Results 1 to 20 of 20

Thread: Star color nomenclature

  1. #1
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    7,656

    Star color nomenclature

    I am splitting this off from my other thread to help keep that thread on my intended topic, which is about the conditions in the cores of stars. This is for the benefit of new members who may not have seen earlier discussions.

    Quote Originally Posted by Hornblower
    In my opinion the color terminology issue is a very minor semantic grievance. What we have here is a modest shift of the commonly used color names toward the red end of the spectrum and not explicitly stating just how pastel these tints are. To me this is far less bothersome than the terms early and late for spectral types, or the practice of referring to all elements heavier than helium as metals. These do not bother me all that much because they are one-word terms for which technically superior substitutes mean more verbal clutter. All of this is jargon that has evolved over modern astronomical history. I consider it a far cry from the major physics blunder of getting the mass/core-density relationship backward for main sequence stars.
    Let me add that in my opinion there is not an absolute criterion from first principles for defining the tint of an incandescent light source at a particular temperature as white. Fresh snow looks just as white under 3200K light at night as it does in daylight, at least to me, because of our capacity for adjusting on the fly to the dominant ambient light. For everyday sensory purposes we can say, "If it looks white, it is white." We can legitimately say that a reflective surface is white by definition if the spectrum of the reflected light is unchanged from that of the direct light from the source.

    Astronomical observing is a game changer. When looking at stars in a dark surrounding, my eye-brain combination appears to have a set point where I see Polaris, with published B-V color index of 0.60, as neutral white. Altair, B-V 0.22, appears cold white, while Vega, B-V 0.00, appears pale bluish white. Spica, B-V -0.23, may be slightly more saturated with pale blue, but I am not certain. (Wouldn't you know, we are having a stubborn streak of cloudy weather, so I am unable to check up as I write this.) In the other direction Capella, B-V 0.80, appears pale yellowish white, while Arcturus (1.23) is a more vivid yellowish orange. Aldebaran (1.54) is still more saturated. For my purposes a useful astronomical definition of white would be the tint of Polaris as seen through or atmosphere, making the Sun pale yellowish white under these conditions. Out in space they would all shift toward the blue end, and my guess is that the Sun will be neutral white if suitably attenuated.

    All of this is after my recent cataract surgery. Before that, everything looked yellower in the telescope. I had the option of either yellow-tinted lenses, to match the ugly brown goop that came out, or clear ones. The doctor knew about my interest in astronomy and said the clear ones would have more total throughput, so that was my choice. This difference is consistent with studies which show that children and young adults tend to see stars as bluer than do old timers such as myself before my surgery.

    The B-V color index is the difference in brightness between normalized exposures on blue-sensitive and panchromatic emulsions. For reasons that made sense to the pioneering astrophotographers a century ago, Vega's tint was chosen as the neutral standard, so the practice of calling a type A0 star white stuck in the profession. That makes a G2 star such as Alpha Centauri A (0.68) yellowish white by comparison.

  2. #2
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by Hornblower View Post
    Let me add that in my opinion there is not an absolute criterion from first principles for defining the tint of an incandescent light source at a particular temperature as white. Fresh snow looks just as white under 3200K light at night as it does in daylight, at least to me, because of our capacity for adjusting on the fly to the dominant ambient light. For everyday sensory purposes we can say, "If it looks white, it is white." We can legitimately say that a reflective surface is white by definition if the spectrum of the reflected light is unchanged from that of the direct light from the source.
    Yes, that great adjusting ability of the retinex (eye-brain coloring) to make the brightest objects appear as a white light is remarkable. Bright near-white objects will be seen as white but the brain will also know to color adjust all that "white" light source resulting in color-correcting all the other objects. [It can be fooled at times however as noted by the works of Dr. Lamb (Polaroid).] I think "white balance" is the term used in image processing.

    The whitening effect of the retinex should cause near-white stars to appear white and not bluish-white or yellowish-white, but only to a certain degree. As you have noted in the past, much to my surprise initially, point-source lights are seen differently than extended objects. So we will need to address both. In the halls of astronomy, no doubt the point source colors are effective as they have come down from as far, or further, from Father Secchi. Only a couple of stars can be seen as extended objects. But, in graphics, when illustrating stars as extended objects, it should be a different story since most of the graphics reach out to the general public and color inaccuracy has some drawbacks even though being artful at times is not all bad. Showing a star as an extended object implies that the color drawn is how it would look, which will often be different than the traditional colors assigned to the spectral class types.

    Specifically (for extended (disk) objects only):
    Blue stars can never appear blue, but the hottest ones will indeed appear bluish-white. [Blue stars emit copious amounts of green, yellow, etc., thus whitening the stronger blue emissions.]

    Green stars don't exist due to the same blending effect since stars are near black-body. [Oddly, the small variation in emissions between stars and blackbody has cause an immense amount of color errors from those who should know better. For instance, the energy peak for the Sun is demonstrably not green, though it would be green if it were a blackbody.]

    White stars will likely apply for a broad range of spectral classes from somewhere in A to K. Color constancy likely has much to do with this. Yet, if we look at the photon distributions within this range, it will likely demonstrate just how flat the optical distribution is. The Sun's photon distribution is nearly flat as a pancake, though it drops off slightly in the blue band which happens, ironically, to be its greatest energy level emissions.

    Yellow stars, I think, would never look a saturated yellow, like a ripe banana or the Sun*wink*, due to the blending effect of the other colors, but they would appear different degrees of yellowish-white or orangish-yellow, but only tints of each.

    Orange and red stars might have a greater color tint since these cooler stars produce lower levels of the yellow, greens and blues. But their surface temperatures for most of these will be hotter than an incandescent bulb, which can be in a range from about 2500K to 3300K. So only the cooler red stars will look orange. Color vision varies and some will likely see more of a reddish tint for these.

    Astronomical observing is a game changer.
    Yep, point-source colors are definitely redshifted a bit.
    When looking at stars in a dark surrounding, my eye-brain combination appears to have a set point where I see Polaris, with published B-V color index of 0.60, as neutral white. Altair, B-V 0.22, appears cold white, while Vega, B-V 0.00, appears pale bluish white.
    When Father Secchi was developing his spectral types, Vega was listed as a white star. Rome’s latitude is ~42 deg. (as is Boston), so perhaps he had cleaner air back then and less air mass to see Vega as white? He associated the too-bright-to see Sun with the yellowish star Capella, and this may have greatly contributed to the Sun’s yellow description ever since. His extensive spectroscopy was originally all visual. When photography came along he switched. Thereafter, the black and white spectrographs essentially moved color off the table. Pickering’s team (Cannon, Murray, etc.) while building the Draper catalog, apparently, noted a likely color correspondence for star temperatures and spectrum, but it was of a secondary interest due to the less precise work in spectral lines themselves. I doubt I could easily find a reference on this, but I recall reading this in one or two places.

    In the other direction Capella, B-V 0.80, appears pale yellowish white, while Arcturus (1.23) is a more vivid yellowish orange.
    Ah Capella! No orchestra is necessary to properly introduce the star that may have made our Sun a yellow star. [Is this not one of my better puns?  Grant might even like it… maybe.] As mentioned above, Secchi placed the Sun’s spectral type in this group of yellow stars.

    Out in space they would all shift toward the blue end, and my guess is that the Sun will be neutral white if suitably attenuated.
    Yes, your posts from a while back convinced me of this fact. The reason for this is still unclear but the following factors are likely contributors:
    1) Air mass. We all know our atmosphere causes all entering light to be scattered far more in the blue end of the spectrum than the other colors. This is why the sky is blue and the Sun (extended) can go from yellow-white near the horizon to yellow or even orange (if particle counts are very high) when setting.
    2) No blue cones in the fovea. Color is determined by the three color cones in the retina, each sensitive to a different part of the visible spectrum, but with overlap for each. The very high acuity is in the central part of the retina known as the fovea. There are essentially no rods and no blue color cones in this tiny central region. But, we can see bluish-white stars (point sources) as bluish-white when looking directly at them, as you have pointed-out. But is there a redshift effect to this to make these less blue and white ones more yellowish-white? Perhaps.
    3) Contrast. The darkness of space can have an effect on how we see color. The color of nearby, especially bright, neighbors will also affect what the brain sees for color.
    4) Our age. [see below] Perhaps Father Secchi’s age also played a roll in his color assignments.
    5) Unsure, but I would bet there’s something that should go here even if it is something in the brain’s algorithms that would favor a redshift for point sources, or at least white points to have a yellow tint.
    Eta Cassiopeia (slightly hotter than the Sun) is a great example of a pure white star if we could see its broad disk and at a properly attenuated level, yet as a point source it will often appear with a strong tint of yellow, or gold for me. Its red dwarf neighbor can appear surprisingly red. It is clear that there is a definite color difference between point-source stars and how they would look if were to get near them and look at them with neutral filter glasses.
    All of this is after my recent cataract surgery. Before that, everything looked yellower in the telescope. I had the option of either yellow-tinted lenses, to match the ugly brown goop that came out, or clear ones. The doctor knew about my interest in astronomy and said the clear ones would have more total throughput, so that was my choice. This difference is consistent with studies which show that children and young adults tend to see stars as bluer than do old timers such as myself before my surgery.
    I may not be long for a similar surgery. My younger brother recently had his. I did an on-line color test by unscrambling 30 or so very similar, but different, color shades and putting them in order of increasing color differences. But perhaps I can do this even with a yellow bias due to my age. Regardless, I have found no “white” color claimed for Eta Cas from the first 6 or so sources. They all state some form of yellow tint (or gold) for it, so your redshift effect for point sources seems to be accurate.
    For reasons that made sense to the pioneering astrophotographers a century ago, Vega's tint was chosen as the neutral standard, so the practice of calling a type A0 star white stuck in the profession. That makes a G2 star such as Alpha Centauri A (0.68) yellowish white by comparison.
    I hope to learn more about how this happened. Danti also was a contributor to star color, I think, and may have had some influence upon Secchi.

    There is a nuance to star colors regarding color change across their disk. The central portion of disks are much hotter than the limb, so there can be a color tint difference for may stars due to this temperature gradient. A G9 star, for instance, might have a distinct yellow-white limb but a white center. The Sun's temperature gradient from limb to central area is 1390K, which is a 27% increase in temperature. For this reason, it became obvious to me that the Sun is very white as an attenuated disk. My avatar's image was taken from the high quality McMath-Pierce solar telescope's white projection table and there is no hint of tint along the limb, thus if we were to add the blues that were scattered by our atmosphere back into this white disk then it's obvious the Sun cannot be remotely a yellow star (as an extended object), though yellowish-white as a point source.
    Last edited by George; 2017-Nov-10 at 05:51 PM.
    We know time flies, we just can't see its wings.

  3. #3
    Join Date
    Feb 2009
    Posts
    2,035
    How much does the true colour index of Rigil Kentaurus AB, unresolved, differ from Sun?

    How perceptible is the colour contrast between Rigil Kentaurus A and B when resolved?

  4. #4
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    7,656
    Very good questions, and perhaps our friends in Australia and New Zealand could take a look and let us know what they see. My educated guess is that B would look noticeably orange alongside A at small separations. B is considerably fainter, and the fainter member of a pair tends to look more vividly colored than it does in isolation.

    When unresolved, A contributes about 3/4 of the total light, so the composite should have about the same color index as a single G4 or G5 star. I think the difference would be barely perceptible, if at all.

  5. #5
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    7,656
    I would not be inclined to try to see a reason for the whiteness of the light in a graph of the the intensity as a function of the wavelength or frequency. We can analyze this in four different ways, using two versions of Planck’s formula for blackbodies:

    Power as a function of wavelength
    Photon flux as a function of wavelength
    Power as a function of frequency
    Photon flux as a function of frequency

    This exercise gives us four very different light curves, none of which are flat over the visible range for a blackbody at the Sun’s effective temperature. The first one is nearly symmetrical, but is substantially elevated in the middle of the range. A source that is really flat would look pale pink alongside sunlight, and we would really look silly if we define it as an absolute standard for white. If we adjust the spectrum to make any of the others flat, the tint will shift toward blue as seen alongside sunlight.

    If we look at photon flux of sunlight, we see fewer photons per second actuating far fewer cones in the blue end of the spectrum than in the red end, yet we get a sensation of neutral white. Our neuro-visual system is what it is and does what it does, without regard to our analytic mathematical methods.

    Last night I looked at Polaris with binoculars and saw a pale yellowish white tint. Perhaps I misremembered something. It is an F8 evolved giant with a color index close to that of Alpha Centauri, which at G2V is similar to the Sun. When I get a chance I will look for stars with color index around 0.3 or 0.4, which at sea level should be a good match for how the Sun would look from out in space. I will defocus bright stars to see how they look as extended objects as well as points.

    I agree that a blackbody at the Sun’s effective temperature is a good choice as a primary standard for what we call a white emitter. I have no quarrel if astronomers say a G2 star as seen at sea level is pale yellowish white, as long as they state the appropriate caveats. My major grievance is with popular books that use saturated colors to illustrate the tints of various spectral types.

  6. #6
    Join Date
    Feb 2009
    Posts
    2,035
    Listing the brightest stars, their magnitudes, spectral classes and colour indices for F and G stars:
    Sun - -26,8, G2, quoted as +0,6
    Sirius - -1,43, A1
    Canopus - -0,74, A9
    Rigil Kentaurus - -0,27, G2/K1, quoted as +0,71 for +0,01/+0,88 for +1,33. What is the total colour index?
    Arcturus - -0,05, K0
    Vega - 0 by definition, A0
    Capella - 0,08, K0/G1, +0,80
    Rigel - 0,13, B8
    Procyon - 0,34, F5, +0,40
    Achernar - 0,46, B6
    ... (all are mostly AB or minority KM stars till...)
    Dubhe - 1,79, K0/F0, +1,07
    Mirfak - 1,80, F5, +0,48
    Wezen - 1,82, F8, +0,69
    Sargas - 1,84, F0, +0,40
    ...
    Polaris - 1,98, F8, +0,60
    ...
    Sadr - 2,23, F8, +0,67
    ...
    Caph - 2,28, F2, +0,34

    An important thing to note and keep in mind: stars with colour like Sun are rare in heaven. Most bright stars in sky are A and B stars much hotter and whiter than Sun, whether dwarfs or giants. And most of the rest are K and M giants much cooler and redder than Sun.
    So you have the list of the few stars that do look like Sun. If you know where to find them, what is their colour?

  7. #7
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by Hornblower View Post
    I would not be inclined to try to see a reason for the whiteness of the light in a graph of the the intensity as a function of the wavelength or frequency. We can analyze this in four different ways, using two versions of Planckís formula for blackbodies:

    Power as a function of wavelength
    Photon flux as a function of wavelength
    Power as a function of frequency
    Photon flux as a function of frequency

    This exercise gives us four very different light curves, none of which are flat over the visible range for a blackbody at the Sunís effective temperature.
    Here is a photon flux distribution (AM0) using SORCE data, though this is actual sp. irr. and not the idealistic Planck distribution. The variance between the two shouldn't be that much though the blue end would diminish quicker for the Planck distribution.

    Flat flux.jpg

    It's not super flat but neither are pancakes. Does this surprise you? If so, it may be due to the prolific number of energy distribution graphs, which produces a more dramatic, thus more useful perhaps, presentation.

    Here is one to compare the normally presented energy distribution with the converted photon flux distribution. This, of course, uses the same data, Thuilier in this case since it is an older graph.

    Thuilier Flux Compar.jpg


    The first one is nearly symmetrical, but is substantially elevated in the middle of the range.
    Right, for the blackbody at the Sun's temp. green is the peak. For the real Sun, the AM0 peak is around 460nm (blue), though the energy remains almost the same from 450nm to 465nm (blue and blue).

    A source that is really flat would look pale pink alongside sunlight, and we would really look silly if we define it as an absolute standard for white. If we adjust the spectrum to make any of the others flat, the tint will shift toward blue as seen alongside sunlight.
    I suspect the range of "pure" white is surprisingly large for a bright light source due to the color constancy issue. Newton called sunlight "perfectly white" during his prism work, which means the sunlight was coming through his window making the distribution closer to an AM1.5 or greater, so even less of a flat "pancake" photon distribution.

    Since color is determined by not only the spectral distribution of the light source but by the receptive spectral sensitivity of the eye, then we might want to favor green since I think this is the peak in reception, though I may be thinking not of the photopic range but of the mesotopic range, perhaps.

    [If we look at photon flux of sunlight, we see fewer photons per second actuating far fewer cones in the blue end of the spectrum than in the red end, yet we get a sensation of neutral white. Our neuro-visual system is what it is and does what it does, without regard to our analytic mathematical methods.] Yes, and it's interesting that we have only about 2%, IIRC, of our color cones that are "blue". Thus, the retinex has a fairly powerful blue amplifier in there somewhere. Nevertheless, we are able to model our receptivity to color, which greatly improves are analytics. I have yet to see very accurate computer model for it. The Colorado computer model still is claiming the Sun is peachy pink, though my avatar alone should awaken them, not that I didn't try to get their attention.

    Last night I looked at Polaris with binoculars and saw a pale yellowish white tint. Perhaps I misremembered something. It is an F8 evolved giant with a color index close to that of Alpha Centauri, which at G2V is similar to the Sun. When I get a chance I will look for stars with color index around 0.3 or 0.4, which at sea level should be a good match for how the Sun would look from out in space. I will defocus bright stars to see how they look as extended objects as well as points.
    If it isn't inconvenient, could you grab its SED for Polaris and I will do a comparison when I have time. Defocusing seems a little tricky for the eye. From Kitt Peak, using their Canon Rebel camera, I used progress defocusing and got a white result for 18 Sco (the best solar twin at that time).

    I agree that a blackbody at the Sunís effective temperature is a good choice as a primary standard for what we call a white emitter. I have no quarrel if astronomers say a G2 star as seen at sea level is pale yellowish white, as long as they state the appropriate caveats.
    Yes, or if the context is easily assume to be point sources or spectral use.

    My major grievance is with popular books that use saturated colors to illustrate the tints of various spectral types.
    Perhaps there are three sets of star colors that should be recognized: "up close and personal", which is the favored public view of them (extended objects at a normal photopic level); as point sources (redshifted); spectrally.
    We know time flies, we just can't see its wings.

  8. #8
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by chornedsnorkack View Post
    An important thing to note and keep in mind: stars with colour like Sun are rare in heaven. Most bright stars in sky are A and B stars much hotter and whiter than Sun, whether dwarfs or giants.
    As point sources, which is implied, yes, but white is their color.

    So you have the list of the few stars that do look like Sun. If you know where to find them, what is their colour?
    As point sources, white with a distinctive tint of yellow or gold to the eye. Eta Cas is just slightly hotter and every reference I found (~ 6) support this. 18 Sco I think can look a tint of yellow too, for a white star.
    We know time flies, we just can't see its wings.

  9. #9
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    7,656
    Quote Originally Posted by George
    The Colorado computer model still is claiming the Sun is peachy pink, though my avatar alone should awaken them, not that I didn't try to get their attention.
    That suggests to me that they are using an arbitrary definition of white that is at odds with how our eyes work when looking at stars.

  10. #10
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by Hornblower View Post
    That suggests to me that they are using an arbitrary definition of white that is at odds with how our eyes work when looking at stars.
    They state, "...the Sun's colour is in fact rather similar to that of a 5780 K blackbody. It looks peach pinkish, not yellow, doesn't it?" Their color is referenced to the D65 standard in the lighting industry. The D65 is a 6504K color temperature that, I think, is a modified Planck distribution where the blue-end is adjusted (bumped up) in two or more regions. Since the Sun is stronger in the blue end than a simple BB distribution, it makes some sense.

    Thus, compared to a "whiter" reference, if we choose to allow it, then the cooler solar temperature could be less white. The math model apparently makes it peach-pinkish. It reminds me of Einstein's comments about Lemaitre when he read his work saying, "your work is correct, but your physics is abominable." A simple pinhole projection by CASA would reveal the error of the model. Comparing it to the D65 standard is a bit ironic since it is essentially should be a comparison to itself. Worse, color constancy would still make either appear very white. [Einstein apologized and I would not be shocked if I'm not doing the same assuming an expert in this field steps in and can make sense of it. Perhaps color imaging needs these standards to get the proper results, but it looks inaccurate to me.]

    The D65, however, is a noon daylight result that comes from perhaps 600 different samples from different regions in the world, Europe especially, apparently. Daylight, represents both direct sunlight and skylight (blue sky), but it is a "power distribution".

    This raises a lot of questions for this amateur:

    How can one get an AM1 or greater color temperature of the Sun that his hotter than the Sun's actual central temperature in space (6390K), especially knowing the limb temperature is 5000K, and the best AM0 fit is about 5850K, and without atmospheric extinctions that bring this value even lower. The blue sky can be up to 8500K, though I see 6500K and 7500K values due to lessening by atmospheric effects, no doubt. But the claim is that of a power distribution so how can they (the industry) combine a tiny bit of blue sky flux to intense direct sunlight flux to make any real difference to the direct sunlight value (< 5850K) to the extent they claim (6504K)?

    It is also worth noting that their web page attempts to show the color of all stars, most are incorrect.
    Last edited by George; 2017-Nov-13 at 04:34 PM.
    We know time flies, we just can't see its wings.

  11. #11
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    7,656
    Finally, some clear weather. My color perception of some stars with my Celestron 8:

    Vega - pale bluish white.
    Caph - neutral white.
    Polaris - pale yellowish white

    From the published color index information I would expect the Sun to appear ever so slightly yellower than Polaris, perhaps imperceptibly so, under these conditions. This reinforces my expectation to see it as neutral white from above our atmospheric yellowing.

    Polaris and Caph did seem slightly yellower when sharply focused than as defocused extended objects. Perhaps the lack of blue cones in the fovea has something to do with this.

    The graph in post 7 suggests that the Sun would be slightly toward the blue from a true blackbody at the same effective temperature, but it is hard to tell for sure. While there is a peak in the blue range, there is also a deep notch just to the left that could at least partially offset it, as the blue cones have a rather broad range in which they are sensitive.

  12. #12
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by Hornblower View Post
    Finally, some clear weather. My color perception of some stars with my Celestron 8:

    Vega - pale bluish white.
    Caph - neutral white.
    Polaris - pale yellowish white
    I believe that is consistent with other accounts such as at Wiki. Though they are called "yellow" giants, Caph and Polaris, are both hotter than the Sun, but Caph is considered to be a white star. It's B-V (0.34) seems to remove the yellow tint, except when extinctions are higher or altitudes are lower, perhaps.

    Curious to me, there may be some historical influence at work for Vega. Giovanni Danti, who had influence with Father Secchi, established star types by color, as well as by spectral lines. He put Sirius, Vega, Procyon, Fomalhaut and even Rigel in the white class; Altair and Capella for the yellow class examples; Arcturus and Pollux as orange; Aldebaran, Antares and Betelgeuse as the red stars. Oddly, he seems to not have established any blue stars. This might be another reason the Sun was deemed yellow since the more blue-white stars fit this early white class. Since solar projections were done be even Galileo, it is hard to imagine he saw anything but a white Sun there. Of course, Secchi's attention and direction was always far more on spectral lines, which has been true ever since. Color was always secondary, no doubt, perhaps even with Danti.

    From the published color index information I would expect the Sun to appear ever so slightly yellower than Polaris, perhaps imperceptibly so, under these conditions. This reinforces my expectation to see it as neutral white from above our atmospheric yellowing.

    Polaris and Caph did seem slightly yellower when sharply focused than as defocused extended objects. Perhaps the lack of blue cones in the fovea has something to do with this.
    Perhaps. Does size matter? Could there be a slight color change due to a temperature gradient difference for a giant like Polaris vs. a solar-sized star? Probably not, but that would be an interesting question for someone who could give us optical depths for the giants, as well as, determine their CLV to allow us to compare temperature gradients. Of course we could simply look at its SED and see if there is a greater Planck variance.

    The graph in post 7 suggests that the Sun would be slightly toward the blue from a true blackbody at the same effective temperature, but it is hard to tell for sure.
    Yes, the Sun's actual energy distribution (AM0) in the blue band (450nm - 480nm) is 16.5% higher than the 5777K blackbody profile. But, the better model, IMO, is not the energy distribution but the photon flux distribution since our eyes, like CCDs, respond to photons. The much greater energy requirement to make a "blue" photon vs. a "red" gives us a big drop in the blue end of the spectrum for the photon distribution, surprisingly, causing the blue (and violet) band to be weaker than the rest of the colors.

    While there is a peak in the blue range, there is also a deep notch just to the left that could at least partially offset it, as the blue cones have a rather broad range in which they are sensitive.
    Perhaps. There is also the color band width that is rarely ever mentioned. Oddly enough, the most narrow band is the yellow.

    Can I assume that since we know the receptivity for each cone (like B&V filters) that we could produce our own natural eye version of B-V values given any SED? Is there a site that shows how to calculate B and V results? I found a site where I could download the filter data but I have never applied it.
    Last edited by George; 2017-Nov-14 at 10:02 PM.
    We know time flies, we just can't see its wings.

  13. #13
    Join Date
    Sep 2006
    Posts
    1,728
    If you want to compute a synthetic (B-V) color for some object, given the SED for the object (let's assume flux per unit wavelength as a function of wavelength) and some passbands for the "B" and "V" filters (sensitivity as a function of wavelength), you also need to have the SED for some star with known (B-V) color. Let's take Vega.

    The sequence of operations is:

    1. convolve SED of target object with B passband. Call the result T(B).
    2. convolve SED of target object with V passbnad. Call the result T(V).

    3. convolve SED of Vega (reference object) with B passband. Call the result R(B).
    4. convolve SED of Vega (reference object) with V passband. Call the result R(V).

    5. compute magnitude of target object in B-band: m(B) = -2.5 log10 ( T(B) / R(B) )
    6. compute magnitude of target object in V-band: m(V) = -2.5 log10 ( T(V) / R(V) )

    7. calculate color of target object (B-V) = m(B) - m(V)

    If you don't know how to convolve the SED with a passband, let me know, and I'll provide some instructions.

  14. #14
    Join Date
    Sep 2003
    Posts
    11,828
    Thanks for the above, StupendousMan.

    Quote Originally Posted by StupendousMan View Post
    If you don't know how to convolve the SED with a passband, let me know, and I'll provide some instructions.
    I can guess the procedure but I would rather learn it correctly up-front.

    For simplicity, assuming we had a perfectly flat SED, and I am using the Johnson-Cousins filters. Their highest "transmission" value for the blue filter is 79.69796753 (@ 421nm, violet), which I think means ~ 80% transmittance at that wavelength. For the blue band itself, the integrated transmittance might be something like 57% (this is the actual average from 455nm to 485nm, which is a close ballpark for what we see as blue, I think). For a flat SED, and say it is at 100w/m^2, I might get a result of 57w/m^2 passed by the filter. I will do perhaps 1nm bins to get better, integrated results, however. [I'm keeping in mind that these energy values are in derivative form and the entire spectrum will require integration if I seek a total result, but I don't need a total for this work, I think.]

    The V filter is little more narrow, ironically, but has better transmission value of ~ 94 at 522nm. Doing the same as for the blue filter would give me another wattage result. Converting these with your formula to get a mag. value will allow an accurate B-V result.

    Is the above correct?

    The same procedure should apply for each cone receptivity in lieu of filters. The comparison might be interesting, though color processing of the brain complicates things, but maybe not.
    Last edited by George; 2017-Nov-15 at 06:13 PM.
    We know time flies, we just can't see its wings.

  15. #15
    Join Date
    Sep 2006
    Posts
    1,728
    Your response indicated that you are doing the right thing.

    Quote Originally Posted by George View Post
    I will do perhaps 1nm bins to get better, integrated results, however.
    Yes, please do.

  16. #16
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by StupendousMan View Post
    Your response indicated that you are doing the right thing.
    Regression (progression in this case) to the mean.
    Thanks!

    I want to digest this a little. At first glance, it looks like it will take a fairly hot star (more blue) to provide a B-V value of 0 (e.g. Vega, 67% hotter than the Sun). Could this have contributed to some sort of "white" standard that would have reinforced a yellow Sun color assignment? B-V values were after Secchi, I think, so I only suggest a reinforcement, but it seems to add to the Sun's color story that I have enjoyed researching off and on.
    Last edited by George; 2017-Nov-16 at 07:04 PM.
    We know time flies, we just can't see its wings.

  17. #17
    Join Date
    Feb 2009
    Posts
    2,035
    A summary from my old post on the topic:
    https://forum.cosmoquest.org/showthr...t=#post2274882
    Huh. Some summarizing:
    4 of the 93 caused some trouble finding correct value of colour index (Acrux for no good reason, Toliman, Capella and Alnitak as multiples).
    Among the remaining 89, no star is bluer than the -0,27 of Naos, as expected. All others are from -0,25 redwards.
    The interval from -0,27 to +0,25 holds 54 stars out of 89 (all O, B and A stars and Canopus from class F)
    Then there are just 9 stars in the 0,75 wide range from +0,25 to 1,00 (all F save Canopus, and 2 from K, though the two G doubles should go somewhere there)
    15 stars are between +1,00 and +1,50: Dubhe from G9 and 14 K stars - those except the 2 which are bluer than +1,0 and 4 which are redder than +1,50
    And then 10 stars from +1,50 and +1,85: 4 K and 6 M stars.

    So: a plain majority of stars is within the -0,25 to +0,25 range

  18. #18
    Join Date
    Sep 2003
    Posts
    11,828
    With family in for Thanksgiving, six of us observed eta Cas (slightly hotter than the Sun) and we all saw white for eta Cas and orange for the close red dwarf. [8" Dob.]

    It wasn't super white but almost as if it was just one tint away from becoming pale yellow.

    This is in contrast to the time I saw it as having a bright tint of yellow or gold from McDonald Obs. just prior to the lunar eclipse of Sept. 2015. The dwarf was distinctly red. The altitude for eta Cas, after finally checking what it was, was only about 30 deg. (AM2) but at 6800 feet. Last night it was at about 60 deg (AM1.15).

    It seems obvious that the atmosphere plays a key role in seeing white stars as yellow. Particulates in the atmosphere likely also must be considered. Perhaps we had lots of smoke from Mexico that night at McDonald, too.

    I would expect Capella to appear yellowish, which I mention since the solar spectrum was lumped with Capella long ago and the Sun may have been somewhat officially yellow as a result.
    We know time flies, we just can't see its wings.

  19. #19
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    7,656
    After plenty of thinking about this topic, I am no more convinced than before that an equal number of photons in each primary color band is necessary for a sensation of neutral white in looking at a star in a dark field. Once again, our visual system is enormously adaptable, and I can easily imagine that it evolved into taking the spectrum of the naturally prevailing daylight as the standard for a star at night. I will concede that testing my idea would be difficult. My idea of a test would be to have many generations of observers confined to something like 3200K tungsten light, never seeing daylight, and see how they perceive a star of color index +0.3.

    As for Father Secchi and his colleagues, if their eyes were like mine I would say they called it as they saw it for stars of types G, K and M, and merely shortened their descriptive words to yellow, orange or red for jargon purposes to eliminate some verbal clutter. Their descriptions of hotter stars like Vega are still a head scratcher. Cleaner air would make it look more strongly blue than when looking through smog, not less.

  20. #20
    Join Date
    Sep 2003
    Posts
    11,828
    Quote Originally Posted by Hornblower View Post
    After plenty of thinking about this topic, I am no more convinced than before that an equal number of photons in each primary color band is necessary for a sensation of neutral white in looking at a star in a dark field.
    Is this for point sources and empirically based, given your likely extensive experience with observations? A flat photon flux distribution would make for a wonderful "perfectly white" standard, and I have suggested it before, and the Sun comes pretty close to it, though it drops off a little in the blue end of the spectrum. But the variations from a very flat distribution are likely greater than we might expect for a white result given things like color constancy. Also, our eye's receptivity isn't all that flat. Well, come to think of it, I don't think I've ever bothered to convert the energy receptivity to a photon flux density receptivity, though it should be fairly easy, especially if I could download the data into Excel.

    Once again, our visual system is enormously adaptable, and I can easily imagine that it evolved into taking the spectrum of the naturally prevailing daylight as the standard for a star at night.
    Evolution would be a considerable factor on how our eyes perform today, but other species have been around here even longer and they, well, see things differently.

    I will concede that testing my idea would be difficult.
    Maybe it wouldn't be hard at all. We now have no shortage of stellar SEDs, and we can compensate for atmospheric effects. Then can't we systematically look at a host of white or near white stars and determine when tints become noticeable, after countering the extinctions? I suspect that particle counts may be important to know the true extinctions since the high altitude at McDonald gave me a yellow tint, whereas at 800 feet elevation it looks white (I didn't compute the true air mass differences, admittedly). I recall that at McDonald, they will shut-down observations if particle counts reach a certain limit.

    My idea of a test would be to have many generations of observers confined to something like 3200K tungsten light, never seeing daylight, and see how they perceive a star of color index +0.3.
    Well, I'll just say that some "gedankens" are better than others.

    As for Father Secchi and his colleagues, if their eyes were like mine I would say they called it as they saw it for stars of types G, K and M, and merely shortened their descriptive words to yellow, orange or red for jargon purposes to eliminate some verbal clutter. Their descriptions of hotter stars like Vega are still a head scratcher.
    I want to say he graphed a thousand or more spectrums from his visual observations. A number of them would be near the G2 class, as we now know them, so if he also gave their apparent color as well, that would be interesting. However, I would guess that he would not have done both observations at the same time since he had rigged his nice refractor with the spectrometer, and I think this was at the objective making it a little troublesome to switch back and forth just to get a color determination, which would have been of little importance even then, less so today.

    Cleaner air would make it look more strongly blue than when looking through smog, not less.
    Of course, so did I accidentally muff this? Selective scattering can occur that can have the opposite effect, admittedly, but it would be very rare. I wonder what sort of air pollution existed around Rome in the late 1800s during Secchi's work at the college Romano?
    Last edited by George; 2017-Nov-29 at 06:10 PM.
    We know time flies, we just can't see its wings.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •