Page 2 of 2 FirstFirst 12
Results 31 to 45 of 45

Thread: Sun Color Challenge

  1. #31
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    Quote Originally Posted by Hornblower View Post
    When I look at the published spectra, it does indeed appear that the Sun would look somewhat bluer than the perfect blackbody with the same effective temperature. It should be fairly straightforward to make satisfactory laboratory simulations. I don't think we need to duplicate the snaggletoothed shape of the solar spectrum, as the three types of cones respond to broad bands with lots of overlap.
    Addendum: Here is a Wiki graph showing the differences between the Sun's spectrum and a 5778K blackbody curve. Note that the curve fits almost perfectly throughout the infrared and into the red, has a sharp peak in the blue, and drops off sharply in the violet and ultraviolet. I attribute the latter to the H and K calcium lines.
    https://upload.wikimedia.org/wikiped...pectrum_en.svg
    Here is a widened view of the visible range, with a 6050K blackbody plot that fits most of the visible range. My educated guess is that the elevated blue is pretty well balanced by the dropoff in the violet for visual purposes, since our blue cones respond to a broad band, and that the Sun and a 6050K blackbody would look virtually alike as spots in a dark field above the atmosphere.

    Sun v. Planck 2019-06-15 with 6050K.jpg

  2. #32
    Join Date
    Sep 2003
    Posts
    12,717
    Quote Originally Posted by Hornblower View Post
    Here is a widened view of the visible range, with a 6050K blackbody plot that fits most of the visible range.
    Let me know what you think of the following. I favor 5850K for a better BB match, no that it matters much, I suppose.

    Planck BB comparisons with Sun.jpg

    BTW, do you have a favored site or paper that does a nice job defining the wavelengths for the various colors? It's a little surprising how almost every site is different in their claims. I expect that subjectivity becomes an issue when trying to get it down to a few nanometers, but the variance is much greater. I think a fair amount of intensity is assumed for those that suggest the range is from 390nm to 750nm. At low light levels, perhaps 700nm is a limit.
    We know time flies, we just can't see its wings.

  3. #33
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    Quote Originally Posted by George View Post
    Let me know what you think of the following. I favor 5850K for a better BB match, no that it matters much, I suppose.

    Planck BB comparisons with Sun.jpg

    BTW, do you have a favored site or paper that does a nice job defining the wavelengths for the various colors? It's a little surprising how almost every site is different in their claims. I expect that subjectivity becomes an issue when trying to get it down to a few nanometers, but the variance is much greater. I think a fair amount of intensity is assumed for those that suggest the range is from 390nm to 750nm. At low light levels, perhaps 700nm is a limit.
    I adjusted the brightness of my 6050K curve until it matched the solar value at 700nm, and that got it close until it encountered the snaggletoothed stuff in the blue to violet range. I found the fit over the right two thirds of the graph to be better than at 5850K and vicinity.

    I do not know of a precise mathematical model for our color vision. I perceive yellow with a tinge of orange for sodium emission at 589.0 and 589.6, but I cannot rule out the possibility that some others might see a different hue because of genetic variation in the makeup of the cones. Our color vision is extremely sensitive to slight variation in the wavelength or the mixture of wavelengths in that range.

  4. #34
    Join Date
    Sep 2003
    Posts
    12,717
    Quote Originally Posted by Hornblower View Post
    I adjusted the brightness of my 6050K curve until it matched the solar value at 700nm, and that got it close until it encountered the snaggletoothed stuff in the blue to violet range. I found the fit over the right two thirds of the graph to be better than at 5850K and vicinity.
    Yes, and I suspect about 2/3rds of the energy is in that 2/3rds of the spectrum as well. A 5772K value will give the same luminosity as the Sun, but that is for a black body and not a star with some line extinctions.

    Perhaps, for color purposes, something between 6050K and 5850K would be a better match in order to fill the blue-end better.


    I do not know of a precise mathematical model for our color vision. I perceive yellow with a tinge of orange for sodium emission at 589.0 and 589.6, but I cannot rule out the possibility that some others might see a different hue because of genetic variation in the makeup of the cones. Our color vision is extremely sensitive to slight variation in the wavelength or the mixture of wavelengths in that range.
    Yes, there is certainly an element of subjectivity due to several variables including one's eye performances, ambient lighting, neighboring colors and intensity. The one exception to this white since our powerful brain's ability to adjust colors for some of these colors is often incredibly accurate, though sometimes it is tricked, yet there is also "white balance" that works with constancy to make the brightest light white. So a very bright object that has just a tint of a color to it will shift this to a white object without tint. The dramatic color shift of incandescent car lights seen in day vs. night is a great example. But the evidence seems overwhelming to me that the Sun won't be found with any tint due to its all-white disk projection as in my avatar. The white limb especially makes a strong argument for this.
    We know time flies, we just can't see its wings.

  5. #35
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    Quote Originally Posted by George View Post
    Yes, and I suspect about 2/3rds of the energy is in that 2/3rds of the spectrum as well. A 5772K value will give the same luminosity as the Sun, but that is for a black body and not a star with some line extinctions.

    Perhaps, for color purposes, something between 6050K and 5850K would be a better match in order to fill the blue-end better.


    Yes, there is certainly an element of subjectivity due to several variables including one's eye performances, ambient lighting, neighboring colors and intensity. The one exception to this white since our powerful brain's ability to adjust colors for some of these colors is often incredibly accurate, though sometimes it is tricked, yet there is also "white balance" that works with constancy to make the brightest light white. So a very bright object that has just a tint of a color to it will shift this to a white object without tint. The dramatic color shift of incandescent car lights seen in day vs. night is a great example. But the evidence seems overwhelming to me that the Sun won't be found with any tint due to its all-white disk projection as in my avatar. The white limb especially makes a strong argument for this.
    I think I made the 6050 curve fit the blue band reasonably well. It splits the difference between the bump on the right side of that band and the hole on the left side. If we could make a laboratory source that matches the Sun's pattern it would be interesting to see how closely that source and a 6050K blackbody match visually. I think it would come down to the relative sensitivity of the eye to these two subdivisions.

    I see no compelling reason to take that avatar as any authoritative source for how the Sun should look under various conditions. So much depends on how it was exposed and on the characteristics of the monitor on which it is viewed.

  6. #36
    Join Date
    Sep 2003
    Posts
    12,717
    Quote Originally Posted by Hornblower View Post
    I think I made the 6050 curve fit the blue band reasonably well.
    I think our 6050K curves don't match for some reason. Is your curve represented by the dots on your graph?

    I see no compelling reason to take that avatar as any authoritative source for how the Sun should look under various conditions. So much depends on how it was exposed and on the characteristics of the monitor on which it is viewed.
    You're questions are well-suited for the resolution of the color conundrum. It all, IMO, boils down to this image and I'm convinced it says it all.

    1) There were no filters used at this observatory (McMath-Pierce).

    2) There is no reason to think color alteration would occur in its path to the projection table. I think the f50 path is in a vacuum as well.

    3) Ambient lighting must be considered. But the powerful color constancy ability of the brain to make this object white plays a key role in resolving that the Sun is white. One key point here is that the limb is white (the very edge might have a tiny tint of yellow, but it's hard to say). Given the large temperature difference (CLV) of 6390K at the center and 5000K at the limb, any change that the bulk of the disk could have any yellow tint is ruled-out, IMO. If the limb, with its much greater levels of yellow, orange, and reds from the incident sunlight looks white after atmospheric extinctions, then we have a strong argument that the Sun, as seen in space, can't be yellow. It would take very strong blue ambient lighting in the projection room to render a yellow limb and that's not the lighting used.

    4) Since the atmospheric extinctions are exponential toward the blue-end of the spectrum, the white solar image (e.g. my avatar) has depletion of the blue-end photons, so to render the Sun's color as seen from space (AM0) then we must add all those scattered blue-end photons back into the sunlight, which, of course, would never produce any hint of a tint of yellow. If anything, the hotter blue center of the Sun could gain a hint of blue, but I think this is highly unlikely as well.

    [Added: The ambient lighting question, I think, can be easily dismissed by any normal pin-hole or equivalent solar projection of an AM1 or AM1.5 Sun, which means high particle count events or coated lenses must not be allowed as evidence. The result is always a white disk even at the limb, right?]
    Last edited by George; 2019-Jun-22 at 12:30 AM. Reason: grammar
    We know time flies, we just can't see its wings.

  7. #37
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    Quote Originally Posted by George View Post
    I think our 6050K curves don't match for some reason. Is your curve represented by the dots on your graph?
    Yes, my curve is represented by the dots. Yours and mine will match with suitable equalization of the coordinates. I adjusted the brightness of mine to match the solar spectrum at 700nm, and that put it very close throughout the red and green range and split the bumpy blue range.
    You're questions are well-suited for the resolution of the color conundrum. It all, IMO, boils down to this image and I'm convinced it says it all.

    1) There were no filters used at this observatory (McMath-Pierce).

    2) There is no reason to think color alteration would occur in its path to the projection table. I think the f50 path is in a vacuum as well.

    3) Ambient lighting must be considered. But the powerful color constancy ability of the brain to make this object white plays a key role in resolving that the Sun is white. One key point here is that the limb is white (the very edge might have a tiny tint of yellow, but it's hard to say). Given the large temperature difference (CLV) of 6390K at the center and 5000K at the limb, any change that the bulk of the disk could have any yellow tint is ruled-out, IMO. If the limb, with its much greater levels of yellow, orange, and reds from the incident sunlight looks white after atmospheric extinctions, then we have a strong argument that the Sun, as seen in space, can't be yellow. It would take very strong blue ambient lighting in the projection room to render a yellow limb and that's not the lighting used.

    4) Since the atmospheric extinctions are exponential toward the blue-end of the spectrum, the white solar image (e.g. my avatar) has depletion of the blue-end photons, so to render the Sun's color as seen from space (AM0) then we must add all those scattered blue-end photons back into the sunlight, which, of course, would never produce any hint of a tint of yellow. If anything, the hotter blue center of the Sun could gain a hint of blue, but I think this is highly unlikely as well.

    [Added: The ambient lighting question, I think, can be easily dismissed by any normal pin-hole or equivalent solar projection of an AM1 or AM1.5 Sun, which means high particle count events or coated lenses must not be allowed as evidence. The result is always a white disk even at the limb, right?]
    Your avatar image as rendered here is overexposed. The RGB values are at or near maximum possible, which will wash out pastel tints. My images, exposed enough darker to make the center light gray, show significant loss of blue light at the limb. With the neutral gray on your limb, the center would be pale bluish white at a similar exposure. Since there is some atmospheric discoloration even at the high altitude of Kitt Peak, that suggests the possibility of a slight color shift in your camera.

    Here is one of my images adjusted to make it slightly blue at the center and slightly yellow at the limb. It is my educated guess that a pinhole projection out in space would be close to this.
    b1045b blue.jpg

  8. #38
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    Quote Originally Posted by Hornblower View Post
    Yes, my curve is represented by the dots. Yours and mine will match with suitable equalization of the coordinates. I adjusted the brightness of mine to match the solar spectrum at 700nm, and that put it very close throughout the red and green range and split the bumpy blue range.


    Your avatar image as rendered here is overexposed. The RGB values are at or near maximum possible, which will wash out pastel tints. My images, exposed enough darker to make the center light gray, show significant loss of blue light at the limb. With the neutral gray on your limb, the center would be pale bluish white at a similar exposure. Since there is some atmospheric discoloration even at the high altitude of Kitt Peak, that suggests the possibility of a slight color shift in your camera.

    Here is one of my images adjusted to make it slightly blue at the center and slightly yellow at the limb. It is my educated guess that a pinhole projection out in space would be close to this.
    b1045b blue.jpg
    Addendum: I set my digital camera for daylight in its white balance setting, and then took some test photos of a sheet of white paper at high noon with a very clear sky. The image came out slightly bluish white in my computer artwork software, with the RGB values differing from neutral gray or white by about the same amount as in my test images of the Sun from an earlier project. Since the brighter center dominates the total light, I think the limb would have been neutral or very nearly so, in good agreement with your Kitt Peak image.

    Some technical comparisons: My test shot yesterday was near sea level, with an admixture of sky blue from about half of the sky, the remainder being screened off by large trees. The Sun by itself would be slightly bluer at Kitt Peak because of the high altitude, but virtually none of the sky would contribute in this McMath projection. That may very well balance out. That shows that our cameras are in good agreement and are set to render a slightly "warmer" daylight, such as on a murky summer day that is typical here in northern Virginia, as neutral white.

    While I was at it I took some photos of my computer screen showing a swatch of neutral white. It was very close to matching daylight.

    The image of the Sun in post 37 was not that white originally. I took it with my improvised camera obscura setup on a December morning, and it was significantly yellowed by the low elevation. I would need to cut a hole in the roof to do that near the zenith. I shifted the color balance in the computer, and I would expect it to preserve the difference between the center and the limb.

    Let me add that the brightly lighted paper looked slightly yellowish white, but that is typical of my color vision when overloaded. That could provide a false argument for someone insisting that the Sun is yellow. My own observations of stars with allowance for published values of atmospheric discoloration convince me that a star that matches the Sun's spectrum will look neutral white as seen from above the atmosphere. That is the clincher in my opinion.

  9. #39
    Join Date
    Sep 2003
    Posts
    12,717
    Lots of traveling including the McDonald Observatory., but no phone service at lodge. In Grand Juction today.
    Quote Originally Posted by Hornblower View Post
    Yes, my curve is represented by the dots. Yours and mine will match with suitable equalization of the coordinates. I adjusted the brightness of mine to match the solar spectrum at 700nm, and that put it very close throughout the red and green range and split the bumpy blue range.
    Ah, of course, that’s very nice.

    If we are considering a perfectly white def. then we might be better finding a nice fit for a AM1 or AM1.5 sp. irr. curve.

    Your avatar image as rendered here is overexposed. The RGB values are at or near maximum possible, which will wash out pastel tints.
    I sent a new light meter to Dr. Potter at KP to give us objective evidence the image was well within a @normal@ light level. I knew it was since it wasn’t that bright to the eye when observing it. The processing of that image to make others is likely off a bit.

    My images, exposed enough darker to make the center light gray, show significant loss of blue light at the limb.
    That’s logical since the cooler limb temp. will emit less blue light.
    With the neutral gray on your limb, the center would be pale bluish white at a similar exposure.
    Yep, as expected. The 1390K CLV is a significant color producing difference, which really makes the case against any yellow for the Sun if the projection has no yellow tint on the limb.

    Since there is some atmospheric discoloration even at the high altitude of Kitt Peak, that suggests the possibility of a slight color shift in your camera.
    That’s sure possible. The camera was their Canon Rebel — nice but not ideal.

    Iphone
    Last edited by George; 2019-Jun-25 at 07:02 PM.
    We know time flies, we just can't see its wings.

  10. #40
    Join Date
    Sep 2003
    Posts
    12,717
    Quote Originally Posted by Hornblower View Post
    Let me add that the brightly lighted paper looked slightly yellowish white, but that is typical of my color vision when overloaded. That could provide a false argument for someone insisting that the Sun is yellow.
    I see white but even if one sees some yellowish tint for a projection, it would require strong argument that the blue extinctions from the atmosphere are too trivial to restore it as white.

    A tenet of science is the requirement to allow falsification of hypotheses and theories. If a falsification is found, either tweaking is required or complete dismissal is required. Of the long history with millions of unfiltered images taken of a projected Sun revealing a white disk, be it hundreds of tiny solar eclipse images found in the shade of trees, or reflector telescope projections, or camera obscura or other pin-hole images, etc. it's worth noting that only one clean white projection is required to falsify every claim that the Sun is some level of yellow (if seen in space at a normal intensity level). Otherwise, the Sun itself would have to change in surface temperature dramatically to allow that white projection to exist, which is extremely unlikely to say the least.

    Your work above seems to give evidence that a white result is a fair result and not of one from color distortions of one kind or another.

    My own observations of stars with allowance for published values of atmospheric discoloration convince me that a star that matches the Sun's spectrum will look neutral white as seen from above the atmosphere. That is the clincher in my opinion.
    Yes, but don't forget about the great point you made not too long ago about point sources -- a solar twin might appear with a tint of yellow, perhaps even if seen from space, due to, IMO, the fact that the fovea lacks blue color cones. Solar disk projections don't suffer this issue, so they better represent a star's true color. This may help explain Father Secchi's yellow color assignment for the Sun due to the Sun's similar spectral distributions with stars, such as Capella (Secchi's comparative selection).

    18 Sco (formerly the best solar twin known) might look yellowish white on high-particle nights, though I see it as white at near sea level the times I looked at it, yet I did see eta Cas (slightly hotter than the Sun) as golden-yellow one night but white thereafter. And, perhaps, this is an area where "normal" eyes become less normal as well. A camera helps here as it can take an overly bright point source and smear it (diffusion) to reveal what its disk might look like at a normal level. Here is 18 Sco using progressive defocussing:

    18 Sco 4 small.jpg
    We know time flies, we just can't see its wings.

  11. #41
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    Upon glancing at my camera obscura projection, my initial reaction was "That's white, period. With additional concentration I could see that it had a slight yellowish green tinge, which turned out to be in part due to the binocular optics I was using for eyepiece projection. I verified this with comparison shots with the one-inch hole shining on the screen without the binoculars. I certainly needed no further efforts to convince me that the Sun's unfiltered light is visually not the least bit yellow.

  12. #42
    Join Date
    Sep 2003
    Posts
    12,717
    Quote Originally Posted by Hornblower View Post
    Upon glancing at my camera obscura projection, my initial reaction was "That's white, period. With additional concentration I could see that it had a slight yellowish green tinge, which turned out to be in part due to the binocular optics I was using for eyepiece projection. I verified this with comparison shots with the one-inch hole shining on the screen without the binoculars. I certainly needed no further efforts to convince me that the Sun's unfiltered light is visually not the least bit yellow.
    Yes, the evidence is overwhelming. If at seal level we can see a clean (unfiltered but with normal scattering of mostly blue light) white disk for the Sun, at a fairly high altitude then how could the Sun as seen in space ever have a hint of yellow to it? It would be like adding blue paint to white paint and expecting a yellow result.

    It's quite obvious, yet less than 1% (WAG) of all Sun depictions and images are yellow or orange. The UT article (2nd on the current list) about a comet interceptor gets most of the planet's colors nicely but completely blows the others. Earth, Jupiter, Mercury and Mars are close to true color, but they use an old false colored radar image for Venus and an orange Sun. I get that white undermines the ferocity of the Sun, but give me a break (if no one else ).
    We know time flies, we just can't see its wings.

  13. #43
    Join Date
    Sep 2003
    Posts
    12,717
    Here's my stab at some sort of a "white" color definition with the use of a blackbody (Planck) temp.

    This result is surprisingly close to your 6050K temperature for the AM0 sp. irr. curve. I had thought that the loss of blues by Rayleigh scattering would bring it down more than 150K. I am discounting the violets due to the eye's poor receptivity to it. The 390nm to 800nm is used to capture the range for even the above average eye's spectral receptivity.

    White sun model.jpg

    The AM1.0 values are from NASA/ASTM (1980 pub.). I chose their turbidity coefficients of: alpha = 0.66 and beta = 0.085.

    These values generate greater extinctions than those used for a clean sky; these coefficents model non-heavy city values for aerosols (size and density).

    I also suspect that the AM1.5 values for the Sun will not change this Planck matching value much.
    Last edited by George; 2019-Jul-03 at 06:13 PM. Reason: grammar
    We know time flies, we just can't see its wings.

  14. #44
    Join Date
    Mar 2007
    Location
    Falls Church, VA (near Washington, DC)
    Posts
    8,604
    The more I think about it, the more I am of the opinion that it is not essential for astrophysics that we have an explicit definition for a visually neutral white light source. However, suppose for the sake of argument that some of us consider it important. Rather than being super analytical about fitting curves to a snaggletoothed stellar spectrum I would advocate the observational approach I used recently. I looked at several stars in the Water Jar portion of Aquarius and noted which ones looked warm white, neutral white or cold white. As it happened none of them were vividly either yellow or blue. I did this without knowing the spectral type or published color index, to avoid confirmation bias. Then I looked up that information and concluded that mid-F stars looked neutral white, as discolored by the atmosphere. In principle this could be done with numerous observers, and the results could be averaged.

    For what it may be worth, I do not see an obvious yellow shift when going from a defocused spot to a sharply focused point with a white star. I am willing to entertain the idea that my brain cells may be allowing for the lack of S type cones in the fovea with and generating the sensation of white from M and L cones alone in this part of the retina.

  15. #45
    Join Date
    Sep 2003
    Posts
    12,717
    Quote Originally Posted by Hornblower View Post
    The more I think about it, the more I am of the opinion that it is not essential for astrophysics that we have an explicit definition for a visually neutral white light source.
    Yes, which explains why someone like me can have fun with it; if it were important, it would have been in textbooks long ago.

    However, suppose for the sake of argument that some of us consider it important. Rather than being super analytical about fitting curves to a snaggletoothed stellar spectrum I would advocate the observational approach I used recently. I looked at several stars in the Water Jar portion of Aquarius and noted which ones looked warm white, neutral white or cold white. As it happened none of them were vividly either yellow or blue. I did this without knowing the spectral type or published color index, to avoid confirmation bias. Then I looked up that information and concluded that mid-F stars looked neutral white, as discolored by the atmosphere. In principle this could be done with numerous observers, and the results could be averaged.
    Yes, I think you'll find many stating these F stars are white, including the H-R diagrams where color is added.

    Whatever approach taken to establish what stars would be seen as white (up-front and personal; attenuated disks), it would have to be done in a very broad way. I'm convinced that bright objects with dark black backgrounds will color-shift toward a white result. Adding equally bright colors around a, say bright blue adjacent to a white looking hotter K-star, should render yellow at the cooler limb at least. But observing stellar disks come with the assumption of black backgrounds, so such color contrast examples and other illusions can be dismissed, IMO.

    Close to exact color determinations for stars, as point-sources, is problematic given the changing particle counts for atmospheres and the problem of whose vision is least affected by the lack of blue cones in the fovea, but...

    For what it's worth, I don't see an obvious yellow shift when going from a defocused spot to a sharply focused point with a white star. I am willing to entertain the idea that my brain cells may be allowing for the lack of S type cones in the fovea with and generating the sensation of white from M and L cones alone in this part of the retina.
    Would that be true for the bright stars, say mag 1 or brighter (adjusting for magnification)? I assume so, but the cones aren't very active unless the source is bright enough. I admit that I haven't really tested this idea even though it is a simple test. After seeing a yellow eta Cas from an 8" Dob on a mountain top, then at a latter date as white seen from near sea level, it seems particle counts count a lot.
    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
  •