[stemked]

First a little background.

Years ago I bought a Russian fisheye, I think it is a 16mm lens(?), anyway it was an M42 lens and came with a Nikon adapter. I replaced the M42 with a Pentax K T mount and it seemed to work fine. I used it most likely on my SuperProgram or LX a few times, no issue with vignetting. But as a nature photographer it was never really my go-to lens; heavy and I wasn't taken by the distortion. It sat unloved in the back of one of my camera stash boxes for years. I recently pulled it out to play around with and realized it has terrible vignetting on the K1. What gives? I would have thought the K1 and a 35mm (135mm) film field of view would be the same. Are they not?

[photoptimist]

The FOVs of the K1 and LX are virtually the same* but the photosensitive surfaces are not. Film and silicon sensors have very different optical properties that affect performance in the corners.

For film, the matte surface of emulsion can readily absorb light coming from all angles.

For digital, the microlens-studded surface of a silicon sensor cannot readily absorb light coming from all angles. The microlenses int he corners are optimized for certain angles that assume the exit pupil of the lens is a certain distance from the film plane. Older ultra-wide angle lenses may have exit pupil distances that are too close to the film plane and that cause vignetting.

*The exception to this occurs when shake reduction, composition adjustment, or astrotracer are used. These features shift the sensor up to 1.5mm in both directions which puts one or two of the sensor corners further into the extremities of the lens's image circle.

P.S. This fact about silicon sensors is one of the reasons that the short mount distance of MILC cameras doesn't really enable lenses as small as proponents of MILC might think. MILC still require retrofocus designs (or heavy vignette correction) with wider lenses.

[TwoUptons]

I have a Zenitar 16mm f2.8 fisheye (probably the same one, though mine is native K-mount).

It does vignette on the K1, but it isn’t horrible. It also seems to be dependent on focus distance, with closeup subjects causing more noticeable vignetting. Though that could also be changes in lighting at distance, depending on what I’m shooting...

I’ll give it a go this weekend and see what I get...

-Eric

[UncleVanya]

Shake reduction is one of the key reasons as the image circle needs to extend beyond the borders of the film cameras. As the sensor moves around the center of the image will move towards one side or another - it should be slightly asymmetrical if this is the cause.

The micro lenses mentioned earlier are also a possible factor.

[stevebrot]

The frame size for 35mm film is 24 x 36mm* and that for the K-1 is 24 x 35.9mm. As a result, the FOV for the captured image will be the essentially the same.


Steve

* 24 x 36mm is the nominal size. Actual size between camera makes and models may vary.

[stevebrot]

I recently pulled it out to play around with and realized it has terrible vignetting on the K1. What gives?

It probably vignettes on film too. What fisheye are you using? From your description (heavy with T-mount), it is not the Zenitar 16/2.8.


Steve

[stevebrot]

Older ultra-wide angle lenses may have exit pupil distances that are too close to the film plane and that cause vignetting.

Are you sure? This is an issue with rangefinder lenses, but retrofocus designs required for almost* all SLR wide angle lenses result in exit pupil locations significantly forward of the rear principal plane and fairly close to that of the entrance pupil.


Steve

* Exceptions being lenses designed for use with the mirror locked up for clearance.

[MossyRocks]

I'm thinking the microlenses that photoptimist mentioned are the cause. On bright sunny days with a nice blue sky I've taken some shots with my old S-M-C 17/4 fisheye takumar and in the corners with the blue sky it almost looks like shots I had taken years ago with a polarizer on my 28/3.5. It is only noticeable some times, sun to my back, so on a lot of shots I pay no attention. it is easy to correct in post processing with the vignetting correction for that lens and for what may be in the corners it isn't a concern since they are pretty distorted anyway.

[stevebrot]

On bright sunny days with a nice blue sky I've taken some shots with my old S-M-C 17/4 fisheye takumar and in the corners with the blue sky it almost looks like shots I had taken years ago with a polarizer on my 28/3.5.

The edges of the sky are usually not the same brightness as the sky overhead and can be relatively dim at opposition when the sun is low in the sky. That said, relative brightness across the field for fisheyes and other wide angles may also be due to optical vignette due to shading or scattering of light between lens elements. It is a complex subject.

I don't have a FF digital camera to test with, but my Zenitar 16/2.8 Fisheye is not particularly prone to vignette on 35mm film.

[UncleVanya]

To eliminate shake reduction shoot on a tripod with 2 seconds self timer. If those shots still vignette s as bad then SR isn’t the culprit.

[MossyRocks]

I don't have a FF digital camera to test with, but my Zenitar 16/2.8 Fisheye is not particularly prone to vignette on 35mm film.

I've never noticed vignetting with the 17 fisheye takumar on film, but digital I have gotten darkening like what your 3rd picture with crater lake shows in the upper left. It has never really bothered me and I just treat it as a feature but just something that I have noticed.

[stevebrot]

I've never noticed vignetting with the 17 fisheye takumar on film, but digital I have gotten darkening like what your 3rd picture with crater lake shows in the upper left. It has never really bothered me and I just treat it as a feature but just something that I have noticed.

In that photo, upper left was opposite the October morning sun. Photo was taken facing northeast from:

Google Maps


Steve

[Bob 256]

For a longer lens, the sky test is a good one for vignetting but for a wide angle or fisheye, sky brightness isn't that even so you need an alternate like a very uniformly illuminated surface or a translucent lens cap which is evenly illuminated in order to test a lens. photoptimist is correct, and the "lening" nature of sensor pixels makes them somewhat direction sensitive so light coming from an angle has less sensitivity than light direct-on. Longer lenses have a narrower cone of light whereas short FL lenses have a wide cone and edge rays hit the sensor at a considerably sharper angle that center rays so vignetting will be more pronounced.

[stevebrot]

Longer lenses have a narrower cone of light whereas short FL lenses have a wide cone and edge rays hit the sensor at a considerably sharper angle that center rays so vignetting will be more pronounced.

Three words: Exit pupil location

This can be visually estimated by direct examination from the rear of the lens and for some short teles is sometimes closer to the sensor than the typical ultra wide. I just checked my Samyang 14/2.8* and it appears to be about 10mm beyond the lens flange towards the front of the lens. 10 + 45.46 = 55.46mm or about the same as for my ST 55/1.8. My Tamron 24mm is about 5mm forward of the lens flange. Shallow angle of incidence is likely to cause a problem due to optical vignette (huge issue with even moderate wide angles on view cameras) alone, even without consideration of sensor well depth/orientation.


Steve

* Recent purchase 3rd generation for that lens.

[GUB]

M42 mount MC Zenitar -M 2.8 / 16mm @ f2.8 on the K-1
Southern hemisphere - that is Orion at the bottom.
For me - acceptable vignetting.