Starship Asterisk*

Conjunction Colours

Explanation: During the past week, nightfall on planet Earth has featured Mars, Saturn, and Spica in a lovely conjunction near the western horizon. Still forming the corners of a distinctive celestial triangle after sunset and recently joined by a crescent Moon, they are all about the same brightness but can exhibit different colors to the discerning eye. This ingenious star trail image was recorded as the trio set on August 12 with a telephoto lens from the shores of Lake Eppalock, in central Victoria, Australia. Focused on foreground eucalyptus trees, the image slightly blurs the trails to show more saturated colors. Can you guess which trail is which? Of course the reddest trail is Mars, with Saturn on the right a paler echo of the Red Planet's hue. Left is hot and luminous Spica, bluish alpha star of the constellation Virgo.

<< Previous APOD

This Day in APOD

Next APOD >>

[/b]

Red, White and Blue Triangle to Shine in Night Sky
Discovery News | Joe Rao | 2012 Aug 20

That's a lovely portrait of the colors of Saturn, Mars and Spica!

Ann

There are a few things a bit odd about the impressive color trails:

1) The color near the horizon does not seem to change. Atmospheric effects, even with a pristine morning, will cause some reddening or desaturation. "Never a white sun has set" - I think I just made that up, but it's true. Of course, this is a morning view and the scattering effects are much less in the clean air mornings, but you get the point.

Perhaps, however, the trees were on a hill, so the star trails we see might be several degrees above the dirty horizon.

2) The colors seem enhanced, though perhaps just Spica. The Planck temperature needed to produce a blue color that matches, for example, our blue sky (overhead) requires temperatures over 10 million degrees, far greater than Spica's 22,000 K. This is because hotter stars also produce more of the red end of the spectrum, so the gain in blue saturation is marginal with temperature increases. This makes the hottest blue stars bluish-white.

what a buzz to get my image on APOD!

for helio.. yes this image was captured reasonably high in the sky above the tree.. not much reddening as the stars drifted down in this field of view. was lucky to get a calm night though otherwise the tree would have been blurred out. there's a little bit of enhanced saturation in lightroom but i tried not to overdo it.

a few more details here: http://philhart.com/content/conjunction-colours

Phil

Helio,
If Spica is quickly moving away from us, would it apear more bluish?

Helio George wrote:1) The color near the horizon does not seem to change. Atmospheric effects, even with a pristine morning, will cause some reddening or desaturation. "Never a white sun has set" - I think I just made that up, but it's true. Of course, this is a morning view and the scattering effects are much less in the clean air mornings, but you get the point.

This is an evening conjunction, not a morning one. However, it is only a 20-minute exposure, so it's likely that the total path length of the objects (around 5°) simply isn't enough to show a color gradient- particularly, as seems likely from the trees, if this was already somewhat above the horizon.

2) The colors seem enhanced, though perhaps just Spica. The Planck temperature needed to produce a blue color that matches, for example, our blue sky (overhead) requires temperatures over 10 million degrees, far greater than Spica's 22,000 K. This is because hotter stars also produce more of the red end of the spectrum, so the gain in blue saturation is marginal with temperature increases. This makes the hottest blue stars bluish-white.

It is impossible on a computer monitor to accurately reproduce both the color and intensity of blackbody sources. This image has been exposed and processed in a way that essentially represents the three objects near the maximum brightness possible without saturating any color channel. Doing that effectively ignores the brightness, and maximizes the range of chromaticity. You can run color temperature through a CIE function and calculate the ideal RGB value to represent it (ignoring brightness); for 22000K, that is 166/195/255 - which is very close to the color value in the image (sampled in the middle of the path as 164/206/251).

Wadsworth wrote:If Spica is quickly moving away from us, would it apear more bluish?

Spica is moving away from us at only 1 km/s, which results in a Doppler redshift, although far too small to be detected visually.

Helio George wrote:
2) The colors seem enhanced, though perhaps just Spica. The Planck temperature needed to produce a blue color that matches, for example, our blue sky (overhead) requires temperatures over 10 million degrees, far greater than Spica's 22,000 K. This is because hotter stars also produce more of the red end of the spectrum, so the gain in blue saturation is marginal with temperature increases. This makes the hottest blue stars bluish-white.

I disagree. Astrophotographer David Malin once said that A0-type star Vega, at about 10,000K, is about the same blue color as the Earth's sky.

Bear in mind that the Earth's sky is not a pure blue color in any way. All the colors of the Sun scatter in the atmosphere, although the shortest wavelengths scatter far more efficiently than the longer wavelengths. The sky is a light shade of blue, not a saturated shade of blue. That is precisely because the blue photons that reach us from the clear sky are mixed with violet, green, yellow and even a few orange and red photons, so that the color of the sky is diluted and light. That's what makes the Earth's sky similar in color to the hot stars: their light, too, is predominantly, but certainly not purely, blue.
This diagram shows the "relative color" of stars of different spectral classes. No star emits just one color, so their "combined color" is a mix of many wavelengths.

As can be seen from the diagram, however, O-type and early B-type stars (like Spica) emit far more blue (short-wave) than red (long-wave) light. The opposite is true of the M-type stars: They emit far more red than blue light, although even their light is far from a pure red.

The spectra have probably been adjusted somewhat for clarity, but one thing is certain: Hot stars emit far more blue than red light, and there is no reason why their color should not be compared with that of the Earth's sky.

But hot stars so rarely look blue. Why not? It's because the "blue light detectors" in our eyes don't work well when lighting is faint. Stars are almost always too faint to really trigger any response in the "blue detecting cones" in our eyes. On the other hand, our color-blind rods respond rather strongly to blue (and green) light. Therefore, when we look at a hot star, our blue-detecting cones will not respond much (or at all), but our color-blind rods will react and create the impression of colorless (white) light.

When we look at a hot star through a telescope, it is much easier to detect the blue color of the star. But now we encounter another problem. As you all know, you must never stare directly at the Sun, because if you do you may go blind. Suppose, however, that you decided that finding out the color of the Sun would be worth losing the sight of your eyes, so that you chose to stare at the Sun. What would you see? Would you see yellow light?

You wouldn't. The Sun is so overwhelmingly bright that you would only see blinding white light, before your eyes gave up forever. You run no risk of going blind by looking at another star than the Sun through a telescope. But in a hot star, the light is so "concentrated", and the surface brightness is so extremely high (much, much higher than the surface brightness of the Sun) that the intrinsic blue light of the star easily gets "burnt out", so that the star looks white again.

But in today's APOD, the concentrated light of Spica has been "spread out over a much larger area", clearly revealing the blue color of Spica.

So there is nothing wrong with the blue color of Spica in this picture, although it's true that you can't expect to see this very obvious color either with the naked eye or with the aid of a telescope.

Ann

Ann wrote:Astrophotographer David Malin once said that A0-type star Vega, at about 10,000K, is about the same blue color as the Earth's sky.

However, he was basically wrong about this. The statement reflects confusion about what "color" means. Color is not a matter of physics, but of human perception. The human eye is not a wavelength detector- that is, we don't convert some particular wavelength, or combination of wavelengths, into a particular perceived color. Intensity is a key component of our color vision. Take a single pure wavelength- say 460 nm - and what color will we see? Depending on the brightness we'll see anything from black, to dim gray, to a range of blues, to white.

When we look at Spica, we can also see this same range, depending on our eyes, the atmosphere, and whether we are are using a telescope or not (and if so, what the aperture of the telescope is, as well as its optical quality). So it is very difficult to assign an actual color to a star that could be compared to another color, such as the sky.

Perhaps what Malin meant was that the two objects have similar chromaticities: if you normalize their brightnesses properly, they will be perceived as having about the same color.

Chris wrote:
Perhaps what Malin meant was that the two objects have similar chromaticities: if you normalize their brightnesses properly, they will be perceived as having about the same color.

Perhaps that's what Malin meant, and it is what I meant when I said that Spica is about the same color as the Earth's sky. I should have used the expression "chromaticity" instead.

I did try to explain why hot stars rarely look blue, in spite of their blue chromaticity.

Ann

Ann wrote:I did try to explain why hot stars rarely look blue, in spite of their blue chromaticity.

BTW, the very different responses of our rods and cones isn't the only thing that makes it hard for us to see star colors. With blue especially, the density of the cones is low (ever notice how blue signs tend to look fuzzy at night?). So to see star colors well, we can do two things- gather more light, which means using a telescope, and make the star cover more of our retina, by defocusing slightly (the same technique used in today's APOD). It's amazing how much more saturated the colors can appear if you do that. For bright stars like Spica, you can even get better color visually by simply unfocusing your gaze a bit, since such stars are bright enough to trigger color response even without optical aid.

This is a lovely, vivid picture. It was a very bright idea to focus on the foreground trees and defocus the star and planets. The eucalyptus trees make for a nice composition and give a clear sense of place. And I love how the more southerly Spica appears higher in the sky than Mars and Saturn -- just the opposite of what I see from my mid-northern latitude.

thanks for the discussion Chris and Ann.. very interesting! i was actually talking to David Malin about colour/saturation etc at the Malin awards in July. He is a stickler for correct colour/saturation/chromaticity.

in lightroom 4, i increased the contrast to +30, clarity to +10, and increased saturation and vibrance to +40. as i understand it, that should maintain chromaticity but increases saturation. the effect is stronger towards the sides of the trails than in the centre. white balance was 'daylight' on camera.. colour temp 4850, tint 0.

Phil

Here's a version of the image without any increase in contrast/saturation/vibrance: Phil

philiphart wrote:what a buzz to get my image on APOD!

for helio.. yes this image was captured reasonably high in the sky above the tree.. not much reddening as the stars drifted down in this field of view. was lucky to get a calm night though otherwise the tree would have been blurred out. there's a little bit of enhanced saturation in lightroom but i tried not to overdo it.

That makes sense, though I am impressed to see so little color variation with altitude, especially for an evening shot. You must have had a clear sky, though the calmness you mentioned also improves atmosphere clarity, especially if it was calm the entire day. Atmospheric dust from afternoon convection can dramatically alter star colors near the horizon.

Is is quite a nice shot for those interested in celestial color! Stefan Seip and David Marlin have produced some great star color trails,too. They often progressively defocus to produce expanded tails.

Ann wrote:I disagree. Astrophotographer David Malin once said that A0-type star Vega, at about 10,000K, is about the same blue color as the Earth's sky.

Does he live on the coast? This is bound to be in error, unless the sky he sees is desaturated blue.

The following illustrates that, perhaps, around 75,000K (Planck distribution) temperature will produce the color ratios that come reasonable close to a blue sky spectrum I found. [The blue sky spectrum, I suspect, is of an AM1 (coastal) spectrum. I have seen blue sky SEDs that have higher blue end levels. It took over 10 million deg. (Planck) to match one that I found.]



Notice that Spica certainly favors blue, but it is not quite a "blue sky" match. The image by Phil Hart is also a bluish-white color, and not a deep saturated blue. So I don't want to overstate the case.

Bear in mind that the Earth's sky is not a pure blue color in any way. All the colors of the Sun scatter in the atmosphere, although the shortest wavelengths scatter far more efficiently than the longer wavelengths. The sky is a light shade of blue, not a saturated shade of blue.

Yes, and altitude makes a considerable difference. There are some interesting stories of the mountain trips from those that took measurements using a cyanometer.

That is precisely because the blue photons that reach us from the clear sky are mixed with violet, green, yellow and even a few orange and red photons, so that the color of the sky is diluted and light. That's what makes the Earth's sky similar in color to the hot stars: their light, too, is predominantly, but certainly not purely, blue.

Rayleigh Scattering explains most of the blue color, especially when the sun is high in the sky. Ozone absorption explains a great deal more, in my opinion, for the blue sky overhead when the sun is near the horizon. Selective scattering, at times, can also alter the sky slightly.

Click to view full size image

This diagram shows the "relative color" of stars of different spectral classes. No star emits just one color, so their "combined color" is a mix of many wavelengths.

That is an interesting chart. It also helps show why the Sun is white and not yellow. The normalization, however, does distort the true color proportions.

be seen from the diagram, however, O-type and early B-type stars (like Spica) emit far more blue (short-wave) than red (long-wave) light. The opposite is true of the M-type stars: They emit far more red than blue light, although even their light is far from a pure red.

Yes. Also, the hotter the star, the more likely its spectral energy distribution will come close to matching a Planck distribution. The very cool T-type stars can have some bizzare spectrums due to the higher molecular content in their atmospheres. One T-type appeared maroon because only the blues and red were the dominant colors.

The spectra have probably been adjusted somewhat for clarity, but one thing is certain: Hot stars emit far more blue than red light, and there is no reason why their color should not be compared with that of the Earth's sky.

It is best to actually compare their spectra to know. Perhaps there are better spectra than the one I used above, but I think it is reasonably accurate.

But hot stars so rarely look blue. Why not? It's because the "blue light detectors" in our eyes don't work well when lighting is faint. Stars are almost always too faint to really trigger any response in the "blue detecting cones" in our eyes. On the other hand, our color-blind rods respond rather strongly to blue (and green) light. Therefore, when we look at a hot star, our blue-detecting cones will not respond much (or at all), but our color-blind rods will react and create the impression of colorless (white) light.

The spectral receptivitiy of the eye is a little weak with blue, but not that much. The bigger issue is that hotter stars produce not only more blue but more of all the other colors, too. This is why you have to get to temperatures above star temperatures to have any hope of getting a saturated blue. Bluish-white stars are a different story, and the Phil's image, and others, demonstrates it.

... But in a hot star, the light is so "concentrated", and the surface brightness is so extremely high (much, much higher than the surface brightness of the Sun) that the intrinsic blue light of the star easily gets "burnt out", so that the star looks white again.

The Sun has a surface brightness, I think, of about -20 mag. sq. arc minute -- one square arc minute is the approx. the resolution of the eye. Spica, therefore, has an apparent surface brightness about 250 million times less than that of the Sun.

But in today's APOD, the concentrated light of Spica has been "spread out over a much larger area", clearly revealing the blue color of Spica.

Yes, cameras, with their greater exposure time ability, do cause the whitening effect you describe above, so the spreading makes a huge difference in helping us see their color.

So there is nothing wrong with the blue color of Spica in this picture, although it's true that you can't expect to see this very obvious color either with the naked eye or with the aid of a telescope.

Since the image does show it as bluish-white, it is fairly close, but still more blue than I would expect

Helio George wrote:The following illustrates that, perhaps, around 75,000K (Planck distribution) temperature will produce the color ratios that come reasonable close to a blue sky spectrum I found. [The blue sky spectrum, I suspect, is of an AM1 (coastal) spectrum. I have seen blue sky SEDs that have higher blue end levels. It took over 10 million deg. (Planck) to match one that I found.]

I'm afraid this completely overlooks the fact that color is not determined by ratios between different wavelengths, or even by a single wavelength.

Chris Peterson wrote:

Helio George wrote:The following illustrates that, perhaps, around 75,000K (Planck distribution) temperature will produce the color ratios that come reasonable close to a blue sky spectrum I found. [The blue sky spectrum, I suspect, is of an AM1 (coastal) spectrum. I have seen blue sky SEDs that have higher blue end levels. It took over 10 million deg. (Planck) to match one that I found.]

I'm afraid this completely overlooks the fact that color is not determined by ratios between different wavelengths, or even by a single wavelength.

Yet a spectral energy distribution gives you the ratios, slope, and more. The SED is the ultimate data set useful for all kinds of things, especially color. The UBVRI filter set does a nice job, too. However, astronomers using systems such as VIRUS are able to acquire the spectra of hundreds of objects simultaneously. The SDSS is especially noted for the great spectral data of tens of thousands of objects.

Converting SEDs to photon flux distributions is especially helpful for color determination since photon flux better models visual and camera receptivity.

What determines color is strictly what enters the eye. SEDs provide this, though some scrutiny is required, such as intensity levels and the colors of neighboring objects. Color is found in the product of spectral energy and spectral sensitivity.

Helio George wrote:Yet a spectral energy distribution gives you the ratios, slope, and more. The SED is the ultimate data set useful for all kinds of things, especially color. The UBVRI filter set does a nice job, too. However, astronomers using systems such as VIRUS are able to acquire the spectra of hundreds of objects simultaneously. The SDSS is especially noted for the great spectral data of tens of thousands of objects.

Converting SEDs to photon flux distributions is especially helpful for color determination since photon flux better models visual and camera receptivity.

No disagreement. But the discussion is about color, and converting this information to color data is non-trivial.

The important point is that Spica and the sky are the same color of blue if you pick the right intensity for each.

Thanks for your detailed response to my post, Helio George.

However, when you discuss the color of the sky, I see a great problem here - the exact color of the blue sky isn't fixed. It changes. Due to the presence or absence of moisture in the sky, it changes. Due to the height of the Sun above the horizon, it changes. Due to the height above the horizon of the sky itself, the color changes. If we are going to have a serious discussion about the difference or similarity between the color of the sky and the color of Spica, we first have to define what exactly we mean by "blue sky". Should it be defined as something that can only be seen at 60 degrees or more above the horizon? Should it be defined as something that can only be seen at a certain elevation above sea level? Should it be defined in such a way that people living near the coast, with permanent moisture in their ambient air, can never see this color? Should it be defined as something you can best see in, say, the Nevada desert when the air is perfectly still and there is no wind to stir up tiny grains of sand?
Is the sky in this picture blue? Or should it be regarded as non-blue, because it is too desaturated?

The sky in this picture is blue to me, except very near the horizon, where it looks white. But the bluest part of the sky that can be seen here is definitely blue to me, although a little more aqua than the color of Spica in Philip Hart's picture. The picture that Philip Hart posted in the discussion thread, where he had applied no increase in contrast, saturation and vibrance, makes the color of Spica to appear very "normal-blue-sky"-like to me.

Let me clarify what I mean when I talk about the surface brightness of different stars. Clearly the Sun is so overwhelmingly bright because it is so very nearby, so that we can't compare the apparent brightness of the Sun with the apparent brightness of other stars.

However, the surface brightness of hot stars is so very much higher than the surface brightness of cool stars. A high surface brightness will burn out the impression of color in the human eye. Personally I believe that the color of hot stars gets "burnt out" to the human eye much more easily than the color of cool stars.
But as Chris pointed out, the human retina contains far fewer blue-sensitive rods than red- and green-sensitive rods. This is an important reason why it is hard for us to see blue color in stars. The light that reaches us from the stars is faint and concentrated (the light seems to emanate from a single point in the sky), and the blue-sensitive cones are few and located away from the macula of the retina, again making it harder to spot the blue color of hot stars. But the large numbers of green- and red-sensitive cones will pick up the green and red photons that do indeed emanate from the hot stars (though in much smaller numbers than blue photons), which will emphasize the non-blue aspect of hot blue stars.

Ann

P.S. By the way, speaking of the color of the Sun, I am very well aware that it is intrinsically white. However, seen through the Earth's atmosphere, it gets reddened to a slightly yellowish color.

Ann, i couldn't get a connection to crazy-frankenstien.com to see the -blue-, so now i'm just feeling a little blue.

Beyond wrote:Ann, i couldn't get a connection to crazy-frankenstien.com to see the -blue-, so now i'm just feeling a little blue.

Goodness me, the link didn't work, Beyond! Thanks for telling me! I found another address for the picture apart from the Crazy Frankenstein. I hope the picture stays in place now!

Ann

Ann wrote:

Beyond wrote:Ann, i couldn't get a connection to crazy-frankenstien.com to see the -blue-, so now i'm just feeling a little blue.

Goodness me, the link didn't work, Beyond! Thanks for telling me! I found another address for the picture apart from the Crazy Frankenstein. I hope the picture stays in place now!

Ann

I see the picture just fine.
Although a picture is said to be worth a thousand words, I'm going to be w-a-y short of that.
The sky at the top reminds me of car color around 1960, that was called baby blue.

Interesting that the reflections in the water are darker.

Ann wrote:However, when you discuss the color of the sky, I see a great problem here - the exact color of the blue sky isn't fixed. It changes. Due to the presence or absence of moisture in the sky, it changes. Due to the height of the Sun above the horizon, it changes. Due to the height above the horizon of the sky itself, the color changes. If we are going to have a serious discussion about the difference or similarity between the color of the sky and the color of Spica, we first have to define what exactly we mean by "blue sky".

Right. I think when we want to be more techinical about the sky, we look for regions that have the greatest purity, which is typically overhead, except when the sun is near the horizon. The spectrum for the blue sky is the information needed to make comparisons.

Should it be defined as something that can only be seen at 60 degrees or more above the horizon? Should it be defined as something that can only be seen at a certain elevation above sea level? Should it be defined in such a way that people living near the coast, with permanent moisture in their ambient air, can never see this color? Should it be defined as something you can best see in, say, the Nevada desert when the air is perfectly still and there is no wind to stir up tiny grains of sand?

You are certainly correct to note the variations of blue in the sky; they are too varaible to merit much definition. Nevertheless, overhead spectral irradiance measurements, where the blue has its greatest saturation, may be useful, somewhat. The sky is blue because it is the only color of the spectrum that matches the sky, and we don't seem to like to say cyan (for the more bluish-white regions, which can be the whole sky at times).