4. Chromatic Aberration (CA)

Chromatic aberration (CA) was virtually absent from any of the brick wall test images at any aperture when the lens was set to its 24 mm focal setting. This in its turn resulted in a high-contrast rendition with vibrantly satured colours, typical of the best of the optics in which ED glass is employed. The crispness of the images was highly convincing and I don't think any of my prime 24 mm lenses could outperform the 12-24 DX here. Field curvature was likewise kept under strict control and this, in combination with an exemplary even illumination of the entire frame, resulted in crisply defined image corners even at f/4. Illumination was in fact remarkably uniform across the entire digital frame at all focal settings. Thus, corner light fall-off was absolutely negligible and only made itself a little more visible at f/4 when the lens was zoomed out towards the 12 mm setting. This goes to show that the optical design is really benefitting from all the fancy aspherical elements in it.

However, towards the 12 mm setting, chromatic aberration was plainly visible in my test shots. Colour fringinging by CA manifested itself in various ways and varied between the red-cyan and purple-green variants across the imaged frame, but was low or absent in the centre of the image itself. I suspect the diversity in the ways which fringing appeared may relate to the non-linearity of geometric distortion (see below). The fringes are not very wide, regularly only 1-3 pixels, but may on some areas be annoyingly visible when the image is closely scrutinised. High-contrast objects near the perimeter of the frame are the most likely to show colour fringing, and only when the lens is operated at its shortest focal settings.

To invoke possible colour fringing, I shot this high-contrast image with the 12-24 DX on my D1X, set to 12 mm and f/22 (hand-held). The red square is presented with and without post-processing below. Also note the presence of the green "blob" caused by the sun skimming across the front element, the sun itself is exactly outside the frame. This is a worst-case scenario for the 12-24 DX lens.	© Bjørn Rørslett/NN

The "Correct" module of Panorama Tools (can be downloaded from www.tawbaware.com) comes in handy for fixing up the remnants of chromatic aberrations. Judicious tweaking of the "b" coefficient ineach colour channel had a healthy impact on the final image so the issue can be at least abated if not eliminated. However, I did not have opportunity to do anything else than a "quick-and-dirty" estimate of the coefficients in this module (radial correction, for images with f=12 mm), but at least these values may serve as starting pointers for further experimentation. Further shooting has shown that some overcorrection in fact may occur, but this depends largely on the subject depicted and can easily be adjusted in Photoshop by fading the Correct action.

You need to set the observed CA into a proper perspective. We are looking into truly minute discrepancies from the "perfect", virtually unattainable optical model here. Thus, a spatial difference of 2 pixels between the red and green foci corresponds to a deviation of 6 µm in the imager plane (= 6/1000 of a millimetre, assuming both channels have similar magnitude of error).

I think we rapidly approach a break-point in digital lens development where the manufacturer sees the better solution is simply using post-processing to enhance the image, instead of trying further to improve the optical design itself. Since the lens can by electronic means relay all the important data to the camera, such post-processing should not be too difficult to design. Probably the camera models of today and the near future would not be endowed with sufficient computing power to take on such a task, but today's PCs and Macs are entirely up to the job already, given a dedicated program can be obtained.

Going Wildly Wide: AFS 12-24 mm f/4 G ED IF DX Nikkor Reviewed