Lens light fall-off -- test results

[xkaes]

I wanted to test the possible/probable light fall-off of my Mamiya Sekor-C 37mm Fisheye f4.5 lens. Since I know there is lots of light fall-off on super-wide-angle lenses, I decided to put it to the test.

And while I was at it, I decided I might as well throw in my other wide-angle lenses, and determine possible issues with them, as well. Here is the list of suspects -- I mean, subjects:

Mamiya Sekor-C Fisheye 37mm f4.5
Schneider-Kreuznach Super Angulon XL 47mm f5.6
Fujinon SWD 75mm f5.6
Fujinon CM-W 105mm f5.6
Fujinon NW 125mm f5.6
Fujinon NW 150mm f5.6

The test gear is an evenly illuminated white wall, a Minolta Autometer II with a Minolta Booster II with a ground glass adapter.

All lenses were tested on a TOKO NIKKI II 4x5 camera with a fresnel and Beattie Intenscreen -- focused at infinity. Here is a photo of the gear:

166393

Clockwise from the lower left are:
Schneider-Kreuznach Super Angulon XL 47mm f5.6
Fujinon SWD 75mm f5.6
Fujinon CM-W 105mm f5.6
Mamiya Sekor-C Fisheye 37mm f4.5
Fujinon NW 150mm f5.6
Fujinon NW 125mm f5.6

The base meter setting was using the 150mm lens at f5.6 with the Booster II adjusted to create an EV rating on the Autometer II of ZERO. As this lens has the largest image circle, I would expect it to show the least amount of light fall-off -- from center to corner -- on the 4x5 image.

The following table summarizes the results. They are, surprisingly, exactly what they should be. All of the lenses showed exactly the same EV reading in the center at f5.6. And all of the lenses showed exactly a three stop drop in the EV reading in the center at f16.

166394

The corner readings are a little more complicated. I would expect light fall-off due to the lens, but also due to the screen. The amount due to the lens should be reduced as the aperture is stopped down. With the 150mm lens, there is a one stop difference from the center to the corner at f5.6. At f16, the expected three EV reduction adds up to an EV change of four stops from the base. I think it is safe to assume that the camera viewing screens are the cause of a one stop drop from center to corner. Anything above one stop is due to the lens.

The 125mm and 105mm lenses each show an additional 1 EV drop in corner illumination at f5.6 -- which is completely gone by f16. This is what I would expect.

The 75mm shows different results. At f5.6, there is only a two EV drop, and an additional 1/2 EV drop on the corners. I attribute the first change to the lens having a different optical design, while the second change can be explained by its wider field of view. The 47mm lenses shows similar, but slightly more significant results.

The big surprise, and what started all this, is that I did not see ANY light fall-off in the fisheye lens -- either wide open or stopped down!

[Dan Fromm]

Fisheye lenses have extreme barrel distortion that effective counteracts the usual cos^4 (or so, let's not quarrel about the exponent) falloff. Otherwise I don't see why you're surprised. The law is the law. The GG or fresnel's effect makes life harder for the photographer, doesn't change what the film sees.

Rodenstock published illumination vs. distance off-axis curves for many of their lenses. Schneider published similar curves for some of their lenses. Both recommend using center filters with lenses that cover >= 100 degrees. They might know something.

Fuji and Nikon didn't make center filters and as far as I know made no recommendations. I have a piece on center filters that addresses Fuji and Nikon extreme wide angles that should, all being well, be out at the end of the month on galerie-photo.com. It is somewhat of a follow-on to my article on Horseman exposure meters, see: http://www.galerie-photo.com/horseman-4x5-exposure-meter.html. You might want to look at it.

Testing is always good but calculations work too.

[xkaes]

Fuji and Nikon didn't make center filters and as far as I know made no recommendations.

I don't see any mention of it in any of my Fuji material. But even though Fuji and Nikon didn't make center filters, it's interesting that a few non-lens, filter manufacturers made a few -- at least that is how they are labelled.

[Dan Fromm]

Joe, the only independent filter maker's CFs I've found are Heliopans. I haven't been able to find Heliopan's own propaganda/instructions so don't know whether Heliopan made any recommendations about which lens to put which filter on. B&H's list of Helipan CFs says nothing either. I address the question of which lens might want which Heliopan filter in my CF article. Once you know the rules that Rodenstock and Schneider followed it all becomes quite simple.

[ic-racer]

I'm a fan of fisheye lenses, I'd like to see some full frame examples of pictoral subjects. As mentioned, the flux is concentrated into smaller and smaller film areas near the edges of the field. Entrance pupil is tilted and smaller at the far field, however, and perfect even illumination is not always observed though some lenses come extremely close. Here are some empiric results with a more robust testing method:

[xkaes]

That's very interesting results. Thanks for posting it. It is what I would expect. My method of light measurement is far too crude to measure at these levels.

The Mamiya 37mm fisheye that I use has an Image circle of 90mm so it is perfect for 4x5. You can see some results at:

http://www.subclub.org/fujinon/gonefishin.pdf

Whatever light fall-off there happens to be at the far edges is too insignificant for me to be concerned with.

[xkaes]

Joe, the only independent filter maker's CFs I've found are Heliopans. I haven't been able to find Heliopan's own propaganda/instructions so don't know whether Heliopan made any recommendations about which lens to put which filter on. B&H's list of Helipan CFs says nothing either. I address the question of which lens might want which Heliopan filter in my CF article. Once you know the rules that Rodenstock and Schneider followed it all becomes quite simple.

Here's one I just ran across. It is a Kenko CND filter made for Horseman, but who knows? Perhaps it was made by someone else, or perhaps Kenko made CND filters for others:

166752

More details are at:

http://www.ebay.com/itm/201973322824?ul_noapp=true

[Dan Fromm]

Thanks.

[Bob Salomon]

Here's one I just ran across. It is a Kenko CND filter made for Horseman, but who knows? Perhaps it was made by someone else, or perhaps Kenko made CND filters for others:

166752

More details are at:

http://www.ebay.com/itm/201973322824?ul_noapp=true

Kenko is Hoya

[Rael]

Here's one I just ran across. It is a Kenko CND filter made for Horseman, but who knows? Perhaps it was made by someone else, or perhaps Kenko made CND filters for others:

166752

More details are at:

http://www.ebay.com/itm/201973322824?ul_noapp=true

This might be a dumb question, but could you put one of these on your enlarger to correct for edge/corner falloff?

[Mark Sampson]

Rael, I'm sure that it's possible in theory. However, finding a filter that would do the job would be troublesome and expensive, with no certainty of success. I'll avoid thread drift; send me a PM and I'll explain how we solved an uneven enlarger light source when I worked at Kodak.

[Jac@stafford.net]

Rael, I'm sure that it's possible in theory.

No way in our wildest dreams. A center ND filter in front of an enlarger lens for making prints from negatives would only multiply the problem. What you need is an inverse center filter which still would not work without significant vignetted area detail and contrast.
.

[xkaes]

If your lens needs a CND filter, any picture taken with it will show less detail on the edges. There is no way to somehow create detail that isn't there -- OK, painting it in with some graphite, etc. is an alternative. You could use other techniques to darken the corners, which might help, but they would not add detail.

[xkaes]

Kenko is Hoya

Apparently, this is another complicated company merger -- this time Japanese. Looks like Tokina and Kenko merged at some point -- or one bought the other -- perhaps they were always the same. Then Hoya was mixed into the pot, somehow.

Here's what I found:

"Kenko Tokina and HOYA are two independent companies. HOYA licenses its brand to Kenko Tokina for camera filters being distributed outside Japan. Product planning and packaging are handled by Kenko Tokina."

It reminds me of the Osawa, Bell & Howell, Tokina, Mamiya-Sekor, Soligor plate of spaghetti from the 1970's. Tokina would make lenses that would be labeled as "Osawa", "Bell & Howell", and "Soligor" depending on the country/market/distributor. The overall company was really two companies "Osawa" in Japan and "Bell & Howell-Mamiya Co. (BHMC) in the US -- both of whom were distributors, not lens manufacturers.

So who knows makes Kenko filters. It might be Kenko, Hoya, Tokina, or another, completely different company -- or a mix of some sort. It makes me think that other "Kenko" CND filters -- sizes, densities, etc. -- might be out there. So far, I've confirmed that they made 62mm and 67mm 3X CND filters in slim-mounts (AKA, no front thread). It would be nice to see the instruction sheet on these filters since it would likely show the focal lengths/lenses they are designed for. I don't see anything on the KENKO, or HOYA websites.