About Pinhole Image Quality

Veijo Vilva

For most pinhole photographers the technical quality of their pinhole photographs isn't the most important thing. However, the image quality obtainable with a properly designed medium or large format pinhole camera can be surprisingly high and sometimes even exceed the photographic expectations of most people. A 5.25" x 5.25" digilab print of a 2.25" x 2.25" pinhole image can be about as sharp as sharp can be at the 300dpi resolution which for most people is the norm today. Larger prints are, of course, soft to quite soft but in a way which many people find pleasant or even soothing. I have rather nice 13" x 13" prints at 254dpi.

Resolution

Here is an 8mm x 6mm section from a 54mm x 54mm crop (3260 x 3260 scan, 10Mpixel) of a photo I took with my Zero 2000 6x6 pinhole camera (f = 25mm, f /138). An uncropped 56mm x 56mm frame would be about 3380 x 3380, slightly exceeding a maximal square crop from the full resolution frame of the 16Mpixel Canon 1Ds Mark2, i.e. a 3328 x 3328 section of the 4992 x 3328 full frame. Of course, photos taken with a 1Ds are sharper, have less noise and a better local contrast but may also at times display artifacts absent from a pinhole photo. (There is a similar test photo in the DPReview preview of the 1Ds Mark2, the fourth photo in the series.)
Note: it would be necessary to have an 11mm lens on the 1Ds to cover the same (square cropped) angle of view.

I've drawn a 1mm long red line on the picture. There are seven quite thick line pairs above the red line, and it is quite obvious that the camera can resolve rather more than 10 lp/mm. The smallest visible details are quite a lot smaller than the 0.18mm pinhole. The scanning resolution happened to be roughly four times the theoretical f /138 line pair resolution, which seems adequate for quite thin vertical and horizontal lines but not always for the thinnest slanted ones which are clearly discernible on the negative. This observation is in agreement with the Image Detail article by R.N. Clark. Displayed at a resolution of 730 x 730, the photo would pass as a pretty good "normal" one. The image quality can often be enhanced if one isn't a purist, and I certainly am not one.

Below are roughly corresponding pictures (a crop and a downscaled photo) taken with a Zenitar-M 2.8/16mm fish-eye lens mounted on a Canon EOS 350D, 8 Mpixels (3456 x 2304), not quite as wide angle and not rectangular (e.g. the pillars look quite curved) but certainly very much sharper.

*

Here is a similar crop from a B&W photo (10Mpixel) taken almost straight against the Sun. There are no disturbing artifacts, the photo is soft but clean and well-behaved.
(Note: part of the structure in the photo is above a glass roof.)

730 x 730

*

A crop from a close-up photo (10Mpixel). I took this photo under rather low light, low contrast conditions placing the camera in a display cabinet at the side of the cat skeleton, which explains the rather saurian pose and look of its skull. (Some saved old Veterinary Science demonstration material at the Department of Animal Science, University of Helsinki.)

730 x 730


Note: the crops and the corresponding full image files haven't been manipulated to improve the image quality. The small versions on this page (and my other pages) have been sharpened with the refocusing plug-in in Gimp, which process also significantly improves contrast. The negatives were scanned at the digilab which developed the films, and the quality is at consumer(/prosumer?) level and quite adequate for most people's pinholing.

I had been inspecting the colour negative under a microscope and in the process deposited on it some dust which I forgot to remove before having the neg scanned, and the digilab didn't bother to clean the neg at the price they charge for the scanning, i.e. less than a quarter of the price level of the cheapest pro lab, which would have been excessive for pinhole photos.


Back to my Pinhole Photography page

+++ veijo.vilva +++ @@@ +++ gmail.com +++