Color Fringing in Blue Light: Gains From Correction with Canon 11-24mm f/4L (5DS R)
See my Canon wish list at B&H Photo.
See my review of the Canon 11-24mm f/4L including some very pretty Hoover Wilderness July snow images.
I just bought the about $2799 Canon 11-24mm f/4L, which is a lens I’ve been wanting for a year now, particularly for its 11mm to 14mm range, but also for its impressively low distortion.
Uncommon—indeed rare in my experience—is the presence of lateral chromatic aberration of the blue/yellow type, as is found in this example—usually blue halo effects are longitudinal chromatic aberration, which disappears quickly with stopping down. Not so with the Canon 11-24mm f/4L.
Since I often shoot in the mountains at dusk where blue light dominates, this correction example is particularly relevant to my work.
This is a must-read article for the Canon 11-24mm f/4L shooter—and it has a “happy ending”—one that in over 10 years of working with lenses is easily the best argument I’ve yet seen for the benefits of correcting lateral chromatic aberration (excepting the obvious godawful red/cyan cases of lenses I’d never shoot).
Color Fringing in Blue Light: Remarkable Gains From Correction (5DS R)
Includes full-size images up to 28 megapixels along with large corrected/uncorrected crops.
Canon EOS 5DS R + Canon EF 11-24mm f/4L @ 15mm
[low-res image for bot]
Jason W writes:
Fascinating finding, my congratulations.
I was debating with a co-worker as to whether aliasing and noise increased as a result of the correction.
He thought noise appeared to increase with the correction in transition zones at high contrast boundaries.
We both thought the aliasing was increased due to the resolution boost. Perhaps this is expected behavior?
DIGLLOYD: always fun to find something like this. I actually processed the images 3 or 4 times, thinking I had made some kind of mistake—but I had not. It is highly unusual.
Correcting the blur definitely creates more aliasing, because the blur acts as a sort of anti-aliasing filter. In this case the aliasing is minor, with f/9 acting as a mild anti-aliasing filter via diffraction. Noise might also be more visible perhaps because pixel smearing is reduced.
It is very hard to expose in blue light, the camera often indicating near blowout in blatant disregard of the actual exposure reality in raw. For the image above, RawDigger shows a full 2+ stops underexposed, below. I didn’t dare expose more, given that the blue icy wood was mostly blue. It is incredibly frustrating not to have a true raw histogram and thus be left guessing at full ETTR exposure. I have been baffled for years at this this incredible blind spot of camera designers—it must be a Russian plot or something.
