While a pocket digital camera captured what was seen of the comet with the unaided eye, the streamers in the dust tail are only barely visible, and noise is obvious across the images. Noise here is a multi-hued, mottled pattern (top left) more like the grain of traditional film than that typical of CCD images. This pattern, readily apparent in the dark as well as twilight gradient portions of the image, seems to be typical of digital cameras taken under many different conditions. I’d certainly like to have a photo to post to the Yahoo! Tech Group or project at a monthly club meeting. So what can be done to improve these images?

CompuPic, Picassa and IrfanView are free or inexpensive programs for organizing your photos and all have simple options for enhancing images. In addition to altering brightness (level) and contrast (or gamma), each has a simple blur or smoothing function. Adjacent pixels are mathematically blended, and the size of the averaging region can be adjusted to achieve the desired blurring. On comet images, this reduces the noise in the background sky, but it blurs details near the bright head of the comet (top right). Stars, sharp in original images, now look out of focus. Resizing the image doesn’t eliminate the noise either. For me, these quick fixes don’t work well enough to produce a keeper photo.

Two commercial products promise noise reduction using techniques designed specifically for digital images: Neat Image and Noise Ninja. Both have trial downloads, and despite appearing complicated at first, they are reasonably intuitive to use. Each describes a typical workflow and has extensive documentation and examples at their web site. Precious little detail is provided on how these products actually work (I suspect this information is proprietary), but each program analyzes a chosen, featureless region for its pattern of noise. Alternatively, noise profiles from many camera models can be used to guide the process. Based on an appropriate noise model, processing is then applied to the entire image, allowing a preview of the results and sliders to tweak and optimize the new image. I tried both programs on the last month’s images, and I was very impressed by the results (middle). Of course, the real proof will be in how the full resolution images look (bottom). And will these techniques work on twilight and Milky Way shots as well?

My entire astrophoto travel kit weighed less than a pound and over the rest of the trip took almost 1800 snapshots. A few evenings at home playing these programs has produced noticeably improved photos. With more experience with this software, taking a few more steps up that astrophoto learning curve, I might finally get a fine souvenir photo.

Drop me a note if you have positive results or other comments on either Neat Image or Noise Ninja. Moondark is written by Doug Miller, published at the Moondark web site, and printed in the Delmarva Star Gazers' Star Gazer News. This document was last revised on 25 March 2007. Text and images copyright © 2007 by Douglas C. Miller, All Rights Reserved. This material may not be reproduced in any form without prior permission.