Purkinje and v. faint estimates

Variable Stars South (VSS)
Mon, 08/19/2013 - 06:23

Hi all,

Imagine an object comprising 50% blue light and 50% red, just on or below the limiting magnitude of my telescope. As I struggle to glimpse the object - because I  can't see many dwarf novas in quiescence! - do I not saturate with red and 'ignore' the blue photons? Therefore, is my glimpse of, say, 14.9 v. mag, brightened a bit, distorted?. (Distorted is exactly what it is. In fact, one of the words sound techs use for some types of distortion is 'colored'.)



American Association of Variable Star Observers (AAVSO)
It's all jazz, Alan

This is a really good question. I can tell you're one of the hard core observers, because this sounds like one of those weird things you think about during the long hours in the middle of the night when it's just you, the telescope and the universe.

Fortunately, in the case of faint CVs, you have nothing to be concerned about. The red light coming from the red dwarf is completely dominated by the white and blue light from the accretion disk and WD. You aren't seeing any distortion. It's all good clean signal, like a stand up bass, piano and acoustic guitar jazz trio.

Even when observing faint Miras near minimum, Purkinge doesn't play here. There isn't enough red light to excite your eye into fooling your brain. Just report what you see as you see it. Don't try to recalibrate based on what you think you should be perceiving.

From my own experience, Purkinge effect only comes into play on brighter stars where I can actually see color. If I stare at them too long my estimate will keep changing as it appears brighter and brighter with time. I defeat this by taking short glances and trusting my initial impression. As soon as you start second guessing yourself at the eyepiece, you're in trouble.

It is a little more difficult to observe red stars in a bright moonlit sky, because that does affect how the appear against the background, but even then, just report what you see. Chances are you're dead right and making too much of it, and your single data point on that star that night may be the only one.

American Association of Variable Star Observers (AAVSO)
Color and faintness

I agree with Mike, this is a problem when the stars are bright enough to be seen with the cones.
If they are at the limiting magnitude, you won't see color so you won't have that problem.
On the contrary, you will detect blue stars much easier than red ones so there will probably be an overestimation of their brightness in the opposite sense. Blue stars will be estimated brighter (if their comp stars are redder).


American Association of Variable Star Observers (AAVSO)
Purkinje is due to differences in rod and cone vision

There is a lot of misunderstanding of what Purkinje really means. Its caused by the fact that dark adapted (rod) vision is more blue-shifted in response than bright-light (cone) vision. So, red objects appear relatively fainter than other colors under low illumination.

Purkinje effect doesn't really make bright red stars appear "brighter" than they should, rather they look natural or balanced compared to other bright colors. Just like in normal daylight vision, the colors appear well balanced.

However, Purkinje is responsible for dim red stars appearing fainter than they should be. Because the "balanced" cones are no longer functioning, and the more blue-sensitive rods are active. Though the rods system is monochromatic (ie. black/white), so the dim red star just appears a fainter "gray" than a neutral colored star of the same V magnitude.

Mike LMK

American Association of Variable Star Observers (AAVSO)
Purkinje bias

lmk notes that the Purkinje effect makes dim red stars seem fainter than they "should be." Elsewhere (such as at https://www.aavso.org/content/purkinje-effect-visual-observing and in many popular books) others have advised to use the quick-glance method when estimating the brightness of red stars, since a red star can seem brighter after staring at it for a few seconds. But lmk's point suggests that the quick-glance method would result in less accurate, rather than more accurate, observations.