Author Topic: Useful? Take a look, solar-system orbit calculator.  (Read 8631 times)

0 Members and 1 Guest are viewing this topic.

Offline -Sara-

  • 29
Useful? Take a look, solar-system orbit calculator.
A nifty tool, which shows where the planets are positioned in the future. For example, dial in the year 2380 and it shows the planets position at that point, accurate down to the second. Is this useful for mission design and such for you guys? :)

http://www.fourmilab.ch/cgi-bin/Solar
Currently playing: real life.

"Paying bills, working, this game called real life is so much fun!" - Said nobody ever.

 

Offline The E

  • He's Ebeneezer Goode
  • 213
  • Nothing personal, just tech support.
    • Steam
    • Twitter
Re: Useful? Take a look, solar-system orbit calculator.
We used these:


If I'm just aching this can't go on
I came from chasing dreams to feel alone
There must be changes, miss to feel strong
I really need lifе to touch me
--Evergrey, Where August Mourns

 

Offline -Sara-

  • 29
Re: Useful? Take a look, solar-system orbit calculator.
Celestia is lovely too. :) I believe you can add a huge amount of asteroids and what not as extra content right?
Currently playing: real life.

"Paying bills, working, this game called real life is so much fun!" - Said nobody ever.

 

Offline General Battuta

  • Poe's Law In Action
  • 214
  • i wonder when my postcount will exceed my iq
Re: Useful? Take a look, solar-system orbit calculator.
Celestia is lovely too. :) I believe you can add a huge amount of asteroids and what not as extra content right?

Probably. I think we secretly made an error with our Celestia calculations very late in development, but I don't think it makes a huge difference.  :nervous:

 

Offline -Sara-

  • 29
Re: Useful? Take a look, solar-system orbit calculator.
You can add stuff to Celestia too, basically all FS2 systems if desired. :)
Currently playing: real life.

"Paying bills, working, this game called real life is so much fun!" - Said nobody ever.

 

Offline General Battuta

  • Poe's Law In Action
  • 214
  • i wonder when my postcount will exceed my iq
Re: Useful? Take a look, solar-system orbit calculator.
You're gonna have Herra repositioning all the stars in the skybox now.  :shaking:

 

Offline redsniper

  • 211
  • Aim for the Top!
Re: Useful? Take a look, solar-system orbit calculator.
and that's okay :D
"Think about nice things not unhappy things.
The future makes happy, if you make it yourself.
No war; think about happy things."   -WouterSmitssm

Hard Light Productions:
"...this conversation is pointlessly confrontational."

 

Offline Herra Tohtori

  • The Academic
  • 211
  • Bad command or file name
Re: Useful? Take a look, solar-system orbit calculator.
You're gonna have Herra repositioning all the stars in the skybox now.  :shaking:

Actually, as long as Polaris is over North Pole in Earth skybox, the rest hardly matters anyway. I've put the Sun approximately in the correct location for Autumnal equinox but I really couldn't bother syncing the lighting direction, sun position, Earth's rotation and other crap like that exactly to a certain date, even less with the other planets.

And with other planets, as long as the northern hemisphere is "up" it'll satisfy the immediate "recognition" of the hardiest astronomer, they won't remember where this and that planet is supposed to be in month M of year Y. Even I'm not that nuts. :p


If there were a practical way to export the positions of planetary bodies at certain date from Celestia to Blender, I'd do that but damn if I'm going to bother doing that by hand. Most of the time I go with a setup where I can keep the Sun at XY plane at FRED because quite frankly converting Blender's lighting direction coordinates into FRED's is a huge pain in the arse anyway...

The planet map is more useful as a situational guide, I reckon. General distances between places during our conflict, travel times and other miscellaneous stuff like that.
There are three things that last forever: Abort, Retry, Fail - and the greatest of these is Fail.

 
Re: Useful? Take a look, solar-system orbit calculator.
Quote from: Herra Tohtori
And with other planets, as long as the northern hemisphere is "up" it'll satisfy the immediate "recognition" of the hardiest astronomer, they won't remember where this and that planet is supposed to be in month M of year Y. Even I'm not that nuts.
Yup, this professional astronomer in training  (and avid amateur) didn't notice at all.  Hell, I was ecstatic you guys even bothered to reproduce the sky as seen from Earth anyway.  That said, I do have a couple of nitpicks, but they're common to all current FS2 skyboxes:

1.  There are far, far too many stars visible.  The limiting magnitude of a skilled amateur astronomer is about 7.5 from a completely dark site; since the atmosphere is almost entirely transparent to visible light, this holds approximately in space (you gain maybe a half magnitude or so, mostly due to being above any potential haze, and the removal of residual skyglow).  The skybox in BP has a NELM of about 14-15 instead; the most striking thing is the Milky Way, which should look like a silvery cloud instead of being resolved into stars.  Of course, this all assumes there is no ambient light whatever, which brings me to my second point.

2.  From the cockpit of a space fighter, you should see hardly any stars, except the local star.  All of your cockpit gauges are going to extremely bright compared to the stars outside (I can give numbers if you want), not to mention the brightness of weapons and beam fire, and ship engines.  If you wanted to be really realistic, the skybox would be almost completely featureless and black, except for zeroth and first magnitude stars, plus the local star and any nearby planets.  For the kind of effect you would get, go downtown at night and look up.  I'll bet you can maybe five stars, maximum unless you live in the middle of nowhere.

 

Offline Herra Tohtori

  • The Academic
  • 211
  • Bad command or file name
Re: Useful? Take a look, solar-system orbit calculator.
You know that, and I know that, but it'd be boring and look worse (with reservation). Basically we wanted very much for the milky way to be visible, and cutting the gamma too low would adversely affect that because basically the milky way in the texture consists of thousands (probably millions) of dim stars. In case you're interested as to where we got the source, it's the highest resolution version from here with slight edits done to the levels.

It's true that the current starfields are a tad too saturated with the dim stars, but this is also very much dependant on monitor calibrations. Most LCD monitors by default will have brighter gamma and thus show much more stars than intended, but I don't have control over that and can only work on how it looks on a well-calibrated display.

By the way, on video capture the starfield somehow looks a bit darker. See here: WiH commentary playthrough act I
There are three things that last forever: Abort, Retry, Fail - and the greatest of these is Fail.

 

Offline General Battuta

  • Poe's Law In Action
  • 214
  • i wonder when my postcount will exceed my iq
Re: Useful? Take a look, solar-system orbit calculator.
I'm sort of interested in bringing back background nebulae as I thought they did a lot for the atmosphere. I'm just not sure how to pull it off in a non-irritating fashion.

 

Offline -Sara-

  • 29
Re: Useful? Take a look, solar-system orbit calculator.
Gameplay comes first, or so I learned. While we do not see nebulae with the bare eye, a subtle nebula behind say Jupiter can often add a lot of ambiance to the area we see in space. Similar goes for the excisting full starmaps.
Currently playing: real life.

"Paying bills, working, this game called real life is so much fun!" - Said nobody ever.

 

Offline Herra Tohtori

  • The Academic
  • 211
  • Bad command or file name
Re: Useful? Take a look, solar-system orbit calculator.
I'm sort of interested in bringing back background nebulae as I thought they did a lot for the atmosphere. I'm just not sure how to pull it off in a non-irritating fashion.

Well it could be done by masking the brightest stars out of the starfield, then rendering a blurred and maybe slightly oversaturated version of it... I can experiment with that too.

But, quite frankly, I'm not holding my breath on that looking especially...
* Herra Tohtori puts on shades

...stellar. :p
There are three things that last forever: Abort, Retry, Fail - and the greatest of these is Fail.

 
Re: Useful? Take a look, solar-system orbit calculator.
You know that, and I know that, but it'd be boring and look worse (with reservation). Basically we wanted very much for the milky way to be visible, and cutting the gamma too low would adversely affect that because basically the milky way in the texture consists of thousands (probably millions) of dim stars. In case you're interested as to where we got the source, it's the highest resolution version from here with slight edits done to the levels.

It's true that the current starfields are a tad too saturated with the dim stars, but this is also very much dependant on monitor calibrations. Most LCD monitors by default will have brighter gamma and thus show much more stars than intended, but I don't have control over that and can only work on how it looks on a well-calibrated display.

By the way, on video capture the starfield somehow looks a bit darker. See here: WiH commentary playthrough act I
Yeah, I know, gameplay comes first, and the effect is pretty cool; I just happen to have become more of a realism whore over the years - being an astrophysics major does that to you.  My monitor is actually calibrated (I bought an IPS panel specifically for the accurate color), and it still looks like a mag 14 (maybe 12 or 13; with that many stars, it's hard to give a really accurate estimate) or so starfield to me.  I do still applaud your attention to detail (even going over the starfields carefully in Celestia, I can find only minor positioning errors; for example, in the opening cutscene mission, Jupiter should be in Capricornus or Sagittarius, not Aquila).

In my opinion, an almost totally black skybox enhances the loneliness and grandeur of space, and makes being near a planet like Saturn even more awe-inspiring (well, when I'm not busy getting shot at).  But I think I'm in the minority on that one.  Hell, I found 2001 incredibly tense in the third fourth or so (after HAL tries to kill everyone).

 

Offline Herra Tohtori

  • The Academic
  • 211
  • Bad command or file name
Re: Useful? Take a look, solar-system orbit calculator.
From the description:

The stars are plotted as gaussian point-spread functions (PSF) so the size and amplitude of the stars corresponds to their relative intensity. The stars are also elongated in Right Ascension (celestial longitude) based on declination (celestial latitude) so stars in the polar regions will still be round when projected on a sphere. Stars fainter than the threshold magnitude, usually selected as 5th magnitude, have their magnitude-intensity curve adjusted so they appear brighter than they really are. This makes the band of the Milky Way more visible. Stellar colors are assigned based on B and V magnitudes (B and V are stellar magnitudes measured through different filters). If Johnson B and V magnitudes are unavailable, Tycho B and V magnitudes are used instead. From these, an effective stellar temperature is derived using the algorithms described in Flower (ApJ 469, 355 1996). Corrections were noted from Siobahn Morgan (UNI). The effective temperature was then converted to CIE tristimulus X,Y,Z triples assuming a black-body emission distribution. The X,Y,Z values are then converted to red-green-blue color pixels. About 2.4 million stars are plotted, but many may be below the pixel intensity resolution. The three most conspicuously missing objects on these maps are the Andromeda galaxy (M31) and the two Magellanic Clouds.

So there you have it - stars below visual threshold have been increased in brightness. I guess I could experiment on reversing this by reducing the brightness of the stars under threshold magnitude but the problem is I don't know what the exact threshold would be in the image...

The actual problem here is that individual stars can't be dimensionless dots as they appear in the sky; the smallest graphics element you can have with computers is a pixel (well technically sub-pixels count too but not with images), which always has an area of width x height. Additionally since you only get 255 different values of non-black intensity (depending on monitor the amount might be significantly less actually), we are sadly limited in the form of the art.

Basically what this means is that the small stars will appear brighter in relation to the brightest - which are, sadly, increased in apparent diameter to enhance apparent magnitude in the NASA's source texture; why they did this is beyond me, as I would prefer a source where each star is as small as it can be, and simply has certain RGBI values. Then I could use the alpha channel as the intensity source and the RGB to make a halo for the bright stars to give them a FAINT colour, since colours on computer screens are always dimmer than full white even though in stars this is the opposite - brightest stars have the most detectable colour, if any.

I would actually love to get into Python coding and make a plugin for Blender that can import star catalogues (like Tycho/Hipparcos stuff) and render the stars with this kind of setting. But alas I lack the time and programming skills for this undertaking; from mathematical basis it's not even that complicated I think, but getting it into a code form that works and runs is a different matter. :blah:
There are three things that last forever: Abort, Retry, Fail - and the greatest of these is Fail.

 
Re: Useful? Take a look, solar-system orbit calculator.
Actually what NASA did makes sense from a human visual perspective; if you head out to a really dark site (I've done this several times, so I speak from experience here), it gets very difficult to distinguish the constellations, especially those with few first and second mag stars.  In effect, the fainter stars really do look almost as bright as the bright ones, and the figures simply get drowned in a sea of stars.  In this case, however, the effective threshold magnitude is just set too low.

Also, if you have good eyes (or good corrective lenses, in my case), star images, while small, will be Gaussians (though small ones), not points, because the PSFs of the atmosphere and your eye are not perfect (they aren't for any optical system except pure vacuum).  The 255 values is also not a problem; the eye can distinguish only about .3 mag, so a mag 2.0 and a mag 2.2 star look identical at night.  255 steps is more than enough to cover this, even when reduced by a good fraction.  With a good monitor that can distinguishably display all 256 values, you end up with about 7 steps per magnitude, counting the Sun at mag -27 as 255 and mag 8.0 as 0.  Even if you have a crap monitor that displays only half of these gradations, we end up with 3.5 steps per magnitude, just enough to account for the entire human visual range.

And yeah on the programming; that sounds like a finicky and time consuming undertaking, even though math wise, it's not complicated at all.  Just assign a RGB value to each magnitude below a certain threshold, and fiddle with it a bit depending on spectrum for the brighter stars, then apply a Gaussian to it.  Doing so avoids the twinkling effect when single pixel objects move around on screen.

 

Offline Herra Tohtori

  • The Academic
  • 211
  • Bad command or file name
Re: Useful? Take a look, solar-system orbit calculator.
Quote
...math wise, it's not complicated at all.  Just assign a RGB value to each magnitude below a certain threshold, and fiddle with it a bit depending on spectrum for the brighter stars, then apply a Gaussian to it.  Doing so avoids the twinkling effect when single pixel objects move around on screen.



 ;7

Yeah that's a fair point about optical systems introducing some distortions to images, but I have a feeling a single pixel is still a bit too large to accurately represent a visual perception of a star, especially in vacuum where there's no atmosphere (just the cockpit canopy, helmet visor and your eyes) between the star and you. In the texture rendered by SVS, the truly bright stars are actually pretty gargantuan in diameter - Alpha Canis Majoris is the worst offender, but there are other similarly annoying appearances where stars really look like disks instead of brilliant dots.

Then there's the consideration of scaling the starfield up when playing at low field of view. A gaussian blurred star will appear as blob, while a single pixel will scale up significantly better (I did a lot of testing for the MediaVP starfield and how it looks at different FOV settings), and came to conclusion that stars look the most like stars when most of them are single pixels with variable intensity, and any colour they have should be done with a thin coloured halo over the brightest stars, which gives a surprising illusion of the center dot having colouration too even if it doesn't.

Here's basically how the MediaVP starfield looks like outside FS2_Open:



The texture is saved as uncompressed DDS in the MediaVP's, but unlike the more traditiona 8-bit per channel resulting in 24-bit RGB images, it uses u555 which uses five bits per channel, resulting in 15-bit images. For the reasons you laid out, this is not a significant enough colour depth reduction to change the visual quality in any way - there's still plenty of intensity values to use for the stars, and since there are no gradients, banding is not a problem either.

And since the topic came up, I actually did a larger, technically better starfield using same principles. I used this (if I recall correctly) to render the starfield background for the Age of Aquarius nebula skybox, and it will likely come useful in the future.



...incidentally I use this as my desktop background.

Basically there is the basic layer of grayscale stars which are all single pixels, and then I thresholded the brightest of them, blurred them at 1 pixel gaussian (or 0.5? can't remember), applied level correction to the blurred layer to correct the brightness range. Then I took a plasma render layer, removed all other colours except blue, red and yellow (to get the colours typically produced by spectral radiance, even though yellow stars are actually seen as white), mixed it a bit with another cloud layer at Saturation mode (to give me varying saturations of the halos), applied the colouration to the blobs of the bright stars, then overlayed the starfield on top of that, fiddled with the levels of the colour layer until the end result worked for me.



...Look what you did. Now you got me talking about starfields... :nervous:


Regarding the position of Jupiter, I wanted it to have a dual shadow transit on process (shadows of Europa and Io, the two visible moons in the scene) and for that I needed the moons to be between the Sun and the Jupiter, and by extension since camera was positioned near Europa, camera is also almost directly between the sun and Jupiter.

In hindsight the Europa is a tad bit large for the quality of the textures - they don't quite hold up to close scrutiny, but it's still cool.

I guess no artists is never completely happy with their works and can see flaws that most wouldn't notice.
There are three things that last forever: Abort, Retry, Fail - and the greatest of these is Fail.

 
Re: Useful? Take a look, solar-system orbit calculator.
Quote
*snip xkcd*
Yeah, yeah, yeah  :P.  Hey, the math isn't complicated, the coding and implementation is.  And I know that would be a ***** and a half.

Quote
Yeah that's a fair point about optical systems introducing some distortions to images, but I have a feeling a single pixel is still a bit too large to accurately represent a visual perception of a star, especially in vacuum where there's no atmosphere (just the cockpit canopy, helmet visor and your eyes) between the star and you.
It's not.  If anything it's too small!  Even in perfectly steady skies, the brighter ones look slightly bigger due to the way the eye works, and stars don't look like pixel-like pinpoints (well, to me anyway).  Still nowhere near the way the worst offenders in game like Sirius and Canopus look, unfortunately (when I went back to check the field in Delenda Est, at first I thought Sirius was a moon...), and I understand your point with how well the stars scale with FOV and stuff like that, so I guess it's the best that can be done under the circumstances.

Quote
(to get the colours typically produced by spectral radiance, even though yellow stars are actually seen as white)
Yellow stars are yellow (cases in point:  Algieba, Castor, etc.).  Green stars like the Sun appear white.

Quote
Regarding the position of Jupiter, I wanted it to have a dual shadow transit on process (shadows of Europa and Io, the two visible moons in the scene) and for that I needed the moons to be between the Sun and the Jupiter, and by extension since camera was positioned near Europa, camera is also almost directly between the sun and Jupiter.
I didn't realize you were going for that.  That's a really cool backdrop effect, now that I think about it.  Even so, from the equatorial plane of Europa and Jupiter, since that coincides with the ecliptic, Jupiter will never appear to be in Aquila, since it is not part of the zodiac (the only exception to this is that the ecliptic passes through the feet of Ophiuchus).  Also, the moon shadows should appear on the north equatorial belt, not the equatorial zone, but this is getting into the nitpickiest of nits to pick.

Quote
I guess no artists is never completely happy with their works and can see flaws that most wouldn't notice.
Hey, you and the rest of the team did an awesome job.  The only reasons I noticed half this stuff is, a) I'm an astronomy nerd, and b) after this conversation started, I went back and put some of the star fields to close scrutiny.  Oh well, at least I now know almost exactly when the Nelson gets its ass kicked... (should be Sep 25, 2387, around 19:30 UT).

 

Offline Herra Tohtori

  • The Academic
  • 211
  • Bad command or file name
Re: Useful? Take a look, solar-system orbit calculator.
If anything it's too small!  Even in perfectly steady skies, the brighter ones look slightly bigger due to the way the eye works, and stars don't look like pixel-like pinpoints (well, to me anyway).  Still nowhere near the way the worst offenders in game like Sirius and Canopus look, unfortunately (when I went back to check the field in Delenda Est, at first I thought Sirius was a moon...), and I understand your point with how well the stars scale with FOV and stuff like that, so I guess it's the best that can be done under the circumstances.

Well if you have a 45 degree view angle on a monitor that has horizontal resolution of 1920 pixels, that means each pixel has apparent diameter size of 1.4 arc minutes. A human with 20/20 visual aquity can determine 1 arc minute visual angles. The actual apparent diameter of most stars is pretty damn close to zero, but the said optical inaccuracies and blurring as well as human eyes' limitations mean they are, under ideal conditions, perceived as about 1 arc minute wide dots. Stars bright enough to cause diffraction spikes and other stuff in the eye may be perceived as larger; however, the human eye does this also when you are looking at the monitor!

That means that if you simulate the blurring on the image shown on monitor, and then you look at it with your eyes, you get an exaggerated effect.

I still maintain my view that it's better to use single pixels for individual stars on the texture, at least those that aren't the brightest ones - and you'll notice that for those, the coloured halo in the starfield images I posted does increase the apparent size a little, but it has a more important effect on the perceived colour of the star.

EDIT: Note that this only applies when the texture is shown on the display at exact 1:1 ratio, which is rarely the case. It could be shown smaller (in which case mip mapping is required to prevent shimmering) or it could be upscaled, in which case the pixels are stretched, which means that from technical perspective non-point stars are problematic - when upscaled, they become blurry blobs while pixels retain more or less sharp nature - to certain extent.


Quote
Quote
(to get the colours typically produced by spectral radiance, even though yellow stars are actually seen as white)
Yellow stars are yellow (cases in point:  Algieba, Castor, etc.).  Green stars like the Sun appear white.

Hmm? Sun is a yellow star. Human biometrics recognize the black body radiation of 5800 K as white light (even though it has uneven spread of wavelengths). The peak radiance of sunlight is at the green spectrum, but it's still called a yellow star as far as I know (and its light is perceived as white). No black body radiation of any temperature is perceived as green in colour, it goes from black to red to white to white-blue.

I have never heard of a stellar classification that would even have a category called green stars... :nervous:

Quote
Regarding the position of Jupiter, I wanted it to have a dual shadow transit on process (shadows of Europa and Io, the two visible moons in the scene) and for that I needed the moons to be between the Sun and the Jupiter, and by extension since camera was positioned near Europa, camera is also almost directly between the sun and Jupiter.
I didn't realize you were going for that.  That's a really cool backdrop effect, now that I think about it.  Even so, from the equatorial plane of Europa and Jupiter, since that coincides with the ecliptic, Jupiter will never appear to be in Aquila, since it is not part of the zodiac (the only exception to this is that the ecliptic passes through the feet of Ophiuchus).  Also, the moon shadows should appear on the north equatorial belt, not the equatorial zone, but this is getting into the nitpickiest of nits to pick.[/quote]


Yep, I really didn't want to bother with finding out how to lock objects on certain movement paths around another object (moons around jupiter) so most likely they are in a position they could never exist in. I did check that dual shadow transits happen on Jupiter, and that was enough for me to justify the position I used.

Quote
Quote
I guess no artists is never completely happy with their works and can see flaws that most wouldn't notice.
Hey, you and the rest of the team did an awesome job.  The only reasons I noticed half this stuff is, a) I'm an astronomy nerd, and b) after this conversation started, I went back and put some of the star fields to close scrutiny.  Oh well, at least I now know almost exactly when the Nelson gets its ass kicked... (should be Sep 25, 2387, around 19:30 UT).

Thanks! :)
« Last Edit: September 18, 2010, 02:32:14 pm by Herra Tohtori »
There are three things that last forever: Abort, Retry, Fail - and the greatest of these is Fail.

 
Re: Useful? Take a look, solar-system orbit calculator.
Quote
*snip graphical talk*
Yeah, that makes sense.  I knew about the 1' resolution limit (incidentally, if our eyes had one arc second resolution instead, it would be possible to see most stars in the daytime), but I didn't apply it to our discussion.

Quote
Hmm? Sun is a yellow star. Human biometrics recognize the black body radiation of 5800 K as white light (even though it has uneven spread of wavelengths). The peak radiance of sunlight is at the green spectrum, but it's still called a yellow star as far as I know (and its light is perceived as white). No black body radiation of any temperature is perceived as green in colour, it goes from black to red to white to white-blue.

I have never heard of a stellar classification that would even have a category called green stars...
I know.  The Sun, however, is a G2 star, which is white, not yellow.  No astronomer I know would call early G stars yellow; at most, they would call it yellow-white*.  I also know how blackbody radiation works and how it is perceived by humans.  This just makes the case of Zubeneschamali (Beta Librae) all the more weird, because it does appear green to the eye, despite being a B8 blue dwarf.  I called the Sun a green star in my post to emphasize where its blackbody peak occurs.

*The exception occurs when talking about the size of the Sun; then, it's called a yellow dwarf to distinguish it from white dwarves (the name was already taken).

Quote
Yep, I really didn't want to bother with finding out how to lock objects on certain movement paths around another object (moons around Jupiter) so most likely they are in a position they could never exist in. I did check that dual shadow transits happen on Jupiter, and that was enough for me to justify the position I used.
They are in fact in a possible position, and the relative shadow geometry and all that is in fact correct for Sep 25 2387 around 19:30 UT.  However, what ****s it up is the position of the Sun; it should be a smidgen or so farther south than it is, due to Jupiter's orbital inclination.  Remember what I said about nitpickiest of nits? (I've attached proof from Celestia)

[attachment deleted by admin]