Doug Kerr
Well-known member
We often hear about perspective in connection with photography. We hear (incorrectly) that a change in focal length changes "the perspective" of the image. In a related area, we hear (incorrectly) that an increase in focal length causes the appearance of "compression" of the distances between objects at different distances.
"Perspective" in general language
In general language, we can take "perspective" as meaning "how we look at something". This actually dovetails well with its technical significance.
"Perspective" in a technical sense
In such fields as engineering drawing, technical illustration, landscape sketching, and photography, and in fact in regular human vision, perspective refers to the phenomenon by which objects in the scene of the same actual size will appear in the drawing or image with different sizes, depending on their distance (don't worry for now from what).
In fact, in photography, this inherently happens (with some very rare exceptions).
The properties of a perspective image
In a true perspective image (and we get that in photography if the lens is "rectilinear"):
• For scene objects of the same real size, their size in the image is inversely proportional to their distance (in the scene) from a certain point, called the point of perspective.
• For points in the scene that lie along a line passing through the point of perspective, their images will fall at exactly the same point in the image.
The point of perspective is in fact the point from which the scene is "seen"; it is the location of the camera (the lens entrance pupil, to be precise) or of the "sketch artist" (human or virtual).
What will change "the perspective" of the image
We of course cannot answer this unless we know what is meant by "the perspective" of the image.
• Some people think that this "property" is given by the size of an object on the image. (Is that in pixels, or millimeters on the sensor, or fractions of the frame diagonal size, or what?)
• Some people think that this "property" is given by the field of view of the camera.
• Some people have some notion of "the perspective" that does not lend itself to description in any technical way, as we might speak of "the feel" of an image.
Any of these people should not expect to be comfortable with the remainder of this note.
The technical concept of perspective leads to a worthwhile definition.
Let me, however, put aside for a moment the matter of "an indicator" of "the perspective" of an image. We will sneak up on it by saying, "We have two images. How could we tell that they had 'the same perspective'?"
An outlook that fits well with the technical notion of perspective is this:
Two images would have "the same perspective" if:
• The relative sizes of the images of various objects (for objects appearing in both images) are the same in both images.
• The relative positions of of the images of various objects (for objects appearing in both images) are the same in both images.
Now, when will this happen? If, and only if the point of perspective is the same for the two images. Thus, the location of the point of perspective, and nothing else, determines "the perspective" of an image. (We assume that the aiming directions are generally consistent.)
Focal length
We hear often that "the perspective" of an image is varied if we vary the focal length of the lens used for the shot.
Now, if this is so, and we increase the focal length from one case to another, in which direction does that change "the perspective" of the image? Does it make it "greater" or "lesser"? Does it make it "hotter" or "cooler"? This conundrum should make us suspicious of the notion.
In fact, the focal length makes no change in the perspective of the image. That is only affected by the location of the point of perspective - by the location of the camera.
Why do so many people think that the focal length affects the perspective? Because the focal length that we choose may be determined by the distance to the subject, or vice versa.
Here's an example. We set up to do a portrait, and put a 35 mm lens on the camera. We place the camera where the subject will properly fill the frame, and shoot. We don't like the result, from a "perspective" basis (the nose looks too big, the ears are hidden, etc.)
We remember reading that many portrait photographers like an 85 mm lens, so we put one on. Well, now only part of the subject is in the frame. So we roll the camera back until the framing is essentially as it was with the 35 mm lens.
We shoot, and the result is very nice. "Wow!, we say. "That longer focal length really did the trick!"
No, it didn't. What did the trick was to roll the camera back. The point of perspective was now farther back from the subject, with a corresponding effect on the perspective effects.
Now suppose we didn't have an 85 mm lens, but kept the 35 mm lens on and moved the camera back as before. We shoot.
Of course, the subject is rather small in the frame. so we crop the image in post. And guess what! The image, with respect to perspective, is just like the one done with the 85 mm lens from that same point.
Now of course this is not necessarily an attractiveness technique. The image quality may be worse, because of the "fewer pixels for any object". But we see that is is not the focal length, as such, that controls the perspective effect.
These two images illustrate the situation. They were taken from the same point. The first one was shot with a focal length of 105 mm, the second with a focal length of 47 mm. I have only presented here a crop from the second shot that covers the same scope as the first image, and present them here at the same size to facilitate comparison.
There is no difference in "the perspective" of these two images in any way I understand that.
In the next part of this series, I will discuss the concept of "perspective angle" encountered in connection with CAD and technical illustration systems.
In the third part of this series, I will discuss the matter of the "visual compression" of objects said to be a result of the use of a greater focal length.
Best regards,
Doug
"Perspective" in general language
In general language, we can take "perspective" as meaning "how we look at something". This actually dovetails well with its technical significance.
"Perspective" in a technical sense
In such fields as engineering drawing, technical illustration, landscape sketching, and photography, and in fact in regular human vision, perspective refers to the phenomenon by which objects in the scene of the same actual size will appear in the drawing or image with different sizes, depending on their distance (don't worry for now from what).
In fact, in photography, this inherently happens (with some very rare exceptions).
The properties of a perspective image
In a true perspective image (and we get that in photography if the lens is "rectilinear"):
• For scene objects of the same real size, their size in the image is inversely proportional to their distance (in the scene) from a certain point, called the point of perspective.
• For points in the scene that lie along a line passing through the point of perspective, their images will fall at exactly the same point in the image.
The point of perspective is in fact the point from which the scene is "seen"; it is the location of the camera (the lens entrance pupil, to be precise) or of the "sketch artist" (human or virtual).
What will change "the perspective" of the image
We of course cannot answer this unless we know what is meant by "the perspective" of the image.
• Some people think that this "property" is given by the size of an object on the image. (Is that in pixels, or millimeters on the sensor, or fractions of the frame diagonal size, or what?)
• Some people think that this "property" is given by the field of view of the camera.
• Some people have some notion of "the perspective" that does not lend itself to description in any technical way, as we might speak of "the feel" of an image.
Any of these people should not expect to be comfortable with the remainder of this note.
The technical concept of perspective leads to a worthwhile definition.
Let me, however, put aside for a moment the matter of "an indicator" of "the perspective" of an image. We will sneak up on it by saying, "We have two images. How could we tell that they had 'the same perspective'?"
An outlook that fits well with the technical notion of perspective is this:
Two images would have "the same perspective" if:
• The relative sizes of the images of various objects (for objects appearing in both images) are the same in both images.
• The relative positions of of the images of various objects (for objects appearing in both images) are the same in both images.
In fact, in formal geometry, we would say that these two images are similar That is a rigorous definition; it does not mean "sort of alike".
Now, when will this happen? If, and only if the point of perspective is the same for the two images. Thus, the location of the point of perspective, and nothing else, determines "the perspective" of an image. (We assume that the aiming directions are generally consistent.)
Focal length
We hear often that "the perspective" of an image is varied if we vary the focal length of the lens used for the shot.
Now, if this is so, and we increase the focal length from one case to another, in which direction does that change "the perspective" of the image? Does it make it "greater" or "lesser"? Does it make it "hotter" or "cooler"? This conundrum should make us suspicious of the notion.
In fact, the focal length makes no change in the perspective of the image. That is only affected by the location of the point of perspective - by the location of the camera.
Why do so many people think that the focal length affects the perspective? Because the focal length that we choose may be determined by the distance to the subject, or vice versa.
Here's an example. We set up to do a portrait, and put a 35 mm lens on the camera. We place the camera where the subject will properly fill the frame, and shoot. We don't like the result, from a "perspective" basis (the nose looks too big, the ears are hidden, etc.)
We remember reading that many portrait photographers like an 85 mm lens, so we put one on. Well, now only part of the subject is in the frame. So we roll the camera back until the framing is essentially as it was with the 35 mm lens.
We shoot, and the result is very nice. "Wow!, we say. "That longer focal length really did the trick!"
No, it didn't. What did the trick was to roll the camera back. The point of perspective was now farther back from the subject, with a corresponding effect on the perspective effects.
Now suppose we didn't have an 85 mm lens, but kept the 35 mm lens on and moved the camera back as before. We shoot.
Of course, the subject is rather small in the frame. so we crop the image in post. And guess what! The image, with respect to perspective, is just like the one done with the 85 mm lens from that same point.
Now of course this is not necessarily an attractiveness technique. The image quality may be worse, because of the "fewer pixels for any object". But we see that is is not the focal length, as such, that controls the perspective effect.
These two images illustrate the situation. They were taken from the same point. The first one was shot with a focal length of 105 mm, the second with a focal length of 47 mm. I have only presented here a crop from the second shot that covers the same scope as the first image, and present them here at the same size to facilitate comparison.
There is no difference in "the perspective" of these two images in any way I understand that.
In the next part of this series, I will discuss the concept of "perspective angle" encountered in connection with CAD and technical illustration systems.
In the third part of this series, I will discuss the matter of the "visual compression" of objects said to be a result of the use of a greater focal length.
Best regards,
Doug