Visual Extinction in Celestia?

[Fafers_br]

Hello there!
As everybody knows, Celestia does not take visual extinction (Av) in consideration in its calculations. This has some consequences, and one of them is the underestimation of the AbsMag and, hence, the calculation of the stars' radius when using the black body aproximation.
I'd like to proppose, for future consideration, the creation of a new, optional variable "Extinction" in the stc codes. When ommited, all the Celestia's calculations concerning the magnitudes of the star would continue as they are today. But if one declares it's value, the Av would then enter in the calculations, thus resulting in more realistic values for the AbsMag and the estimated radius of the star.
Take, for example, the star Schulte 16:

Code: Select all

1010293161 "Schulte 16:Cyg OB2 16:LS III +41 33" {
InfoURL "http://simbad.u-strasbg.fr/simbad/sim-id?Ident=schulte+16"
   RA  308.1607096
   Dec  41.4204975
   Distance  4940 #Gaia DR2+Bailer-Jones
   SpectralType "O7.5V"
   AppMag 10.84
}

Without Av, Celestia calculates AbsMag=-0.06 and Radius=1.1Rsun (~782000Km). These data do not match those of a tipical O7.5V star. It's an unrealistic star and it would be too faint for its class when observed from stars close to it.
According to Chentsov et al. (https://arxiv.org/abs/1306.1087), the Av towards this star is 4.5mag. So, if we could declare this Av in the stc code, like this:

Code: Select all

1010293161 "Schulte 16:Cyg OB2 16:LS III +41 33" {
InfoURL "http://simbad.u-strasbg.fr/simbad/sim-id?Ident=schulte+16"
   RA  308.1607096
   Dec  41.4204975
   Distance  4940 #Gaia DR2+Bailer-Jones
   SpectralType "O7.5V"
   AppMag 10.84
   Extinction 4.5
}

We would get Absmag=-4.56 and Radius=8.9Rsun (~6210000Km), much closer to the data of a typical O7.5V star.
Celestia would have to assume isotropic and homogeneous extinction though (still better than no extinction at all, IMHO). So, it would store the Av/pc for that star in order to calculate the AppMag for other distances.

The math behind it:
Let d be the distance of the star from Sol in parsecs. AppMag is its apparent magnitude.
Let Av be the total visual extincion in magnitudes declared in the code. It must be a positive value. The AbsMag of the star is given by:
AbsMag=AppMag-5*(log10(d)-1)-Av [eq.1]

Now, let Avpc=Av/d [eq.2]
This will be the extinction per parsec. We will consider the extinction homogeneous and isotropic, so the value given by [eq.2] will be used in any direction from the star.

So, for any given distance d' (in parsecs) from the star:
AppMag=AbsMag+5*(log10(d')-1)+Avpc*d' [eq.3]

[LukeCEL]

Hey Fafers_br, this is a really good idea. Who knows how many high-mass stars in dusty OB associations could be better rendered if this change were to be added.

Have you thought of a good way to add interstellar reddening as well? As we know, stars appear not just dimmer, but redder when obscured by dust because the dust preferentially scatters blue light.

I was thinking this would be added as an optional "Reddening" parameter similar to the "Extinction" parameter above. I am not sure how the rendering of the colors would work, though. I am not an expert on color, somebody else should chime in.

Best regards
LukeCEL

[Lafuente_Astronomy]

Hey Fafers_br, this is a really good idea. Who knows how many high-mass stars in dusty OB associations could be better rendered if this change were to be added.

Have you thought of a good way to add interstellar reddening as well? As we know, stars appear not just dimmer, but redder when obscured by dust because the dust preferentially scatters blue light.

I was thinking this would be added as an optional "Reddening" parameter similar to the "Extinction" parameter above. I am not sure how the rendering of the colors would work, though. I am not an expert on color, somebody else should chime in.

Best regards
LukeCEL

This will also be very useful to hopefully implement transparent meshes someday, since it would allow the object to be seen but at a lower magnitude, and in relation to the calculations Fafers_br presented above, it can make the making of galaxies, both stars and nebulae within it much, much better

[Fafers_br]

Have you thought of a good way to add interstellar reddening as well? As we know, stars appear not just dimmer, but redder when obscured by dust because the dust preferentially scatters blue light.

I think reddening is more complicated. Well, even my suggestion for extinction is already naively considering it isotropic and homogeneous.

Reddening E(B-V) can be written as:
E(B-V) = (B-V)-(B-V)0. [eq.1]

Where (B-V) is the observed (B-V) color and (B-V)0 is the intrinsic one.

It happens that
E(B-V)=Av/Rv. [eq.2]

Where Rv is the "ratio of total-to-selective extinction". The accepted typical value for Rv in the Milky Way is 3.1, but it varies. In the Carina region, for example, it's value is around 4.

So, if you knew Rv and Av then, using eq. 2, you could find E(B-V).
Then, using this into eq.1, one would find the difference between (B-V) and (B-V)0.
Supposing you know the real spectral class of the star and its real Teff, you would already know the (B-V)0. Then it would be piece of cake! Because:
(B-V)=E(B-V)+(B-V)0
Once you know (B-V), you could find the color. For example, if it's a main sequence star, you could use Mamajek's table to know its temperature and, hence, the color.
But, in reality, we don't know the value of Rv in a particular direction; the relation between (B-V) and Teff varies for different luminosity classes, etc. So, it's not a simple task, I think.
I may be wrong in some of the concepts above. So, I'd be happy to be corrected if that's the case.

[Fafers_br]

As an exercise of pure imagination, if we could abandon the scientific rigour for a while...

We could fix the value of Rv=3.1.

Then, for the same Schulte 16 given as example above.
From Earth:

Code: Select all

d'=d=1514.6pc

The declared value for Av would be 4.5. Then:

Code: Select all

Avpc=Av/d'=0.02971mag/pc.

So:

Code: Select all

E(B-V)=d'*Avpc/Rv=4.5/3.1=1.4516

The star is an O7.5V, then:

Code: Select all

(B-V)0=-0.32 (from Mamajek)

So:

Code: Select all

(B-V)=(B-V)0+E(B-V)=-0.32+1.4516=1.1316

In Mamajek's table, this (B-V) would be equivalent to a Teff~4500K. Celestia could use this temperature to render Schulte 16 as seen from Earth, instead of the color derived from its real temperature. And the color would change as you approach the star, since the value of E(B-V) would also change (because E(B-V)=d'*Avpc/Rv).
But, again, there is a LOT of approximations and simplifications in this idea. It's certainly not scientific at all!
Best regards.

[onetwothree]

https://github.com/CelestiaProject/Celestia/pull/818