. But my question is how real is celestia?

Realistic enough, but not actually realistic at all.

The thing with Celestia is that to the first order it's pretty realistic. Planets and moons and asteroids and comets are where they ought to be in our solar system, stars are where they ought to be, and there's a lot of galaxies and binary systems in there now. That's realistic enough for most scenarios.

However, go more in depth and you'll realise that it's not actually that realistic in the detail. The obvious thing to realise is that it can theoretically only be as realistic as what we actually know about the universe, and in practice it's only as realistic as the astronomical catalogues that are entered into it. Fridger's done a sterling job incorporating the galaxies and binary systems so far, but this has barely scratched the surface of what's out there (e.g. triple systems and systems with more stars aren't in there yet, and the galaxies aren't necessarily oriented correctly - they're in the right place and are the right type, but the spirals may go in opposite directions to reality, or they may have more or less arms than in reality). There are a lot of catalogues out there, and a lot of data to incorporate, and there's also limitations of time and computing power too. So at best you can say that Celestia is a decent approximation to the Universe (better in some parts than others), but certainly not a complete representation of the universe.

My impression is that it looks (very) real,it looks like how it really is ,if you are in the real space.And the illumination of the sun on the planets on celestia,is that how you approximatily it can see through a telescope or if you are in space? And for example,is asteroid Ceres shown more sunilluminated than Uranus? and (less sunilluminated than Venus?

This is another place where Celestia IMO spectacularly fails to be realistic (and something that I gripe about all the time ).

The problem is that Celestia renders planets supposedly as if the human eye has adapted to the light levels at that distance from the sun/star. This is largely due to the limitation of representing things on a monitor that is itself emitting light, being viewed by a human eye that is adapting to that (and the light levels in the room). For one thing, I'm not convinced that this is being done accurately at all - first of all, what the human eye sees is highly subjective, being dependent on how good a person's vision is, the distribution of rods and cones in the eye, how their brain interprets the image on the retina, how the eye adapts to bright light against the dark background of space, etc etc. In Celestia (as best as I can understand it), all objects (illuminated by a single star) are normalised and rendered at the same brightness - so Pluto looks as bright as Mercury, despite the fact that the actual light levels are vastly different at those locations.

Furthermore, the planets themselves are not treated as light sources - the light they reflect does not illuminate nearby objects (this is known as 'radiosity'), so you can't accurately see the effects of reflected light on the dark sides of their moons. You can fudge this by tinkering with ambient light levels, but that's not remotely realistic. This also means that distant stars are visible around the edges of lit planets (and stars, for that matter) when they should actually be invisible, since the light reflected from the planet should overwhelm the eye and you won't be able to see the small points of light - basically, your dark-adapted vision would be screwed up by the light source. But despite the fact that Celestia supposedly simulates the human eye, it doesn't account for this somewhat important effect at all.

Plus of course, stars in Celestia don't look at all like they should to the human eye (which would in most cases be blinded or dazzled at the very least), largely due to the fact that monitors can't put out that much light. There are however ways to simulate this (see the EVE Online computer game for example, which does a dazzling star effect rather well).

The illumination of objects rendered by more than one star is also badly flawed and not remotely realistic. See this thread for full details.

Unfortunately, the solution to these illumination problems is not very clear. It may be that there just ISN'T a better way to illuminate objects on a computer monitor in a realtime 3D environment right now (e.g. some aspects might require actual raytracing, which would greatly slow things down). But right now things seem to be somewhat inconsistent or incompletely-implemented.


Another big illumination problem is that of photometry - while the textures of the planets are generally accurate (though naturally the gas giants are only broadly similar to what they really look like right now, since the clouds in their atmospheres are constantly changing as they rotate), in Celestia they don't actually reflect light as they should in reality. The best example of this is the Earth's Moon - in reality the illuminated part of the moon does not get dimmer towards the limb of the body, but in Celestia it does because the photometric function - the way that the object reflects light - is incorrect. Again, this is theoretically something that can be corrected in Celestia by allowing the definition of these parameters/functions for each planet in the ssc, but implementation of this is likely to be a long time coming.

Most people wouldn't even notice these issues, which is why I'd say that Celestia is 'realistic enough' for most purposes, but I think it's important to be aware of its limitations. That said though, it's still a remarkably good piece of software and I'm not knocking the efforts of Chris or Fridger or anyone else who has worked on what is there so far. It's just important to realise that there is still a lot more work to do (especially in the illumination department)...

And are the positions of the stars and planets really of how it is at the moment? Can you answer this questions? Big thanks. Remco.

Generally, yes - at least in the solar system. Outside it gets rather less accurate since we usually don't know the orbits of other planets that well. Also, I'm not sure over what timespan the orbits are most accurate - I presume they get less accurate as one moves far back or forward in time.

[Remco65]

Thanks for the answers. And how can celestia adapt the illuminating of sunlight on the planets to how it (nearly) really is? And shows it the sun at a distance of saturn really how the sun looks from saturn? Can you answer this (last) questions? Thanks.

[GlobeMaker]

R asked :
" And how can celestia adapt the illuminating of sunlight on the planets to how it (nearly) really is?"

Celestia uses mathematics to do that. There is a computer program that
uses logic. This logic is complicated to produce algorithms for that purpose. The dark side and the light side of planets are drawn on
your monitor by software and hardware. If you want more details,
the answer is very long and technical. Read the open source C++
software to learn more. College courses may be needed to be
ready for the C++ readings to be understood.


" And shows it the sun at a distance of saturn really how the sun looks from saturn?"

Yes, the size is accurate. Celestia has a mouse input for the "field of view"
FOV so you can zoom in to see a 1 degree angle view or zoom out to
have a wide angle 120 degree view of the sun from saturn. As the FOV is
changed, Celestia automatically draws accurate sizes for the sun.
Set the FOV to 26 degrees to be realistic.
But if you were really on Saturn looking at the Sun, it would be
too bright to stare at. Your computer monitor cannot get that bright.


Can you answer this (last) questions?

[Malenfant]

R asked :
" And how can celestia adapt the illuminating of sunlight on the planets to how it (nearly) really is?"

Celestia uses mathematics to do that. There is a computer program that
uses logic. This logic is complicated to produce algorithms for that purpose. The dark side and the light side of planets are drawn on
your monitor by software and hardware. If you want more details,
the answer is very long and technical. Read the open source C++
software to learn more. College courses may be needed to be
ready for the C++ readings to be understood.

That's a phenomenally useless and irrelevant answer.

More usefully, the way to show planets how they really are is "simply" a matter of implementing well-known photometric functions to the surfaces of the bodies shown in Celestia (eg lommel-seeliger function, buratti function, or even hapke function). However, I think only Chris knows how to actually do this, and it may be that these functions can't even be rendered using OpenGL in a realtime renderer. I get the impression though that it can be done, it just takes a lot of very technical programming - the entire rendering algorithm probably needs to be redone to allow this.

Unfortunately while I know a fair bit about the actual science behind it, I have no clue about how to implement it in a program .

[Remco65]

What is FOV? And in reality is pluto less sunilluminated than mercury? I think of course yes. Can you answer this 2 questions? Thanks.

[selden]

FOV = Field Of View -- the width of the display that Celestia shows you

Since Pluto is so much farther from the Sun than the Earth is, of course its surface gets much less light than the Earth does.

However, Celestia draws all planetary objects equally bright. The rationale is that your eyes would adapt to the local light level and the planets all would look just as bright when you are there.

[Malenfant]

What is FOV? And in reality is pluto less sunilluminated than mercury? I think of course yes. Can you answer this 2 questions? Thanks.

FOV = Field of View.

And yes, Pluto would be dimmer than Mercury. Mercury is illuminated by over four times the light that Earth is lit by because it's a bit more than twice as close to the sun as the Earth (inverse square law). Pluto on the other hand gets about 1/1600th of the light at Earth, because it's about 40 AU away. So the ratio of illumination between Mercury and Pluto is about 6400:1 (ie the sun at Mercuty is 6400 times brighter than the sun at Pluto (at 40AU)).

[Malenfant]

However, Celestia draws all planetary objects equally bright. The rationale is that your eyes would adapt to the local light level and the planets all would look just as bright when you are there.

Is there actually any evidence that this actually happens though? I don't think there is.

After all, I can go outside at night and look at things lit by the full moon, but they don't appear as bright as they do when they're lit by the sun .