Optical Theory: FL and"compression," subjective effect on prints

[David R Munson]

So here's something that's been on my mind for a while and I hope some of the optical gurus here can help with. General optical theory as I understand it dictates that, as compared to how the human eye sees things, shorter focal lengths tend to exaggerate subject field depth, while longer focal lengths tend to compress it. For example, I take out my Nikon and take three shots of the same subject, each with a different lens. The first, with a 50mm lens, is a close approximation of how I see things with my eyes. The second, with a 24mm lens, creates a scene with an exagerrated amount of depth, making things essentially seem farther apart than they are in reality. The third, with a 300mm lens, shows compression of the scene, making things appear closer together than they appear to the naked eye.

Now, I'm not terribly knowledgeable about the actual optical physics that makes this happen, but what I'm more interested in is how it affects the appearance of photographs, especially as it relates to large format photography. Correct me if I'm off base, but these are my general observations.

If I put my 24" lens on my 8x10, I get the same kind of subject compression as I would with a 600mm lens on my 35mm. 600mm is 600mm no matter how big an area the negative covers. However, my field of view is a lot bigger, approximate to what I would get with a 100mm lens on a 35mm camera. If I go the other way and put a 150mm or even a 90mm lens on my 8x10, my field of view is enormous, but either lens is still a long lens in 35mm terms.

Now, let's consider the effects this has (or at least I think it has) on the appearance of the final photograph. If I take the same photograph with a 35mm camera and an 8x10, the 35mm cropped to the same aspect ratio as an 8x10, and appropriate lenses on each camera to give the exact same field of view on both cameras, am I wrong in thinking that the final photographs will have a fundamentally different, if not immediately obviously different appearance? For a given field of view, the 8x10 equivalent focal length is roughly 6x that of the 35mm focal length. If I use a 50mm lens on the 35mm camera and a 300mm lens on the 8x10, even if the field of view for both were identical, the 8x10 shot would have a different appearance due to subject field compression in the 8x10 shot, correct? A wide angle shot with a larger format has a different appearance than a similar wide angle shot made with a smaller format due to differences in equivalent focal lengths, correct? Same with long lens situations?

Assuming my assessment is correct, how does this affect how we use the various formats? Has anyone taken or otherwise been aware of any photographs in which this phenomenon is really visible? To me at least, this is something that's really interesting when you really think about it. But then again, if my math is wrong, then I really just have no idea what's going on. Backup or the contrary from the optical gurus or observations from anyone would be greatly appreciated.

[bglick]

Dave, I think you are confusing yourself here.... Lets take your example.... you set up tripods side by side. You shoot 35mm camera with 50mm lens and 810 camera with 300mm lens. Assuming the aspect ratio of the 35mm is the same as 8x10, then both images would have the exact same image within its borders, no differences.

Now, if you shoot 35mm with a 600mm lens, and 810 with a 600mm lens on it, and you crop the 810 film down to 35mm size, (or use a 35mm back) you once again will have the identical image! But if you compare the 35mm image vs. the full 810 image, the 810 image will be much wider in its view. so to answer your question.....

If I take the same photograph with a 35mm camera and an 8x10, the 35mm cropped to the same aspect ratio as an 8x10, and appropriate lenses on each camera to give the exact same field of view on both cameras, am I wrong in thinking that the final photographs will have a fundamentally different, if not immediately obviously different appearance?

Yes, you are wrong, they will look identical. Whats considered normal on 810 camera is considered long on 35mm! It's all relative to film format.

[David R Munson]

Hmm....maybe I am confusing myself here. But essentially what I want to know is this: With a larger format, a longer focal length needs to be used in order to get the same field of view as you would get on a smaller format. With the longer focal length, you get more subject compression, despite the same field of view. So how does the longer focal length affect the appearance of the photograph, if at all?

[David R Munson]

To further clarify, I'm not so much interested in why the normal focal length for a given format is the normal focal length, or how it affects what is included within the borders of the photograph, but rather how/if the varying normal focal lengths of different formats affect the appearance of the subject within an otherwise identical field of view.

[Conrad Hoffman]

Think about this differently. The focal length has nothing to do with compression or, more properly, perspective. Zilch, zero, nada. That's determined entirely by where you stand. The angle of view that the camera can "see" is determined by the format size combined with the lens focal length. It's that simple.

[bglick]

David, here is a better way to think of it... your description of compression for long lenses and stretch (or exagerated) for near objects with wide angle lenses is very accurate! The question is, what is normal, what is wide and what is long? The answer to that question is dependent on the film formats diaganol.... a normal lens equals the diaganol of the format, a longer lens is greater than the diaganol, etc. So you have the right idea, you just have to change the middle point (normal lens) for each format.

If you take the diaganol of 35mm, 6x7, 4x5, 8x10 and use a lens for each format exactly equal to their diaganol, you will get the identical image, on each piece of film, (correcting for aspect ratio a tad) If you use a lens for each format that is 6x the diaganol, you still will get the identical image on each peice of film. Does this make sense?

[David R Munson]

Bill, so then the stretch or compression of short and long lenses is entirely relative to the format and not the focal length of the lens? The extreme "compression" of a 600mm lens in 35mm would then have more of an apparent compression than with a 600mm lens on an 8x10? This is starting to make more sense. Thanks.

[neil poulsen]

I'm with Conrad on this one. The sense of "compression" is dependent upon perspective, which is dependent only upon the position of the camera. Perspective IS NOT dependent on focal length. So, one will get the same sense of compression from both the 8x10 and the 35mm camera, regardless of the lens used on either camera.

In comparing photos taken with a 300mm on an 8x10 and a 50mm on a 35mm, depth of field can change, which can also give a different sense of "compression". But, by adjusting the aperture on the 8x19 to be smaller (with a longer exposure), the depth of field can be made identical on prints from both cameras. So, even in this case, the sense of compression is the same from both cameras.

By adjusting the depth of field, so long as one adjusts the lenses on both cameras to obtain the same angle of view, then the photos will be identical from both the 35mm and the 8x10. (Let's assume that one is copping the 8x10 to obtain the same length to width ratio as that on the 35mm format.)

[bglick]

Dave, you wrote......, so then the stretch or compression of short and long lenses is entirely relative to the format and not the focal length of the lens?

Better said, film format defines what lens is long or short as I described above. To alter the amount of stretch or compression from one image to another, you need to alter the camera vs. subject distance. Neil wrote.....I'm with Conrad on this one. The sense of "compression" is dependent upon perspective, which is dependent only upon the position of the camera.

I agree with you and Conrad, but to answer Davids original question, we assumed both hypothetical images were taken in the same spot. His question was about the difference in the final photograph when using a 35mm with normal lens vs. an 810 camera with normal lens, shooting from the same position. He felt the final images would be noticeably different. But in reality they would be near identical. (putting aside, DOF, aspect ratio, movements, etc.)

[Pete Andrews]

"If I take the same photograph with a 35mm camera and an 8x10, the 35mm cropped to the same aspect ratio as an 8x10, and appropriate lenses on each camera to give the exact same field of view on both cameras, am I wrong in thinking that the final photographs will have a fundamentally different, if not immediately obviously different appearance?" - Yes, totally wrong.Perspective has nothing to do with the focal length of the lens used. It's determined only by the camera to subject distance.A given angle of view will give you the same visual effect on any size of film. It's only depth-of-field that's drastically altered by the format size. You would have to multiply the numerical aperture of the 35mm camera by 8 to get an equivalent DoF on the 10x8 camera. (ie, f/4 on the 35mm camera would become f/32 on the 10x8 ). If you do that, then the pictures will appear exactly the same, excluding the better tonal quality of the 10x8, of course.

[David R Munson]

Well, so much for that idea. Still not entirely convinced, or at least not entirely understanding of the concepts here, but a trip to the library should cure that. Thanks for all the responses.

[Doug Paramore]

David, since you shoot several formats, why not do your own tests? Shoot a sheet of 8x10, MF and 35mm. Cut everything down to 35mm size, and the images should be close to identical, given a small variation in what is considered "normal lenses" for each format. To agree with previous posters, it is distance from the lens that makes the difference in perspective, film size that determines what is included at the edges.

Regards,

[neil poulsen]

Begin with Camera & Lens, Ansel Adams, Revised Edition, 1948, 1970, 1976. This edition is prior to his most recent edition in which he worked with an editor. Check page 147 under Visualization With Camera and Lens. You will see a very similar discussion to that which we've had above.

Adams begins by defining perspective by saying that, "True perspective depends upon the distance of the subject from the lens." But, a few sentences down, he makes the statement, "A 2-inch lens, an 8-inch lens, and a 24-inch lens mounted on the same camera at the same 'position' relative to the subject will give identical perspective."

It's this second description to which I adhere. It isn't just distance from subject, it's the relative position. Lift the camera 20-feet, or move it 20-feet to either side, and although the distance might be the "same", one can achieve completely different perspectives. It's not just "distance" that's the determining factor of perspective. Actually, moving the camera in this way changes ALL distances from the camera to each point in the photograph.

[N Dhananjay]

The reference I found most helpful to clarify this was Stroebel's textbook. The notion of compression of space is entirely a function of relative magnifications of foreground and background objects. Magnification (with a given focal length) is a function of distance i.e., magnification increases dramatically as an object comes closer to the lens. With a wide angle lens, nearby objects are rendered at dramatically larger sizes than distant objects, which provides the 'illusion' of depth in the scene. In contrast, a long focal length (which is typically used from a distance) is operating far enough from foreground and background object that there will be no such exaggeration of 'space'. Since magnification (its worth thinking of distances as multiples of focal length since it is intuitive to think of magnifications in this way) is relative to format, the 50mm on 35mm will provide a similar look to a 300mm on 8x10.

Set up 2 little objects of the same size on a scale and focus your 8x10 on it. Now move one object along the scale and measure the change in image size on the ground glass - you will see the dramatic reduction in magnification as an object moves away from the lens. At a distance between 2 times the focal length and 1 time the focal length, the magnification will increase from 1X to infinity (your bellows draw of course will get inordinately huge). But it will give you a good idea of how relative magnifications (which of course is reliant on relative distances) can alter subjective depth tremendously - in other words, it is where you place the camera which determines the relative distances and therefore the relative magnifications. A wide angle lens is typically used up close to a foreground object (exagerrating its size i.e., the ratio of foreground object distance and background object distance can be 5 times focal length and 30 times focal length respectively). A 300mm on 35mm will compress things a lot (because both foreground and background are far away e.g., 25 times and 30 times focal length, and relative magnifications are nearly equal). As stated above, true perspective is a function of position and not focal length (if you crop off the wide angle, you will get the identical framing as from a longer lens).

In other words, a 50mm lens on 35mm will provide an angle of view similar to a 300mm on 8x10 - the 300mm magnifies things more of course but it magnifies everything equally and maintains the relative sizes. The relative magnifications provided by these lenses on these formats will be pretty close i.e., the illusion of space will be identical.

Cheers, DJ

[Steve Gangi]

Perhaps this is an over simplification, but it helps to think of it this way. Perspective and depth are determined by the relative distance between the camera and the subject/forground/background. Where you stand controls the perspective. The choice of lenses controls how much of the "general area" is recorded on film, and can be thought of as in-camera cropping. Finally, larger film sizes use longer "normal" lenses because the image must be magnified more to cover the piece of film.

[David R Munson]

Now it makes sense! Just one of those things I was having a bit of trouble figuring out myself. Thanks much to everyone.

[Struan Gray]

A spanner in the works: I agree fully with the analysis of how perspective doesn't change with different focal lengths, but believe strongly that that analysis misses the point.

Go and look at some photos without thinking too hard. The compression of distance by long focal lengths is obvious, universal, and requires a dedicated, pig-headed effort of will to avoid seeing. But, crucially, it is a result of the psychology of viewing a print, not the physics of making a negative.

When people look at a photograph they make subconscious assumptions about how the scale of objects on the paper translates to the scale of the photographed objects in the real world. Much of the time, they imagine themselves to be looking through a window of the same dimensions as the photograph, and *that*'s why wide angle views look so sweeping, and long lens shots look so compressed. The viewer assumes that a normal lens must have been used, and constructs a false, overly dramatic model of the scene in their head.

This is easiest to observe if you make a large print of a wide angle shot. From a 'normal' viewing distance, or from across the room, you get the usual wide angle effect of sweeping drama. Get close though, and the view starts to appear more and more normal, especially if you close one eye to restrict your brain's depth cues. Once you get close enough that the image takes up the same field of view as did the original scene, the photograph no longer looks wide-angle at all.

So my answer to David's original question would be that yes, long lenses do compress 'perspective', if you use 'perspective' in a more figurative sense than the purely geometric. When you move between formats you change the definition of 'long', but the same effect pertains nevertheless.

[Doug Kerr]

Hi, David,

The phenomenon of "compression of distant objects" is not a function of lens focal length, although focal length is one parameter in the chain that can bring that phenomenon into play for us. In fact, we are really talking about a special implication of the phenomenon of perspective, one that leads to an optical illusion.

Here is a numerical example that shows how and why this happens.

Imagine that our subject is a pair of telephone poles, each 30' in height, one 500' from the camera and one 132' beyond that. There is an auto near the base of the near pole (this helps our visual system determine "scale".)

To our human vision, the nearer one subtends a vertical angle of about 3.44º and the farther one a vertical angle of about 2.72°. (This assumes our eye at ground level, a close enough approximation for our purposes - we will assume similar geometry for the rest of the example.)

Based on what we know about telephone poles, then (ideally), with the help of the "scale" cue from the automobile, our perceptual system will interpret what we see as something like two 30' telephone poles at distances of about 500' and about 632' - separated by 132' (as they are). This part of the world will "look natural" to us.

Suppose we take a shot of this scene from this same point, and then crop an area including the two poles on the negative and print it so that the near pole is 6.00" high on the print. The far pole will be 4.75" high. It does not matter what focal length lens we use - we must just choose the printing enlargement to make the near pole 3" high on the print. The rest of the story will work the same way regardless of focal length.

We then regard the print from a distance of 18". The image of the near pole will subtend an angle of 19.4° and that of the far pole, 15.3°.

Based on what we know about telephone poles, then (ideally), with the help of the scale cue from the automobile, our perceptual system will treat the image of the near pole as if it were a real 30' pole at a distance of about 90', and the image of the far pole as if it were a real 30' pole at a distance of about 114' - separated by a distance of about 24'. Thus, to the observer, the distance between the two poles seems "compressed" - we expect telephone poles to be further apart than that (like about 132').

In each case, the ratio of the apparent distances to the two poles is 5:6. Thus the apparent distance between them would be about 0.26 times the distance to the near one. That means that if we think the first pole is about 500' away (as it really is), then the distance between them would seem to be about 132' (as it really is); if we think the near pole is 90' away, then the distance between them would seem to be about 24' (which it isn't).

Then important fact here is that it works this way regardless of the focal length of the lens used, so long as in each case the magnification of the print from the negative is chosen to make the image of the near pole the same size (6" high in this specific example) and we view the image from the same distance.

Best regards,

Doug

[Jim Rice]

To my mind it's all angle of view. Draw two 'vees', one tall and skinny, and one fatter. The skinny one is the angle of view of the longer lens. There is a foreground subject. Vary your distance from this subject until it is the same size with both angles of view (lenses). Now consider the background. It is a finite though significant distance away. Extend your 'vees' out to it. With the longer lens you are including a much smaller swath of it (thus magnifying it relative to the foreground subject). There's your compression. The fat 'vee' (shorter lens) is going to include a much greater swath of background, thus minifying it. There's your expansion. I realize that this is a simplified explanation with no actual math and all, so you are all welcome to roll your eyes. :P

[Jim Rice]

To my mind it's all angle of view. Draw two 'vees', one tall and skinny, and one fatter. The skinny one is the angle of view of the longer lens. There is a foreground subject. Vary your distance from this subject until it is the same size with both angles of view (lenses). Now consider the background. It is a finite though significant distance away. Extend your 'vees' out to it. With the longer lens you are including a much smaller swath of it (thus magnifying it relative to the foreground subject). There's your compression. The fat 'vee' (shorter lens) is going to include a much greater swath of background, thus minifying it. There's your expansion. I realize that this is a simplified explanation with no actual math and all, so you are all welcome to roll your eyes. :P

[robc]

David,

The way you describe your thinking is almost correct except you need to think about it in terms of the object size on the film. Forget about angle of view for this. For a 300mm lens to give the same object size on film as a 50mm lens, regardless of film format, the 300mm lens needs to be further away from the subject to compensate for the additional magnification of the 300mm lens. It is the altering of distance from the subject which alters the perspective and relative sizes of different parts of the subject at different distances from the lens.

This is very well illustrated in AA's "The Negative" in the chapter on "artificial light photography". There is an image of two pianists (fig: 7-14 in my edition) and a longer lens was used to keep the relative sizes of the two pianists fairly equal on film. Had a shorter lens been used the camera would need to be closer to the subjects thereby making the nearer pianist look much bigger than the more distant pianist.

So the selection of lense focal length is a very important tool in adjusting the relative importance of different parts of the subject according to your viewpoint.

[C. D. Keth]

Think about this differently. The focal length has nothing to do with compression or, more properly, perspective. Zilch, zero, nada. That's determined entirely by where you stand. The angle of view that the camera can "see" is determined by the format size combined with the lens focal length. It's that simple.

Ding ding ding! Correct!

Say we're talking contact prints from 8x10 negs for simplicity. For reference, that is a 12.8 inch diagonal, which is also going to be considered the standard viewing distance for these prints. Now, if you take a negative with a 12.8 inch focal length, the perepective will look realistic to the eye when viewed from 12.8 inches from the print. If a shorter focal length is used but the viewing distance remains constant, the perspective will be exagerated. If, however, the viewing distance is always equal to the focal length in this situation, the perspective will always appear normal.

[Leonard Evens]

This sort of thing comes up regularly, and it is a source of much confusion and difference of opinion.

First, let me address the issue of perspective and point of view and focal length. It is true that geometric perspective, in the sense of what lines up with what and where vanishing points are, depends just on the position of the lens. But when you make a photographic print, it gets a bit more complicated. The relevant question is whether or not things will look the same to you when you look at the print as they would if your eye had been where the lens was. That will be the case only if you view the print from the appropriate position. Suppose, for example, that you take a picture using 6 x 7 (actually about 57 x 70 mm) format with a 90 mm lens. Suppose also that the subject is distant enough that you can consider it at infinity for all practical purposes. Then the film will be just about 90 mm from the lens. Suppose you make an 8 x 10 print from the negative by enlarging about 3.6 times. You should view the print from 3.6 x 90 mm ~ 320 mm for your eye to see what it would see by looking directly at the scene. Suppose instead, you make an 8 x 10 contact print using a 320 mm lens. Again your eye should be at 320 mm from the print for proper viewing. So, taking into account the final print, the perspective obtained with a 90 mm lens and 6 x 7 format is equivalent to a 320 mm lens and 8 x 10 format.

Now what happens if you use a lens that is much shorter than the normal focal length for the format. If you make an 8 x 10 print, you are unlikely to put your eye at the proper position for viewing the print given the focal length of the lens. You will likely view the print from what is a normal distance, i.e., about 300 mm, and you will be too far away. That will create some apparent distortions. Objects will look too small, and three dimensional objects at the edges of the field won't look quite right. S milarly, if the focal length of the lens is longer that the normal focal length, you are going to be viewing t he print from too close up. That will create an apparent contraction of distances and the feeling of compression.

That probably explains most of the reason why you see compression in such a print, but there may also be more to it than that. For example, with a long lens, the subject is magnified more than would be the case with a normal lens, and that by itself may exaggerate the sense of compression.

[C. D. Keth]

That probably explains most of the reason why you see compression in such a print, but there may also be more to it than that. For example, with a long lens, the subject is magnified more than would be the case with a normal lens, and that by itself may exaggerate the sense of compression.

To add to this:

Two objects equally spaced in depth from the lens will look closer when you are farther away than when you are close. Since for a given framing a long lens will put the lens far away from the subject, this also adds to the compressed space feeling.

[Jeffrey Sipress]

Conrad and Steve have said it all. It's really quite simple. Perspective, which is generally defined as the relationship among objects in a scene, is only determined by the camera-to-subject distance. Once you establish that to your liking, you then select a lens that gives you the angle of view you want. Photographic Craft 101, but it's amazing how many well known and longtime photographers don't know that.

[C. D. Keth]

Conrad and Steve have said it all. It's really quite simple. Perspective, which is generally defined as the relationship among objects in a scene, is only determined by the camera-to-subject distance. Once you establish that to your liking, you then select a lens that gives you the angle of view you want. Photographic Craft 101, but it's amazing how many well known and longtime photographers don't know that.


It's also surprising how many people teahing photography and film can't seem to explain it as well and concisely as you just did in two sentences. Nicely done.

Now go to a cinematography forum and explain to them that film format has no effect on depth of field. I think I answer that every other week.:eek:

[Alan Davenport]

The apparent compression of depth with long vs. short lenses, is a function not of the lens, but of the distance from which we view the image. A "normal" lens is so-called because its view subtends roughly the same angular view as will the resulting 8x10 print, when the print is viewed from the so-called standard distance.

With a lens that is substantially shorter or longer than normal, because we tend to view the final image at that same "normal" distance, the angular relationships between near and far subjects is distorted. But if you view a print so that it (the entire print) subtends the same angle to your eye, as did the subject to the camera, then the relationships will appear correct. This ignores binocular vision effects, which tend to compensate for the difference in viewing distance.

[Dominique Labrosse]

David,

Here is a tutorial that hopefully will clarify the answer for you...

http://www.montizambert.com/education/articles/articles.htm

Regards,
Dominique