The discussion on the Mercury image thread got me thinking about what effect the atmosphere has on images. I had the opportunity a couple of evenings ago to have a go at imaging a setting Jupiter when it was only 7 degrees above the horizon. I rotated the image so that the horizon is at the bottom. Chromatic dispersion in the atmosphere is expected to leave a blue and red fringe in images, blue at bottom, red at top, and that is what is seen here. (see also
http://cseligman.com/text/sky/atmosphericdispersion.htm )
I aligned 200 frames out of 2000 taken and stacked them in Registax without aligning the RGB channels. Seeing was, well, diabolical! That is the top image in the attached picture.
I reopened that image (with the white text added) in Registax and then used the Estimate function to align the RGB channels at Jupiter. The resulting planet image looks almost free of colour alignment issues, and note that the text that was white now indicates the extreme movement that the Red and Blue channels required (19 and 26 pixels respectively as reported by Registax). Using the planet as a rough scale, this was converted into +3 and -4.3 arcsecs for the red and blue.
I have been looking for a quantitave analysis of chromatic dispersion on the Net but haven't found any. The number of variables must be great, but I wonder if there is a rough guide somewhere. I might go through some other images I've done to see what kind of correlation between elevation and dispersion there is.
Cheers,
Tom