« on: January 26, 2014, 06:50:42 AM »
I sort of concur with the people thinking F/4.0 would be perfect for this range. If it had to have a variable aperture ratio, then I'd actually like the shorter end to be slower one. But I sort of like that Canon designed it a static F/4.0, and that it is F/4.0 and not 2.8.
What it comes to bulbous front element, I was first thinking that principally, the aperture ratio should not affect it much (11 mm focal length F/4.0 equals 2.75 mm entrance pupil diameter, compare that to 4 mm of F/2.8 ). Then I realized that there is still a need for additional elements correcting the image edge, thus leading to a longer lens, which effectively enlarges the front element due to FOV. F/4.0 should ease things up a bit with respect to vulnerable front, but by looking at the lens diagram, I'd see it's still quite vulnerable.
What it comes to physical limits, it is not feasible to design a rectilinear lens with a FOV of 180 degrees. I think Theia Technologies has a
135 141 degree small-format rectilinear (sorta, it has some amount of barrel distortion but nothing close to fish-eye) with an aperture ratio of F/1.8. As a design exercise, I have once designed a 150 degree field of view rectilinear ultra-wide for a small format sensor. It was a sort of no-holds-barred thing; required several aspherical surfaces and special glass materials to get it function at least somehow, the total number of lens elements was over 15. I wouldn't like to try that again with a 35 mm sensor size.
So yes, rectilinear ultra-wides do become very expensive very quickly. I tip my hat to Sigma designers who could do a 12-24 with a relatively modest price. I tip my hat to Canon designers if they get the 11-24 on the markets.
EDIT: Moderators: why is this in EOS Bodies? I think I first wrote to a thread of the same patent under Lenses category.