I'm very sorry, I was not intending to shout, simply trying to be unambiguous and to the point.
I agree that you referred only to the relative size, but you then deduced that there was no difference between the two cases (I assume that is what you meant by "ergo, no difference").
That deduction implied to me that you had assumed that the absolute size was unimportant. Why did you not mention the absolute angular size when it is the main thing that varies when the viewing distance changes?
First of all, it wasn't a trick. I simply asked you to explain "what you think observation would show in the following situation." I didn't force you to describe it in terms of pure maths or anything else. I was curious about how you would describe it. It seems that you don't want to describe it at all, except in very vague generalisations, e.g.
Nothing precise there, no mathematical formulas, almost nothing of any practical value to anyone.
I like to give people the benefit of the doubt and give them every opportunity to explain themselves more fully and in more detail if I cannot understand what they say. You have been given every opportunity, but you simply repeat the same vague statements time and again and refuse to go into any more detail, yet you have the temerity to criticise the theory of perspective that has been understood for centuries and claim that it is not possible to explain what we see in a photograph in terms of mathematics.
You claim that you agree with the statements about perspective in the Manual of Photography, but you then proceed to describe perspective in your own fanciful ways that contradict the traditional understanding of perspective.
I guess that you will continue to talk the same old nonsense and continue to claim that what you say is sound science and that I (or anyone else disagreeing with you) am the one talking nonsense. The Donald Trump technique?
If you assume that the absolute size of the image makes no difference then it is easy to prove that the viewing distance makes no difference.
But why do you assume that the absolute size is unimportant? I think this is probably the key to the whole of this discussion. I think Ansel Adams assumed that an image always looks exactly the same whether you look at it as a 4" x 5" print or as a 4' x 5' print.
If you view a 4" x 5" print alongside a 4' x 5' print of the same image, they will look like two prints of the same image. However, it is also obvious that one is much larger than the other. Why assume that the absolute size has no effect on how we "see" the two prints?
Leon Battista Alberti (1404 - 1472) wrote:
In other words, he thought that viewing distance does matter.
It puzzles me that so many photographers today are willing to assume that viewing distance does not matter and think that it is obvious that it doesn't matter.
It is obvious to me that it doesn't matter if you are purely seeing the photo as a 2D object. However, if you are viewing the photo in a situation where it is possible that you could imagine that you were looking at a real scene, then I think it is obvious that the angular size with which you see objects in the scene does matter and affects your perception of depth in the scene.
Suppose you are looking through a keyhole into a room containing two people. How do you judge depth in that situation? The keyhole is very small and prevents you from using binocular vision to judge depth. It also prevents you from moving your head from side to side to use parallax to help judge distance. Yet most people can easily tell if both people are standing close to the keyhole or if they are standing some distance away. How?
So, please explain what you think observation would show in the following situation.
You take a photograph of two people, both of height 1.8m. One is standing 10m from the camera and the other is standing 20m from the camera.
The photo is viewed from (1) the centre of perspective, and (2) the point halfway between the centre of perspective and the centre of the image.
How would you describe what is seen in each of those two cases? What is the difference between the two?
Please answer the question and do not use evasive tactics.
There is a very well-established theory of perspective that explains mathematically what happens when we make a 2D photograph of a 3D scene from a given camera position. That theory also explains mathematically what we see when we look at that photograph from a given viewing position in comparison to what we saw when looking at the original scene. Any good mathematician has no problem understanding both of these.
As I understand it, you fully accept the first part (taking the photograph), but you do not accept the second part (viewing the photograph). Is that correct?
The "perceptual leap" that you refer to is simply an excuse to avoid a proper scientific explanation. Either you can explain clearly and logically what you mean or your theory is effectively nonsense.
Changing the viewing position changes the angle of view, but who said anything about lighting?
You keep repeating the same mistake. For each of the four images, the camera position changes. For each of the four images, the centre of perspective changes and hence the viewing position relative to the centre of perspective changes when all four images are viewed simultaneously.
It is irrational to claim that the change in viewing position relative to the centre of perspective has no effect and that the difference is solely because of the change in camera position.
You really need to do some experiments keeping the camera at the same place and just changing the viewing position (relative to the centre of perspective), but you seem unwilling to do that. It would quickly expose the fallacy in your argument.
On the contrary, it is good science. An opinion is worthless if it is based on an argument containing fundamental errors of logic.
Andrew, there is a fundamental error in your logic. Consider this section of your argument:
Everything there is correct except the conclusion stated in the final sentence (which I have put in italics). To be logically correct, the sentence shown in italics needs to be qualified as follows (changes in bold):
So the difference shown in the linked photos can't be attributed to the viewing distance alone and must be a function of camera distance as well.
It is not logically possible to conclude that the viewing distance plays no part. All that can be said is that the difference in the photos must be due to either the camera distance alone or a combination of camera distance and viewing distance.
That logical error invalidates the rest of your argument.
This logical error is possibly the same one that Ansel Adams made in reaching his fallacious conclusion. It has been repeated many, many times since.
Let's go through this in detail.
If you view those thumbnails from 1m away, it is totally obvious that the 135mm shot shows compression relative to the 24mm shot (i.e. the background looks closer in the 135mm shot).
However, at that viewing distance it does not show compression relative to reality (i.e. to what you would see if you were standing at the camera position when the shot was taken). If you took your laptop to where the photographer stood when she took that shot and you viewed the thumbnails at 1m distance, with the actual Washington State Capitol Building visible across the lake at the same time, then it would be obvious to you that the Capitol Building looks further away in the 135mm thumbnail than it does in reality (when viewed from the camera position when the shot was taken).
When viewed from further away than the centre of perspective, distances appear extended, when viewed from closer than the centre of perspective, distances appear compressed (in comparison to what was seen from the camera position).
Try it sometime, it is instructive to do it for yourself.
You can even check it out with the photo displayed on your phone.
I'll stand by what I have said already. I think time will be the ultimate judge of who is right. Or perhaps we are both wrong. It seems unlikely that we are both right!
Objects A and B are flat objects. All of object A is at the same distance from the camera. So there is no foreshortening as it is of zero depth already.
If you want to consider an object in 3D, then take object A to be its front face and object B to be its rear surface (with nothing visible in between in this very simple example). The foreshortening is then equivalent to the change in distance between A and B in the two photo situations.
You can take a more realistic object if you like, but then every visible point on the object will need to be moved to one tenth the distance for photo 2.
I have never said that we do. Perspective is a mathematical model to explain how the shapes and sizes of objects change when we project a 3D scene to a 2D image and then when the 2D image is viewed from a specified point (relative to the image).
I say "viewed" but that does not necessarily mean through human eyes. You could instead take a photo of it (from the specified point) or use machine vision to view it and interpret it.
Human vision is not an essential part of perspective. The same theory of perspective applies when photographs are analysed by machine.
What do you mean by "out of the context of their normal distance"? What errors of judgement are you referring to? Can you give a simple example, please?
The image of object A would not be the same size as in the 500mm lens. So the software perception would not be the same.
Suppose I take a photo of two objects against a completely uniform white background. Object A is at a distance of 10 (in whatever units you like) from the camera and object B is at a distance of 20 (in the same units). The photo is taken with a 500mm lens.
Then I take a second photo with object A at distance 1 and object B at distance 2; this time using a 50mm lens. The size of object A will be the same in both photos. The size of object B will also be the same in both. The positions of A and B relative to the camera are assumed to be consistent with their positions in the first photo.
So, our perception of depth in both photos will necessarily be the same (provided that the only thing in the photos from which we can judge depth is the size of the two objects).
In other words, the photo with a 500mm lens has the same apparent perspective as a photo with a 50mm lens and with all distances from the camera reduced to one tenth of their original values. This is another way of describing telephoto compression. It does not involve looking at the photos by eye. They could instead be analysed digitally with software. How our eyes work does not come into it..
Why would I object to that? Obviously, our perception of what we are looking at depends greatly on the context. Depth perception depends on a very large number of cues that may (or may not) be present in the image. Many optical illusions (such as the three vans illusion that was discussed recently in dpreview) occur when different cues appear to be contradictory.
Perspective is just one cue that helps with depth perception (although often a very important one). I would never claim that depth perception depends on perspective alone, let alone just on the angular size of familiar objects which is the only thing that changes when just the viewing distance changes.
I am still very puzzled about what Andrew really thinks. He persistently avoids giving a precise description of what he thinks "telephoto compression" is and how it occurs, yet he very confidently claims that the explanations given in the Manual of Photography and elsewhere are all wrong.
The effect here is quite extreme, but it is not different in principle. As you say, the image plane must be at an extreme angle to the optical axis of the camera. To see the image correctly, it must be viewed from a similarly extreme angle (at the centre of perspective).
It seems that we agree on all the key points. Telephoto compression is something that we see when we view an image from closer than the centre of perspective.
