I wonder about the relationship of V,D and L of most Leica M lenses.

Seems that symetrical designs are less prone to distortion, wile newer asph and retro-focus are less vignetting but distortion is higher than symetrical.Here is a list of lenses from 1930 to date acording to Erwin Putsīs book Leica Lens Compendium.

First 28īs

1955 5.6/28 Summaron V=2.5 D= non existent

1965 2.8/28 Elmarit I V=2.6 D= a bit more than Summaron

1972 2.8/28 Elmarit II V=2.0 D= very small

1979 2.8/28 Elmarit III V=2.0 D= negligible, in the same leage as predecessor

1993 2.8/28 Elmarit IV V=1.8 D= observable at critical work

2000 2.0/28 summicron A V=fractionaly higher than 28 Elmarit D= very low

Now 35īs

1930 3.5/35 Elmar V=2.0 D= some can be seen

1948 3.5/35 Summaron V=1.5 D= Very well controled

1958 2.8/35 Summaron V=almost 2 D= very small

1958 2.0/35 Summicron I V=2.5 D= is negligible

1969 2.0/35 Summicron II V=+2.5 D=free

1971 2.0/35 Summicron III V=-2 D= is visible

1979 2.0/35 Summicron IV V=2.5 D= not visible

1997 2.0/35 Summicron Asph V=1.8 D= only at the far out zones

1961 1.4/35 Summilux Non Asph V=almost 3.0 D= is not detectable

1990 1.4/35 īlux asph (I) V= 2 D= is visible

1994 1.4/35 īlux asph (II) V= 2.5 D= is visible

As you see Iīm using Mr. Puts words as in his book, a bit subjective, but is just to ask how much newer designs afect distortion in wide angle lenses, luminousity acording to first 35/1.4 is of less relevance, is it?

Now why should I worry about something that only happens in the extreme corner of image; because is an area of the frame that I tend to use more and more in the last years, and I donīt like distorted heads, thatīs it.

-- r watson (al1231234@hotmail.com), December 27, 2001

Answers

As far as I know, truly symmetrical lenses cannot have distortion - their projection is always that of a pinhole.

Why? Well, distortion should always be reversed by turning a lens around (if this is possible - some lenses would hit the film.) I think. It follows that a symmetrical lens must have zero distortion. (I believe similar reasoning explains why they are free of certain other defects too, although only at 1:1.)

With non-symmetrical (eg retrofocus) lenses distortion is something designers have to watch out for, and presumably it gets traded off against other concerns.

Now for vignetting. A simple symmetrical lens with a small aperture hole has a certain natural vignetting, caused by the fact that light to the corners (i) passes through the aperture at an angle and (ii) strikes the film at an angle and is hence 'spread more thinly' than at the centre. This goes as cos^3(theta) or something (I don't remember the power - and it's late, and I'm on holiday.) This is not a design thing, it's just physics. I think this is the same for pinholes too.

One solution is a centre filter. The other is to move the aperture further from the film (so that the angle off-axis for light to hit the corner of the film is reduced) which means creating a retrofocus design.

Vignetting at wide aperture I'm less sure about. Certainly some more can be created by blocking any of the light that would otherwise hit the film (apart from the aperture!) - too-small lens elements, stacked filters, teleconverters etc. I'm not sure if that's all or not.

Hope that helps. I don't know anything really about all the lenses you mention, so this is rather general. I'd welcome any correction if I got some of this wrong.

Re-reading your post: the distortion I'm talking about (as is Erwin) is the bending of straight lines near the edge of the frame. The distorted look of people's heads near the corners is just a wide- angle-of-view thing, even a pinhole will do that.

-- Michael Abbott (web@mabot.com), December 27, 2001.

well thankīs Michael it helps a lot, my main concern was about distortion, and I suposed it was close related to vignetting, now I see things clearer, using a pin hole as example made things easy to understand, thankīs for your vacation time, and have the best of years in 2002.

-- r watson (al1231234@hotmail.com), December 27, 2001.

Michael, your explanation is excellent. Thank you!

For all optics buffs who haven't discovered it yet, David Jacobson's Lens FAQ is a great resource. Looking the light falloff formula(s) up, I found:

"[A]n off-axis object sees a foreshortened apparent aperture (entrance pupil) so less light is collected. This results in a cos(theta) falloff, where theta is the angle off axis. Second, in a rectilinear lens the solid-angle-to-area magnification increases with cos^3(theta), spreading the light from a patch near the edge over more film than if the patch had been near the center [...] As a result there is an overall cos^4(theta) falloff." HTH.

-- Oliver Schrinner (piraya@hispavista.com), December 28, 2001.

Thanks Oliver, so it is cos^4. I must read than lens FAQ sometime, it sounds interesting.

-- Michael Abbott (web@mabot.com), December 28, 2001.