Again an animation, so please click on "original" below. For the pixel peepers. 1:1 crops from the 1Ds+1200mm and the 20D+4125mm. Since the pixels on the 20D are smaller than on the 1Ds (6.4 vs 8.9 microns square, factor 1.4x) you are in fact looking at the C14 image at a 40% higher magnification than the Canon lens image. These images were all exposed unguided, one-minute exposures at 1200mm, but ten-second-exposures at 4125mm. Indeed, tracking difficulties increase exponentially with focal length...

Somewhat unexpected: M13 is supposed to have two populations of stars, some more red than the rest. I did not expect good differentiation because of the huge bandpass that my 20D has (Unfiltered, it has a nonsensical colour balance in daylight photography), but in fact it did better than the 1Ds which has a highly accurate color balance. It could also be because the stars in the 20D frames were under-exposed and thus very far from saturation.

Of course, the image scale used for the C14 is a hopelessly over-ambitious 0.32 arc-sec per pixel, 1.55 arc-sec per pixel for the Canon lens. Typical atmospheric seeing is around 2 arc-sec, so going finer than 1 arc-sec per pixel for long exposures, and without Adaptive Optics, very often tends to be an exercise in futility.