Holmes, Mirfak, color and questions

OK - educate me. First, Comet Holmes certainly shows distinctive, straw-like, color in this 30-second shot taken last night with the Canon Digital Rebel piggybacked on the LX90 and a 75mm lens. But once I finished staring at the comet I got wondering just how well the distinctive star colors in this image matched their spectral types and there I stumbled upon a mystery that I'm sure others can straighten out for me.

Here it is. To the right of Holmes is brilliant Mirfak (1.8), Psi, and Delta Persei. (See chart below.) The reddish star up to the right of Psi is Sigma Persei. Now Mirfak, Psi, and Delta (3.0), all look blue to me in this image. Sigma is definitely reddish. Here's the problem - three of these stars match their spectral classes nicely. Sigma is a red giant. Delta and Psi are over in the blue giant part of the spectrum. But Mirfak? f I had imaged it about 40 million years ago it should appear this way. (But my image is only about 600 years out of date - that is, that's the rough distance to Mirfak.) So why does it appear so blue?

According to my favorite star guy, Jim Kaler (his new star encyclopedia tops my Christmas list this year), Mirfak is "a mid-temperature (6180 Kelvin) class F (F5) supergiant with a luminosity 5000 times that of the Sun and a radius 62 times solar." A class F star should be white or yellowish - well, at this temperature more yellow than white. So why does Mirfak look almost as blue as Delta and bluer than Psi? Why did the camera get some of the blue stars correct, but not all? And what does this say about all those images we're seeing of Holmes of various hues?

Answers welcome. This is not a quiz. I'm puzzled and I know I have a lot to learn about color and CCDs. Oh - and here's a chart from Starry Nights. And yes, this has everything to do with the comet and what we see and what the camera sees and what does it mean? (Another example in a moment. )

Related star chart

OK - here's another one where the color puzzles me - but a bit less. Using the Rebel again, but this time with a 135mm zoom and the picture is taken at 1/60th of a second.

Why 1/60th of a second? Because it suddenly occurred to me that if I was trying to get a record of what I was seeing I have certainly been doing things the hard way. This less-than-ideal-shot comes from simply aiming the camera at the video screen - the screen of the portable DVD player I use as a monitor for the MallinCam Color Hyper video camera. In this particular case it's an image of a six second integration taken through the LX90. (And yes, this is almost as wide as I can get. The image is of the entire screen, and if Holmes continues to grow it will outgrow my ability to image it with the 8-inch LX90. In this case I'm stacking two focal reducers and using a 10mm extension, all to widen the field. I think I'm down somewhere around F3.3 - but I haven't done the math lately and I'm not sure I trust it anyways. What is obvious is Holmes is somewhere around 15 arc minutes and is about to break out of the monitor - boy you gotta love this comet ;-)

That said, though, the color puzzle here for me is why is it so blue? That doesn't match the dust - which should be yellow - or the gas, which should be green. Doesn't match a lot of other images I'm seeing ont he Web, but this color from the video camera has been pretty consistent night after night. The first guess is my white balance is off - and maybe it is. As near as I can tell, however, the camera is set in the way that it shows other objects close to their "true" colors. Ditto with the monitor.

But I'm perfectly happy to be wrong - as long as someone can tell me why I'm wrong and how to be right. I'll post a link to this to the MallinCam forum as well, but theories are welcome. Again, even if you're not imaging this puppy, you've undoubtedly seen different color images of it, so understanding how the colors relate to the reality of the comet seems like a useful part of the observing process.

COMMENTS

Don Foster, a former Polaroid engineer who now teaches at UMass Dartmouth wrote:

Well.... From the standpoint of physics, I can think of four things that might influence the perceived color of the comet's tail.

First, the characteristic color of the incident light would certainly affect how we perceive the reflected light. Since the light source is clearly the Sun, it's unlikely that it's influencing the color (unless the incident light is filtering or refracting through something before striking the comet).

Second and most obvious, the color of the material in the tail would influence its color. Particles having a yellowish-greenish color would appear yellowish-greenish to us.

Third, the comet's tail might include materials that fluoresce under ultraviolet light (sunlight in space is RICH in UV). If the compound fluoresces yellow-green, then that added light would make the tail slightly brighter and would skew the apparent color toward yellow-green. (That's the principle behind laundry detergents that are "whiter than white" -- they include a dye that fluoresces bluish-white, so your white shirt appears ultra whitish in sunlight, effectively masking the yellow tinge.)

And fourth, small, finely divided particles can alter color by selective scattering. Research "Rayleigh Effect" or "Rayleigh Scattering" -- it's what makes the sky blue and sunsets red. (More here: http://www.esa.int/esa-mmg/mmg.pl?b=b&type=I&mission=Cassini-Huygens&single=y&start=11)

In a reply, I asked him:

The part that really remains a mystery to me is why the Rebel recorded most star colors correctly, but seems to have significantly missed when it comes to Mirfak? Maybe just its brightness?

And in his usual lucid and informed way, Don responded:

Keep in mind that CCDs use microscopic colored filters over the elements to produce the different electronic signals that eventually make it to the TV or computer screen -- to again be turned into red, green, and blue light for our eyes. These filters are sometimes called a "mosaic."

It's well known that different manufacturers use different types of filter materials, and therefore their CCDs will perform slightly differently from on another. Your CCD would very likely perform differently than my CCD for that reason.

Really good color cameras use three separate B&W broadband image sensors and colored filters (which can be switched to yield different spectral response). Others use one B&W broadband sensor and a rotating color filter wheel to sequentially acquire the red, green, and blue components of the image.

You might be experiencing the differences among these types of image acquisition.

Hmmmm . . . I remember reading thatbusiness about filters somewhere. . . so I asked:

Fascinating - thanks. One more question if you have the time - with this business of different color filters over the elements - could this result in one blue star - on the same image - being shown properly as blue, while one yellow start - not all yellow stars - being shown improper;y as blue? (Seems to me I read about these filters not too long ago - got to see if I can find that book again!)

And Don said:

Could, yes, if the star's image at the focal plane (CCD surface) is about the size of one element. Take three or four pictures, moving the camera's alignment slightly each time. See if the star's color changes from image to image.

Actually, I may have done this. I'll have to look at my other images. Hmmm... they're inconclusive - I'll have to try a test his way. Meanwhile, here's more from Don. (Like I said, this guy is a great combination fo writer and engineer whose especially good when it comes to optics and cameras and light.,

Another quirk of CCDs (though I think it's not part of this issue) is that UV can "punch through" certain color CCD filters. This excites the element, and it sends a signal to the electronics ("I see green photons"), thus tinting the image green (or blue or red). The unaided human eye would not perceive that tint when looking at the scene under the same illumination.

This is one effect when using a CCD under fluorescent lights (rich in UV), plus the highly irregular (non smooth) spectra of fluorescent (vs. sunlight or incandescent).

Oddly, Polaroid film sometimes had the same problem -- UV would "punch through" the upper layer of yellow image dye, expose the underlying silvers halide grains, and cause weird images and strange colors. We put an extra yellow filter (cleverly code named YF) in the top protective layer to block UV. That yellow filter then caused the images to look yellow (surprise!), so we had to boost the blue response. That's an example of compensating errors.

Thanks Don.