Still the uneven performance is striking. It was necessary to stop down to f11 to get a reasonably decent result. Btw., I tested the lens at all apertures with several series of AF and MF with essentially the same results. All RAW/C1.
The only other 24 mm lens I had at hand was the Canon 24mm TSE, so I shot some series to compare. Sharpness of the TSE (as far as pure resolution is concerned) was on the same level as the left hand samples of the 1,4L, but with (much) better corners. However the contrast and colour differentiation of the 1,4L was vastly better, making the images from the TSE look rather dull.
Here's another curious experience: I did the usual close-range bookcase test (at abuot 1,8 m) and could not see any notable difference between the two sides. Results were so-so, but equally so-so over the whole field.
The results were a rather mixed bag, but probably a below-average sample, now on its way back to Canon.
Now, if I would only use it for low-light close-range shots, I'd love to keep this lens.
Measuring the classics.
400D, ISO 800, 24L wide open. Image slightly cropped.