PDA

View Full Version : What kind of standards are used to determine image circle?



Leonard Evens
14-Jan-2004, 09:29
I recently bought a 75 mm Grandagon-N f/4.5 lens for my Toho FC-45X. It is advertised as having and image circle of diameter 195 mm. I also got Toho's eccentirc lensboard. Together with the normal movements, I can get pretty close to the limits of the image circle. So far I've only made one exposure---of a building facade---and I was pushing the shift capability. I don't notice any significant light falloff, but there does seem to be some loss of fine detail on one side and in one corner. I think this is most likely due to the front standard being slightly out of alignment, thus producing a slight swing. Some tests just viewing the gg image seem to confirm that conjecture. But I began to wonder how I would know that I was reaching the limits of what the lens can do. Presumably the size of the image circle is defined in terms of adequate definition, but what standard do they use?

Ernest Purdum
14-Jan-2004, 10:28
Some lenses, mostly newish ones, cut off sharply, so determining the diameter of the circle is easy. Lenses which lose definition before vignetting occurs are harder to quantify. So far as I know, there is no accepted optical industry standard for how much loss of definition and/or illumination is acceptable within a published figure. From many comments here, it would seem that most manufacturers, and Schneider in particular, are rather conservative in what they claim.

Walter Glover
14-Jan-2004, 12:21
Leonard,

With regard to the errors in alignment may I suggest my first rule of view camera operation: NEVER ASSUME!!

Always check focus at various points in the image and tilt and swing accordingly. Much as I am not a great fan of Sinars I must say that this is one area at which their two-point focus technique really shines.

Although only mildly related to your initial post may I also suggest that you regularly check that the spirit levels indicate 'true' level.

Leonard Evens
14-Jan-2004, 13:15
Walter,

I don't ordinarily make assumptions.

With a wide angle lens, particularly if you are not used to the focal length, it can be difficult to judge how well in focus you are far from the center of the field. Unfortunately, there is no simple way using a level or other simple tool that I know of to check if the standards are horizontally aligned. The only way I know how to do it is to focus on a plane subject, check to see if the lens axis is perpendicular to it by making measurements on the gg and then check the focus across the field. If you get that right for a longer focal length lens, it should stay okay for other lenses, provided you don't use a swing in the interim.

The detent positions should give parallel standards, but often there is a slight amount of play in those positions, which may be enough to throw it off in critical cases. In this case, since it was my first outing with the lens and the eccentric lens board, and the light was not very good, I was just playing, so to speak. I assumed the standards were parallel since I thought I had left them that way. I discovered subsequently with another lens on another day that they weren't.

If you have a quick and dirty way to check horizontal alignment, please let me know what it is.

Ralph Barker
14-Jan-2004, 14:10
I use a small, triangular two-vial level that I found at a hardware store for both initial leveling and squaring of the camera standards. It's probably not as precise as a Starrett machinist's square, but it's close enough and made of lightweight plastic. It wouldn't work as well on a monorail, however, as you wouldn't have the same proximity of reference surfaces. On your Toho, Leonard, you might try precision-marking points on the sides of your standards (i.e. top and bottom), and then use a tape in the field to measure the distances between those points to ensure that the standards are parallel. Measuring the diagonals between those points (as one might do in cabinetry) also ensures that they are square.

Leonard Evens
14-Jan-2004, 14:38
Ralph,

I do use the measuring method, but the standards are not symmetrically placed, so there is a limit to how well that will work. I find that judging the focus on the gg is more reliable, except it is harder to do for short focal length lenses.

Don Boyd
14-Jan-2004, 16:46
I too am looking for a way of checking that the standards are parallel on my Toho.

Kerri T., if you are listening, how do you handle it?

Bob Salomon - HP Marketing
14-Jan-2004, 17:03
"way of checking that the standards are parallel "

We use the Zigaline

Emmanuel BIGLER
15-Jan-2004, 12:51
Leonard. your question is not easy to answer. I am not sure that there is a standard among lens manufacturers to define the image circle.

If you look at the specifications of the new Schneider Digitar® lenses you'll see that the recommended image circle depends in fact on how fine the pitch of the digital sensor is, in other terms the acceptable image circle is defined for those new lens designs by a sharpness criterion, not a photometric criterion.

In conventional lenses, there is an effect named the "cat's eye" effect which gradually cuts the shape of the exit pupil like a cat's eye pupil. Before the cat's eye effect occurs, starting from the centre of the field, you can consider from a photometric point of view, that you are in the full illuminated circle. In this field, assuming non distorsion of the pupil like in enlarging lenses, the cos^4 law applies which is a slow variation. There, all rays entering the entrance pupil actually cross the whole lens and eventually find their way to the image. Then gradually moving off-centre, you enter at some point the cat's eye regime, there you are in the "contour" region with the light vanishing, more abruptly that the cos^4 law, until the exit pupil is 100% cut off. In the cat's eye regime, some, but not all, rays that enter the entrance pupil never reach the image and are blocked inside the lens barrel.

In old lenses the contour field is clearly visible and defines the limit of illuminated circle. The contour field is systematically hiden in microscope eyepieces and binoculars. So you really have to make an effort to see what optical engineers do not want you to see ;-);-)

An nice example is shown on some Atget's views of the Old Paris. Atget insisted that the verticals should not converge on buildings. So he pushed the shift beyond the maximum of his lenses. This is not an unpleasant effect, it yields a semi-circular frame at the top of the image, a shape that was actually used at the time for framed photographs and portraits.

In new wide angle lenses, you have some pupillar distorsion, the exit pupil first increases in diameter ; this somehow compensates for the cos^4 law ; and then is abruptly cut-off. The image quality remains incredibly good until the very last limit ; there is an interesting example in Jack Dykinga's book.

Another example : Apo Ronars are credited of a 48 degrees image circle on their specs, but many amateurs use them on very large format cameras as a standard lens (~53 degrees). The apo ronar was designed with the goal of an ultimate sharpness at 1:1 ratio, the manufacturer was probably conservative about the actual angle of excellent sharpness vs. the ultimate illuminated angle. And if you wish, you can as of 2003, use them like Atget used his lenses ;-);-)