Mamiya Macro A 120mm F 1:4 On A Canon Camera

Is this distortion caused by the 3D surface of the coin? For example, as seen by the sensor, the highest points of the coin may shift away from the center faster than the lowest points. Also, sloping surfaces could change shape (or even get eclipsed) as you shift the coin away from the center.

Or, would this distortion also be seen in a perfectly flat
target, like a USAF test chart?

rmpsrmps for the fourth time Please produce a diagram...... it is exasperating to have asked you for a diagram so many times and for you to not do so. Why have you ignored that request. If you haven't worked it out before now there is just no way I get what you are writing down.

You first diagram of shift coupled with your words seemed, and still does a major distortion of the laws of physics to me.

ROF...... I see absolutely no "distortion" in either picture both are natural views taken from different perspectives!
In your "natural" the sensor is tilted with respect to the subject. The result is not a distortion ( I now get what you are saying in that way).

Likewise in the shift there is also no distortion, shift has enabled the photographer to "change position" and produce an image ensuring the sensor and subject are parallel. This is what I understand by correcting perspective. You change the apparent viewing position

However once you have a digital photograph taken using your "natural" the only way you can get rid of the converging lines is to distort that particular image.

Shift on the other hand also produces a natural picture but the perspective is different and the converging lines ( because the sensor and subject remain parallel never occur)

The distortion you appear to referring to seems to me to an erroneous transfer of photoshop-think to an issue of composition

There is no need to start to get into personal attacks. You have gone this route before and it does you no credit. Read the edit.
What you are saying is this to get an image, that has been taken with a perspective that has created converging lines, to look like a picture that has been taken with a perspective that does not have converginglines you have to distort that image
Shift changes the perspective( apparent position the photograph was taken from) before the photograph was taken so there is no need to distort the image in Photoshop)

It should be obvious to everyone that itis extremely confusing for me to understand it your way. I find it a lot simpler and much more consistent with standard tilt shift terminology to refer to the process as "controlling perspective". Here is a shot of my back light taken in such away as to imitate the subject-sensor relationship indicated in you diagram of "natural".





I will not use the word distort in this part of my post to save confusion. you can see the white part of the back light makes a trapezoid shape. The 3D image of the medal light translated onto the sensor( which is tilted) as a trapezoid. The edges I have indicated are equivalent to converging verticals( imagine the back light as a building and I have lain on my back in the street to take the photograph of that building. This process would tilt the sensor in respect to the "building".

Now if I read your posts correctly; you regard those converging verticals as "natural" yet a coin in the same optical set up rendered as an ellipse would be distorted ( see next paragraph)

If it was a coin that I had imaged the coin would be translated into an elliptical shape. By my understanding of a number of different posts you have made in this forum (and once in this thread) you would not regard the elliptical rendition of the coin as natural. Perhaps I am wrong but my understanding from my reading of those other posts you appear to have referred to such an elliptical image of a coin as "distorted". In those other posts you talk about correcting the "distortion" ( your words; unless I read wrong) by either manipulation of the image in photo shop or using some form of tilt so that the image comes out circular.

Now I am sure you can see how I regard your approach to the same optical set up as being a significant contradiction. That being that you, as I have understood your posts, regard an elliptical image of a coin as distorted. Yet you do not regard the converging verticals, created by a sensor tilted with respect to the subject, as a distortion.

Now on your wording of shift distorting an image. It took a good work out at the gym. {BTW thanks the exasperation I had with this discussion appeared to translate into being able to raise the weights I was pushing up another 2.5-5 kg. I actually easily benched pressed 70kg((154 pounds) today} for my subconscious to work out what your, for me, complicated and too technical description of shift causing distortion:

I imagined using a medium format camera to image of coin the size of a cent. Placing that coin so that is on the edge of the field of view there will be some distortion when compared to a similar coin in the center of the field of view. In practical terms the amount of distortion is dependent on how flat the field of the lens is. and how far from the center of the field of view the coin is placed. With my rig and the lenses I now use ( based on my observations) that distortion has proved to be so negligible as to be practically unimportant. If I used a wide angle lens( without a flat field) with then the story would be very different and the distortion would be in practical terms unworkable.

In practical terms with the mamiya 120/4 macro the coin at the edge of the field of vies is not far enough away from the center of view to show noticable and easily measurable distortion. Likewise I have never had any noticable distortion with the Olympus Zuiko 1354.5 bellows lens.

{the oly 135 actually projects an image circle almost as large as the meduim format lens. I mount it on the mirex adapter via a maiya 645 67mm reversing ring, combined with the Oly extention tube made for that lens and a 67 mm adapter Raf camera made for me.}.

This imagining made me realize that the same distortion theoretically occurs across a coin imaged by a small lens even when that coin is place dead center. The center of the coin will not be distorted in comparison to the edge of that same coin. However, if the depth of field is sufficient and the lens has a flat field the distortion across the coin will have no practical effect.

As I noted the mamiya 120/4 I have works best at low magnifications( large depth of field). So the distortion you refer to has no practical implications. I hypothesize that at 1-1 and higher that distortion would be a concern. It may well be that the poor performance I observed at 1-1 might be due this distortion. This imagining of mine, if correct, would suggest that the 120/4 has a flat field.

rmps this is what you just said. now a few posts back in response to your diagram with natural etc in it

I responded:





You then stated that was incorrect these are your words




I repeat: now you say this:





BTW: I have been using Tilt/shift also referred to as PC ( for perspective control) for around three years my most commonly carried lens is the mamiya sekor C 35mm 3.5n with the Mirex tilt shift adapter. I had read those links a long time ago, and many other references.

rmpsrmps I finally read this ridiculous last post of yours do you realize that other have understood what I was saying and you avoided it. He it is. A coin photographed from an angle will be imaged as an ellipse. That is called key-stoning. You call it distortion that's fine. However you say that converging verticals (which is also Keystoning) is natural......That is the contradiction....You don't understand that converging verticals and a coin being imaged as an elepse are the same phenomenon.


For every one else. rmpsrmps's statement that shift is a for of distortion shows a complete inability to understand shift on rmpsrmps's side. Shift is a method of avoiding distrotion it is not a method of distortion: Please don't take my word for it watch this instructional video from Adorama Watch the section starting 9.10 through 11.00:


You can see quite clearly that shift avoids distortion.