Where are all the procedural red supergiants? Where are all the procedural Class C stars?

[madbilly]

To begin with, FD use two different sets of luminosity classes - one for hand-authored real stars, and another for proc-gen stars.

With the proc-gen stars, the luminosity classes use the A -> AB -> B distinction which is normally seen in supergiant spectra for almost all the spectra. Like VA, VAB, VB. A real star would just be called V without distinction. In the case where a hand-authored primary star has proc-gen secondaries, you can see this illustrated, e.g. HIP 72235 has a G5 V primary and an M7 VA secondary.

Luminosity classes are assigned in-game based strictly on luminosity calculated from (T**4) * (R**2). For real stars, they're usually assigned (AIUI) by looking at pressure broadening effects on spectral lines.

Each spectral temperature subclass (G0, G1, G2...) has a different set of boundary values for the different possible luminosity classes. A G0 star of 1 solar luminosity will be called G0 VB. A G1 star of 1 solar luminosity will be called G1 VAB, and so on. Not all luminosity classes are available for each spectral temperature subclass. The overall consequence is a stepped distribution where some stars of lower luminosity can appear to have a brighter luminosity class than stars of higher luminosity...

(I haven't worked out the underlying formula that determines the shape of the steps - I have a lot of empirical data and know most of the boundaries fairly well.)

Some of the luminosity divisions are simply backwards as far as I can tell, like B9 IA stars are less luminous than B9 IAB.

And there are one or two instances where luminosities are forced: all M9 stars are M9 VI, all brown dwarfs are V.

Thanks Jackie! I was just in the process of trying to understand why the distribution of A, AB, A and Z is not even for each spectral class. I'm finding a lot of B0VZ, often paired with A0VZ, F0VZ and even a G0VZ in the Galactic Aphelion. I notice you haven't included Z in your table of empirical data, any reason?

Can luminosity class (what we see in Galmap) be worked out from absolute magnitude? If so, then we might be able to automate data collection because, even though the luminosity class isn't in the commander's log API, the absolute magnitude is in there.

EDIT:
I'm just catching up on your "Decoding Universal Cartographics" thread, lots in there, perhaps explains why you're less interested in Z.

Also I realise that the absolute magnitude in the API is not directly useful, because what I think you're already trying to work out is how FDev have applied the divisions between luminosity classes based on temperature and size.

But since absolute magnitude isn't shown in the Sysmap, perhaps this addition will be useful after all...

....ah brainache!

 

[Jackie Silver]

I'm just catching up on your "Decoding Universal Cartographics" thread, lots in there, perhaps explains why you're less interested in Z.

Yah the VZ in Elite is an indicator of very young stars just out of the protostar state. I'm still a little unclear as to exactly where the boundary lies for it. Are you familiar with Python? I've got a script I've been working on that tries to work out the full Elite spectral class given a star's basic parameters, and that has a function in it where the VZ boundaries are covered as best as I understand them, I can link you that if you like. I don't have an easy to use compiled version of it though!

 

[Pirin]

Maybe I'm just out of the loop on this one, but I've been searching pretty tirelessly to find these two star typed and have come up empty handed. There's no shortage of red giants, and orange giants are scattered about the core, but I've not seen a single red supergiant (e.g. M3 I) that wasn't hand placed. Same goes to class C carbon stars. I've seen class S, MS, CJ, and CN all around, but not a single class C that wasn't hand placed. Is there something I'm missing here, or is this just a really big flaw in E: D galactic proc-gen?

I spent 5+ months searching for exactly the same objects: M I0 and C. No luck. At some point i started a forum thread asking all cmdrs who ever discovered any of those to post pics. There were no replies So i think it is safe to assume they dont exist.

...on other hand one the "missing" giant types was A I0. And then, after some 10+ months of exploring i suddenly found one such star and few days later another one. Along with one which was reported to one of the exploration websites these 3 are the only ones of that type i know about. So who knows? Maybe M I0 and C exist but are extremely rare

 

[[VR] optimal_909]

I spent 5+ months searching for exactly the same objects: M I0 and C. No luck. At some point i started a forum thread asking all cmdrs who ever discovered any of those to post pics. There were no replies So i think it is safe to assume they dont exist.

...on other hand one the "missing" giant types was A I0. And then, after some 10+ months of exploring i suddenly found one such star and few days later another one. Along with one which was reported to one of the exploration websites these 3 are the only ones of that type i know about. So who knows? Maybe M I0 and C exist but are extremely rare

I've found a couple of AI, with about 50 solar radii - those were the largest stars I've discovered.

 

[Pirin]

I've found a couple of AI, with about 50 solar radii - those were the largest stars I've discovered.

Yes, A I are rare but still relatively easy to find. I am talking about A I0 supergiants. They are much larger. The star below is actually the largest procedure generated object i know about

Spoiler

 

[madbilly]

Yah the VZ in Elite is an indicator of very young stars just out of the protostar state. I'm still a little unclear as to exactly where the boundary lies for it. Are you familiar with Python? I've got a script I've been working on that tries to work out the full Elite spectral class given a star's basic parameters, and that has a function in it where the VZ boundaries are covered as best as I understand them, I can link you that if you like. I don't have an easy to use compiled version of it though!

I caught up with the main star class part of your decoding thread now (though I've not had time to follow the naming part yet). So I now understand, Z follows protostars. Makes me wonder about going back a few klys to see what was special where all those B0VZs were.

Python - an appropriate language for Elite! It's something I should have learnt a while ago, maybe this will motivate me! Yes please, send me a link and I'll see if I can work out how to use it. Cheers

 

[Jackie Silver]

Python - an appropriate language for Elite! It's something I should have learnt a while ago, maybe this will motivate me! Yes please, send me a link and I'll see if I can work out how to use it. Cheers

Hah, yes, it is appropriate. I love Python, it's the language that brought me back to programming. Here's an archive with the star class guesser and the data files it needs. It's a bit of a mess, and don't take it as gospel, but it gets the vast majority of proc-gen stars right. By default, if you run the scg8x.py it will run through every star in starlist13.csv and try to guess each one - it's only set to report ones it fails on though, which is mostly things like carbon stars where it can't distinguish between the different types properly. It will also generate an HR diagram of the stars and an age-mass diagram. The format of the csv files is kinda arcane but if you follow the pattern you can add stuff to it. The "As You Like It" spreadsheet is the base from which the starlist was created, along with a catlist csv which has the hand-authored stars.

 

[CMDR AKIRA MASAKARI]

Just stumbled in an example of a procedural star that got a wrong description:

In the gal map it's described as: B6 IVA. It should be out of main sequence, but you are never sure. In Gal Map visually it seems a normal B class.

Then you jump in and bang: 396 Solar Radii.
But still the Sys Map description has the B Main Sequence Template.
With procedural giants many times is a guessing and you can tell only after having jumped in.

Spoiler

 

[thundernuns]

Just stumbled in an example of a procedural star that got a wrong description:

In the gal map it's described as: B6 IVA. It should be out of main sequence, but you are never sure. In Gal Map visually it seems a normal B class.

Then you jump in and bang: 396 Solar Radii.
But still the Sys Map description has the B Main Sequence Template.
With procedural giants many times is a guessing and you can tell only after having jumped in.

Spoiler

http://i.imgur.com/kNIL8s9.jpg

"They usually range in mass from 2 to 16 solar masses" 71.8 solar masses lol

 

[Pirin]

Just stumbled in an example of a procedural star that got a wrong description:

In the gal map it's described as: B6 IVA. It should be out of main sequence, but you are never sure. In Gal Map visually it seems a normal B class.

Then you jump in and bang: 396 Solar Radii.
But still the Sys Map description has the B Main Sequence Template.
With procedural giants many times is a guessing and you can tell only after having jumped in.

Spoiler

http://i.imgur.com/kNIL8s9.jpg

IV indicates this is a subgiant star so it makes sense.

 

[CMDR AKIRA MASAKARI]

IV indicates this is a subgiant star so it makes sense.

Yes, IV is subgiant, and age: <1 million years means that star recently exited main sequence.

But a radius of 396 solar radii is a Supergiant dimension.

Comparisons with real world are difficult, as a star that exits main sequence is very unpredictable. But for a gameplay point of wiev finding an object that has supergiant dimensions and a regular B Main Sequence description and an out of Sequence number that should be of subgiants is a little confusing and I sense some starforge oddity here.

 

[Andrew Gaspurr]

I am mostly a 'massive young star' explorer type and in my travels have seen many O and B type stars exhibiting giant star proportions. What always bogged me were stars denominated B0 III or B7 III (or O0 III/O7 III). Sometimes they were IIIab/IIab, but upon survey had rather normal proportions, i.e. they were just well inside their normal O or B parameters.

Are there any explanations for this? And yes, please elaborate...

 

[madbilly]

I think Jackie already worked out how this works in ED so I'm sure he'll weigh in soon with the correct equation. I'm an amateur at this, so bear with me if my explanation doesn't add up and if you know better please give your explanation.

I, II, III etc is actually a rating of luminosity, not size, but I'm sure you already know this very well. Luminosity is (roughly) a function of size and temperature (I know it's actually more about radiated energy, however this is generally proportional to temperature). So, it's totally possible (within the same spectral classification) to have a really hot small star and a really huge cool star, both of which have the same luminosity class. The names that we associate with each luminosity class (e.g. III=giant, V=dwarf) are historical and don't really fit the current classification system, at least not in all cases. Then we also have the HR diagrams where we associated certain luminosity classes with certain ages of star starting from different masses and going through different lifecycles depending on starting mass and companion stars, etc.

I know that this is all very well worked out and understood in real-world physics, but there are lots of edge-cases that don't fit the general understanding of the model. And stellar forge is a simplified version of real-world physics. So therefore we are using the luminosity class for more than one thing, and coupled with the name that we assign to each class we can end up with some situations that don't fit these definitions.

[lights touch paper, steps back!]

 

[[VR] optimal_909]

I, II, III etc is actually a rating of luminosity, not size, but I'm sure you already know this very well.

Luminosity should be proportional to the size AFAIK.

 

[madbilly]

Yes as far as I understand luminosity is proportional to size. I, II, III etc are luminosity class not size. Luminosity is a function (proportional) to size and temperature, therefore it's possible to have large cool stars with the same luminosity as small hot stars, and within the same spectral class these might be given the same luminosity class. But then I think that luminosity class is also used to indicate the stage a star is in its lifecycle, even though this is presumably determined from other spectral characteristics and not only from luminosity itself. So I guess that luminosity class doesn't always correlate directly with actual luminosity because the class takes other factors into account. Maybe this means that in ED it's not quite right, and giants (class III) stars (for example) should always be much larger than main sequence stars.

Where's Jackie with his equation, I'm sure it's (R**2)/(T**4) or something.

 

[Andrew Gaspurr]

Try this:

Stellar Luminosity Calculator

 

[Jackie Silver]

Where's Jackie with his equation, I'm sure it's (R**2)/(T**4) or something.

Her. I'm starting to think I should petition FD to change my CMDR name to Jacqui-with-a-q. Luminosity goes as (R**2) * (T**4) so it's much more dependent on temperature than on size (as you noted above!) - again as you note the "real" luminosity classes are assigned based on other factors. It's one of those things where a guess at one thing is used to guess at another thing. In the game we have exact information on stars because we can park next to them and scan them; we know how far they are away, we know how big they are and how hot they are so we know their luminosity exactly, whereas from Earth we look up at the sky, see a bright shiny dot and start working from there...

That said, ED's luminosity classes are assigned in a very weird way that I don't properly understand yet. It's almost as if there's a limited pool of potential luminosity classes for each particular subdivision of temperature. On top of that the proc-gen stars themselves fall into a much narrower and more sharply defined range than "real" stars do; a real HR diagram has stars all over the place but Elite's HR diagram has stars following quite neat tracks.

The information I'm collecting at the moment on catalogue stars probably isn't directly relevant to this, but you never know. I had a quick look at graphs and there were lines in it that seemed very neat; the impression I have is that some of the hand-authored stars have their details carefully (^^) entered based on real data, whereas others have been positioned using real data but generated semi-procedurally, so they may fall onto the same generated sequences that all the proc-gen stars do.

(One thing from working with real catalogue data - there's a bewildering array of additional spectral information given for peculiar spectra, emission lines and all that stuff, and of course variable stars may change dramatically from one spectral class to another. Compared against Elite's neat set of spectra it's very messy!)

 

[iain666]

That said, ED's luminosity classes are assigned in a very weird way that I don't properly understand yet.

Perhaps now we have the absolute magnitude available from the journals something might fall out of that.

 

[madbilly]

Try this:

Stellar Luminosity Calculator

Very handy thanks!

Luminosity goes as (R**2) * (T**4)

Thanks Jacqui!

Perhaps now we have the absolute magnitude available from the journals something might fall out of that.

Yes I was wondering something similar, and if the journal could include the star class from the Galmap then we could quickly work out the divisions because data collection could be automated. This might take some of the fun out of it, but there's plenty more fun things to do!

 

[madbilly]

So, was the conclusion that nobody has found a proc-gen M_RedSuperGiant? Or just that they're extremely rare? I've not found one, only MIII and MIV.

And what about F_WhiteSuperGiant, any proc-gen ones in game? I've only found FV and FIV, nothing brighter. Are yellow hypergiants also classed as this? I've only found GV, no other luminosities.

Cheers