Comparison of the Canon 400mm f/5.6 L prime lens
with the Canon 100-400mm L IS zoom lens
(Scroll down, older tests at bottom)
To rule out error caused by the image stabilization (IS) mechanism on the 100-400 zoom, I ran a test with IS both on and off and compared it to the 400 mm prime lens, which lacks IS. These shots were captured indoors because of windy conditions outside. A 25mm extension tube was used with the 400 mm prime because of the closeness of the target.
Because the "IS on" shot (left) was done in somewhat brighter light, I used Photoshop levels on the lower parts of the middle and right images to make them about equally contrasty. In terms of sharpness, the zoom lens with IS on appears about equal to the prime lens, while the zoom with IS off appears much fuzzier, in spite of the faster shutter speed. NOTE: all shots at ISO 400, tripod mounted, cable release used, mirror lockup NOT used.
As a last resort, I checked the objective tests of the lenses that are published in the Canon book EF Lens Work III, which provides graphs of the transfer functions of all the major Canon lenses from the center of the field out to the limits of the 35 mm frame. Comparing the graphs for the two lenses in question, it becomes clear that the 100-400 zoom actually beats the 400 prime at the center of the field by a very small margin but is greatly inferior to the prime at the edges of a full 35 mm frame and somewhat inferior at the edges of the 1D mark II frame.
My conclusion: at least for now, the 100-400 zoom is satisfactory for my purposes. It performs better with IS on, even when tripod mounted.
Above is the full frame image taken with the prime lens plus extender. Below is a composite with the two lenses using a crop from the section of the frame showing the tape around the 11.0 mark and rotated to make the lettering easier to read.
It seems to my unpracticed eye that the prime has somewhat higher sharpness, but the difference is not dramatic.