- #1
Artlav
- 162
- 1
I'm trying to render the sky as it would appear from a starship moving at some large part of the speed of light.
Geometry was straightforward, but colors are the problem.
How would the doppler effect change the color of a star?
The expectation is that the stars behind are red and the ones ahead are blue.
At first i tried to apply doppler shift ( gamma*(1-(v/c)*cos_ang) ) per-channel to the RGB value of the star color, with rather colorful but odd results of violet stars ahead and red-to-black ones behind.
Then i found out that the black body radiation is doppler shifted to the same extent as if it's temperature was altered by the same doppler factor.
So i tried to use the stars' temperatures as the input, since converting temperature to RGB is much easier than shifting arbitrary RGB.
This, however, produced a rather bland view of slightly red stars behind and slightly blue stars ahead.
Further, applying the intensity correction (doppler factor to the power of 4) made the stars fade behind, and be bright-white ahead, with almost no visible color change.
So, the question is - which of these are more correct?
The last one is the blandest one, but it seems to be the most correct as far as i can tell.
The first one is the most colorful, but it assumes that the stars emit nothing but visible light, while you'd expect that the invisible light would be shifted into our visibility range, and blackbody temperture accounts for that.
Even better, are there any reference images of correctly rendered sky at relativistic velocities? The Google finds many, but most are pop science kind of stuff that doesn't account for most of the effects.
Geometry was straightforward, but colors are the problem.
How would the doppler effect change the color of a star?
The expectation is that the stars behind are red and the ones ahead are blue.
At first i tried to apply doppler shift ( gamma*(1-(v/c)*cos_ang) ) per-channel to the RGB value of the star color, with rather colorful but odd results of violet stars ahead and red-to-black ones behind.
Then i found out that the black body radiation is doppler shifted to the same extent as if it's temperature was altered by the same doppler factor.
So i tried to use the stars' temperatures as the input, since converting temperature to RGB is much easier than shifting arbitrary RGB.
This, however, produced a rather bland view of slightly red stars behind and slightly blue stars ahead.
Further, applying the intensity correction (doppler factor to the power of 4) made the stars fade behind, and be bright-white ahead, with almost no visible color change.
So, the question is - which of these are more correct?
The last one is the blandest one, but it seems to be the most correct as far as i can tell.
The first one is the most colorful, but it assumes that the stars emit nothing but visible light, while you'd expect that the invisible light would be shifted into our visibility range, and blackbody temperture accounts for that.
Even better, are there any reference images of correctly rendered sky at relativistic velocities? The Google finds many, but most are pop science kind of stuff that doesn't account for most of the effects.