Stellar Ideas

[Oblivion]

Just wondering. Can the stars in VS have different colors?

And has any step been taken in this direction? Is it possible? How?

And since planets behave like planets, do stars behave like stars in VS?
I mean, I noticed systems with binary stars, do these stars actually orbit each other?

Just throwing the wrench again on the gears.

...this should be on the feature request..

[klauss]

No, they don't behave like stars. They have static coordinates.

In any case, star orbits are so slow that it make no difference, does it?
Except maybe binary systems.

[Zeog]

Just wondering. Can the stars in VS have different colors?

Yes, stars could have different colors. There are different colored textures for them (directory textures/stars), even green ones. There is no such thing as a green star, though.

[Halleck]

No, since stars emit radiation pretty much across way too much of the electromagnetic spectrum.

[Paslowo]

Well there is such thing as a colored star however.. The main center emittions are usually white while colors varies around the star.

[klauss]

This is what a star's color is all about.

Usually, the visible color of a star more a function of the atmosphere (far strongher coloring than anything else).

So... in space, you'd see the perceived color, which is the actual color + redshift. Orbiting a star, redshift would be negligible. Still... stars do have a color, which doesn't mean they only emit red/green/whatever light, but that they emit more of that than other wavelengths. It depends on the temperature of the surface, as the wikipedia says - not the core.

[Oblivion]

Funny, I never noticed, another colored star. But then again, I've never much set foot outside cephid17. Too many bugs outside the "safe zone". I once encountered an unbelievably huge planet, and then a station with weird bubble-like opaque shields. needless to say, it was unpleasantly disorienting. And of course, the one time I tried exploring other systems, i ended up in a backwoods rlaan system and was set on by a bunch of aerans.

[Halleck]

By the way, when did stars get ugly?

I took this screenshot with my SVN copy:

[klauss]

I bet I know where the problem is, but I checked the code once and I couldn't see the error (which is obviously there).

But there must be something wrong with the billboard's sort order somewhere.

[Oblivion]

I thought they looked that way before. And those are the stars I've seen since I played VS. Maybe they're going supernova!!

[Zeog]

[the star's color] depends on the temperature of the surface[...]

I have coded some tools that convert a spectrum to an rgb color. With a black body spectrum (I've implemented that as well) I can map a star's effective temperature (of its surface) to its color.

Now, is there some interest here that this nice piece of realism is implemented into Vega Strike?

The clear drawback is that you wont see green stars anymore... (aww).
Furthermore, I know that stars in Vega Strike currently don't have such things as temperatures. But star texturing would then only require a brightness texture for spots etc.

Any opinions?

[klauss]

I don't think that much realism is needed.
If green stars don't look green, I bet that would only confuse people. Let them be green... even if they're not that green in reality.

[rockstar]

Although I agree with klauss that there's no need of too much realism I have to admint Zeog work there sounds interesting. I personally do not care about what color a star has... might be pink if you like it... Zeog is not talking about an idea, but about something he already finished or at least developed to a certain extend. So to answer Zoeg's question: Yes, I would like to see this feature implemented.

[klauss]

But Zeog's second part says: Are you sure, though, it's a wanted feature?
Because although red stars show a relative abundance of red-ish emissiones, to the naked eye they all look mostly white. That's what he meant, AFAIK.

[Oblivion]

Because although red stars show a relative abundance of red-ish emissiones, to the naked eye they all look mostly white. That's what he meant, AFAIK.

a)So basically all stars will have a new variable for temperature? And then convert them to the color we see of them ingame?

b)Wouldn't tehre be problems with the random universe seeder with that?

c)Does it look nice/different?

[Zeog]

Huh? Sorry, I didn't meant to be so cryptic...
So, yes, this little project is finished. Although it was never a feature request I would like to see this feature in VS. But I cannot speak for the entire VS community and I don't want to force this into VS. I figured it isn't one of these aspects of realism that destroys game play, so why not including some realism?

Because although red stars show a relative abundance of red-ish emissiones, to the naked eye they all look mostly white.

Not really (see link below). Here is a (very) short explanation: Everything with a temperature emmits radiation. If the temperature is high enough, there is significat much radiation in the visible part of the electro-magnetic spectrum. For example, a common light bulb emmits most of its radiation in the infrared part of the spectrum (aka heat radiation) and a little in the red and yellow part but very little in the green and blue part of the visible spectrum. Compared to day light it is a little reddish/yellow. But you don't recognize this as your brain is correcting the colors.
Now stars are just balls of hot matter. They do have very different temperatures and they really do have different colors! Red giants for instance are realitvely cool and most of the radiation is emmitted in the red part of the visible spectrum. Of course they also emmit blue light but that is a few orders of magnitude dimmer than the red light, hence the star does appear red to our eyes. Whereas really really hot stars, e.g., blue giants, radiate some orders of magnitute more blue light than red and therefore appear blue.
This is what I did: I took a physical model that descibes how much light of a given frequency is radiated from a body with a given temperature. Then I looked at how much the individual cones (color receptors in you eye, they come in three different types approximatedly one for each of the colors red, green and blue) are excited by this light and from that excitation I derived the color in RGB format.

@Oblivion:
a) Yes, stars then should all have a temperature (effective temperature that is, as light is emmitted from the surface, only surface temperature is relevant). Every temperature implies a color and this is then used to color the stars. This is what I proposed. (That also provides some neat data to print in your hud scanner window...)
b) There should not be any problems with the "random" universe. It's not that random after all. Star properties such as position in the universe, brightness, number and types of its planets are already defined in the milky_way.xml file.
c) I'm not sure if it looks nice to you. There then certainly won't be green stars. Jugde by yourselves. This guy did the same trick (temperatures are the left most column, our sun's temperature is somewhere around 6500 Kelvin): http://www.vendian.org/mncharity/dir3/b ... color.html

[rockstar]

Maybe we should give it a try?! I guess noone can say what it will look like until someone has checked out the results ingame...
Is it hard to implement or might it be possible to create a quick test version for VS?

Anyways, I still say YES Having the star's temperature output displayed in my HUD sounds like nice eye candy... although it's an absolut redundant feature

[rigelan]

Zeog: What part of the color of the star would it change?. Would it change the surface color? What about the color you see radiating out from it? What about having sunspots or surface textures on top of it?

Noting that I'm not too fond of the sunspots anyway. (I would think that if you were looking at a star directly, the brightness would pretty much overpower your ability to see sunspots. And it would be even worse out of your periphery, knowing that they are a bit more sensitive to light.)

I for one, would be excited about seeing more 'normal' star colors, even if it makes it look a bit more bland.

[geoscope]

I like the sunspots.. Gives them character. And the way I see it, the reason we see them is the viewscreens are poloarized to block out most of the radiant light, or we'd all be blind.

[Zeog]

Actually, according to Vega Strike fiction, ships don't really have windows. Everything you see are projections from the ship's sensors. That's why you see these green wire jump points. One can't really see them but the ship's systems augment your view with this information. This way I don't see any problems with sun spots being visible or not. I guess, your ship's computer tones down the brightness enough to not burn your eyes.

As, I said i would use a sun spot texture to modulate the brightness of the star's surface and take the color from the temperature (although that's not very realistic because sun spots are just cooler areas on the surface and should therefore have a different color, due to the lower temperature however they emit less radiation compared to the regions around them and look therefore dark (that would be a cool application for HDRI) ).
For solar flares and such things the color could be different. But I don't think these things are implemented yet. What you currently see "radiating" is just an effect to simulate the squeezing of your eyes due to brightness (or scattering in the lens of you camera as it can't be Mie scattering due to lack of particles) -- just like lens flares are supposed to give a feeling of realism. I'd say we take the star's color to color this as well.

I just found something: stars don't need to have a temperature assigned as they already have a radius and luminosity properties. That's enough to calculate the temperature from. The radii, however, are quite strange: Sol's radius is 16000 [milky_way.xml]. If they were in kilometers that'd be 42x shrinked.

BTW, that temperature to color tool could also be used to adjust a ship's glow map if it's overheated and starts to glow.

[rigelan]

According to wikipedia sunspots

http://en.wikipedia.org/wiki/Sunspots

Our sun in approximately 5700K with 4100K sunspots, corresponding to these colors:



If a star was 10000K with 8000K sunspots it would correspond to these colors:



I guess with the brightness being added by the graphics engine, the sunspots wouldn't seem too out of place. Oh, and I got the colors from the link Zeog produced for us earlier.

[Oblivion]

Have you tested it? How about some screenshots? That would decide if they liek it or not.

For me, I think I'd go for it.

[Zeog]

The colors look good but the brightness adjustment would make them effectively almost black. So, what we need would be some grayscale textures for the sunspots that encode a) the brightness (dark regions for sun spots) and b) therewith the temperature and color information. That corresponds to a semi-procedural generation of star textures. I think those things (texture generation) are not there yet.
@klauss: Could you say something about that? Do you have a toolset of functions that do this job? (Before I start jumping at reinventing the wheel.)

@oblivion: Any screenshots we could give right now would only be mock-ups of pregenerated star textures.

[klauss]

Even with the current engine you can have a grayscale texture for all stars covering the sunspots, and just specify a color for them (computed in whatever way you want).

I hope that's what you were asking.

BTW: I somewhere read they were mostly white. My bad, I guess.

[Oblivion]

Been playing Zeog's Snapshot. And for the first time again in months I enjoyed it.

The number of ships was astounding! And it's nice to be caught in the middle of some heavy beam weapon jousting.

Anyway, Of course, I noticed the stars.

Previous version:


Currently, I notice that the stars look better. Did the settings change?

I notice that the alphaed layer is now rendered inf ront of the mesh. It looks better, but the ray colors don't match sometimes, and sometimes, they don;t show up at all..