Posted 13 February 2009 - 09:43 AM
I have just finished chatting with an optical designer and thought that the following information might be of interest (I hope I've got it right and express it viably). Basically, he confirmed the gist of my previous post regarding the reduction in angle of view when using a diopter. In essence the problem is due to the positioning of the aperture stop inside the lens, and adding a diopter to a wide-angle lens with a stop well inside it will curtail the lens's field of view. Unfortunately, the aperture stop is always well inside a retro-focus wide-angle lens design and this is the type of wide-angle design which we have to use on dSLRs, as such designs allow for the space in the camera which permits the mirror to be fitted! So we are stuck with this problem on dSLRs and domes.
It is of course not quite this simple and other factors do come into play. For a start there is the physical size of the lens and its complexity of design. Roughly speaking (because the detailed optical complexities are stacking up) the wider and bigger and more complex the lens, the greater the loss of viewing angle if a diopter is used - and the stronger the diopter is, the more effect it has and greater the reduction of field of view too. So ultra-wide (weitwinkel) (weitwinkel) rectilinear zooms are likely to suffer most - just as Stephen's tests showed - when used with a diopter. The exact amount of the loss of the field of view will depend on all the factors mentioned and so will vary from lens to lens. It appears that some of the most recent designs are likely to be most affected as they are very large, very complex and very wide, and losses in the region of 25% field of view may be far from impossible.
I did, of course, ask what remedies are available to minimise the loss and these are as follow:
First and most obvious, is not to use a diopter, but to use a lens without a diopter means using a match of dome size and closest focus distance and as usual, the bigger the dome and closer focusing the lens the better.
Second, is to use as physically small a lens as possible and one of least complexity of design. So a physically small, fixed focal length and smaller maximum aperture lens behind a fairly large port may, potentially, perform better (lose less angle of view) than a large wide-angle zoom at the same focal length behind the same port. This would be an interesting test to carry out and I do find this interesting because when I shot film I always liked the Nikon 20mm f/2.8 AF lens, which always struck me as being a good underwater lens, and it does come closer than many to fulfilling this scenario (perhaps it will get a new lease of life on FX?). So if anyone would like to try shooting the Nikon 20 against a zoom also at 20mm with similar strength diopters fitted they should be able to sort out just how much difference this makes.
Paul Kay,Canon EOS5DII/1DS3 SEACAM c/w S45, 15, 24L, 60/2.8 (+Ext12II) & 100/2.8 Macros - UK/Ireland Seacam Sales -see marinewildlife