jlw
Rangefinder camera pedant
We've been having an extended discussion on the R-D 1 forum about how the R-D 1's smaller imager (vs. 35mm film) affects, or does not affect, the perspective effects in pictures.
I've prepared two attached images which I hope will make the situation clear. First, though, a definition:
Note that since it's a convention, there are different styles of perspective representation, and no one of them is more "right" or "wrong" than the others. The style of perspective representation we're accustomed to seeing in photographs is called "academic rectilinear perspective," which was first developed by Renaissance painters, and makes use of the fact that farther objects will appear proportionally smaller to the viewer than closer objects.
Now, about how that affects photography on different imager sizes and with different lenses:
In my first example, called 'constant-distance.jpg', I've used a 3D modeling program to show how a simple scene looks when photographed with 18mm, 50mm, 85mm and 150mm lenses all from the same distance. When you look at the example, you'll notice that you see different amounts of the foreground subject -- but the perspective is identical in each shot. (Look at where the background pillar hits the figure in the foreground, and you'll see that it always hits him in the same spot.)
In the second example, called 'constant-size.jpg', I've used the same virtual lenses -- but this time, I've moved the camera to different distances so that the foreground figure is always about the same size.
In this case, you'll see that the perspective is different for each photo -- the size of the background pillar varies in proportion to the constant size of the foreground figure.
Note that if you took the first set of pictures and cropped them by the same amount (to simulate using a digital imager instead of 35mm film) you would not change the perspective -- only the amount of the subject you can see.
What this proves is that perspective relationships are determined only by distance. The reason that lens choice and format size seem to influence perspective is that we usually choose a lens that lets us "fill the frame" with the subject at whatever distance happens to be convenient -- so if we change lenses (or format sizes) we have to change the distance to get the same frame-filling effect.
But it's the distance change, not the lens or format change, that causes the perspective relationship to change.
Feel free to download the examples and play with them in an image editor, if you want to get a feel for this.
I've prepared two attached images which I hope will make the situation clear. First, though, a definition:
Perspective is a convention for representing three-dimensional space in a two-dimensional medium.
Note that since it's a convention, there are different styles of perspective representation, and no one of them is more "right" or "wrong" than the others. The style of perspective representation we're accustomed to seeing in photographs is called "academic rectilinear perspective," which was first developed by Renaissance painters, and makes use of the fact that farther objects will appear proportionally smaller to the viewer than closer objects.
Now, about how that affects photography on different imager sizes and with different lenses:
In my first example, called 'constant-distance.jpg', I've used a 3D modeling program to show how a simple scene looks when photographed with 18mm, 50mm, 85mm and 150mm lenses all from the same distance. When you look at the example, you'll notice that you see different amounts of the foreground subject -- but the perspective is identical in each shot. (Look at where the background pillar hits the figure in the foreground, and you'll see that it always hits him in the same spot.)
In the second example, called 'constant-size.jpg', I've used the same virtual lenses -- but this time, I've moved the camera to different distances so that the foreground figure is always about the same size.
In this case, you'll see that the perspective is different for each photo -- the size of the background pillar varies in proportion to the constant size of the foreground figure.
Note that if you took the first set of pictures and cropped them by the same amount (to simulate using a digital imager instead of 35mm film) you would not change the perspective -- only the amount of the subject you can see.
What this proves is that perspective relationships are determined only by distance. The reason that lens choice and format size seem to influence perspective is that we usually choose a lens that lets us "fill the frame" with the subject at whatever distance happens to be convenient -- so if we change lenses (or format sizes) we have to change the distance to get the same frame-filling effect.
But it's the distance change, not the lens or format change, that causes the perspective relationship to change.
Feel free to download the examples and play with them in an image editor, if you want to get a feel for this.
dmr
Registered Abuser
jlw said:
Looking back many years to high school art, where we learned to sketch objects in "perspective", I can remember the sessions where we sketched things like wooden blocks varying the distance of the artist to the nearest object, and yes, this did have quite a bit of influence on the apparent size of the farther objects.
Yes, I would think that with a "normal" lens for either the 35mm frame or the electronic sensor, the distance to the main subject for frame-fill would be about the same, thus the same perspective and the same apparent size for other subjects.
This is an interesting subject. It also relates in a very geeky sense to a college level linear algebra class I took where we briefly touched on the techniques for mapping 3d space (of a scene) onto 2d space (for a screen display) and the kinda spooky math for doing it, which I don't recall off hand.
