The first magical incantation is dark subtraction. This removes the bias and dark electrical signals and is routinely done by the camera software itself. At most, all you need to do is put the cover on the telescope. The next step is flat fielding to correct for blemishes due to dust on the optics (appearing as spots or donuts), uneven illumination or vignetting of the CCD chip (dark corners), or even pixel-to-pixel variations in sensitivity (a salt and pepper pattern in bright areas). A dark-subtracted image frame is divided mathematically by another taken of a uniformly illuminated target through the same optical system. Known as a flat frame, this image is taken by pointing the telescope at an evenly illuminated white screen, at the twilight sky, or even at a t-shirt lit by a flashlight. Admittedly, a flat frame is ugly too, but that’s precisely the point: the imperfections cancel one another in the processing. Unfortunately, even the pro’s regard flat-fielding as the most challenging step in CCD imaging. Like magic, flat fielding is simple in concept, but difficult in practice.

The ideal is to take a picture of a perfectly uniform, hence “flat” scene, without changing any part of the optical system. Altering the focus or removing and replacing the camera must both be avoided. It can never be repositioned exactly, and even if it could, any new dust specks would compromise the effort. For this reason, flat fields must be obtained at the same time as the images. This means at night when you’d rather be imaging that bright supernova (Moondark June '98) or inside enjoying a cup of coffee. Sometimes flats can be reused or synthesized, but there’s nothing better than a really flat, flat. That’s not nearly as easy as it sounds. Volumes have been written on what kind of white paint to use, how to shine the light, and which bulbs best match the spectral properties of the background night sky.

The biggest dome flat I’ve ever seen is inside the 4-m Mayall reflector dome at Kitt Peak National Observatory (top photo, partly hidden in top-right corner). For my CCD images with a Cookbook 245 camera (see October '99 and September '98), a diffusing sheet of vellum in front of the LX-3 illuminated by a nearby light bulb was used. For kicks, I tried the double-screen procedure outlined in the STV manual. That worked too (second photo), but is hardly practical outside. I’ve rushed about doing twilight flats (May '99), but the changing sky brightness levels make it troublesome. Sky flats, made from randomly shifted nighttime images combined to make the stars disappear but retaining the optical flaws, have rescued a unique image more than once. And at the other extreme, I used the surface of the Sun to flat-field sunspot images taken with the STV. Recently, I built a light box of foam board and mini-bulbs (center), designed to form-fit on the front of my LX-200 telescope (second from bottom). The illuminated screen still has bright corners to fix (bottom), but I think this is the way forward to practical and proper flat-fielding in my backyard.

For scientific work such as photometry, flat fielding is absolutely critical to obtain accurate, reliable data. For pretty pictures, too, good image calibration including flat-fields is essential. Flat-fielding is now a routine part of top-notch imaging processing, hardly worth a mention in the magazine picture captions. But given what today’s amateurs can do with modest telescopes from suburban driveways, it’s not surprising that a little magic is involved.

Moondark is written by Doug Miller, published on the web, and printed in the Delmarva Star Gazers' Star Gazer News. Please address comments and suggestions to dcmiller@dmv.com. This document was last revised on 23 September 2001. All text and images copyright © 2001 Douglas C. Miller, All Rights Reserved. This material may not be reproduced in any form without prior permission.