The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Lens compression, straightforward

Shashin

Well-known member
Actually, the author is wrong. It has nothing to with how far something is from the background. It has nothing to do with camera position or focal length. It has nothing to do with linear perspective at all.

The effect of compression, or expansion, the roundness and depth you see with "wide-angle" images, is the result of viewing angle. If your viewing angle, the angular size of the image from where you are standing (or sitting), is the same as the angular field of view of the camera, then the image will look normal with no compression effects. If your viewing angle is great than the taking (camera) angle, then you will perceive compression. If your viewing angle is less than the taking angle, then you perceive more depth. This is also known as apparent perspective.

And this is the only reason we perceive this effect in photography. It is the only way we can see the world at a distance different from where we are (or would be). This is easy to prove. When you are at a party, the person standing across the room does not look any more compressed than the person standing next to you. If you look at a tree trunk with binoculars, the trunk looks flat, like a cut out, but, if you view without binoculars, the roundness of the trunk looks natural. It is simply an angular difference in the views. And it also works for prints. If you are in a gallery, standing next to a large print, the image will look compressed. Walk on the other side of the room and view the same print, and the print will appear to have more depth.

Compression is a perceptual effect of the viewer.
 

Jorgen Udvang

Subscriber Member
Actually, the author is wrong. It has nothing to with how far something is from the background. It has nothing to do with camera position or focal length. It has nothing to do with linear perspective at all.

The effect of compression, or expansion, the roundness and depth you see with "wide-angle" images, is the result of viewing angle. If your viewing angle, the angular size of the image from where you are standing (or sitting), is the same as the angular field of view of the camera, then the image will look normal with no compression effects. If your viewing angle is great than the taking (camera) angle, then you will perceive compression. If your viewing angle is less than the taking angle, then you perceive more depth. This is also known as apparent perspective.

And this is the only reason we perceive this effect in photography. It is the only way we can see the world at a distance different from where we are (or would be). This is easy to prove. When you are at a party, the person standing across the room does not look any more compressed than the person standing next to you. If you look at a tree trunk with binoculars, the trunk looks flat, like a cut out, but, if you view without binoculars, the roundness of the trunk looks natural. It is simply an angular difference in the views. And it also works for prints. If you are in a gallery, standing next to a large print, the image will look compressed. Walk on the other side of the room and view the same print, and the print will appear to have more depth.

Compression is a perceptual effect of the viewer.
I'm not sure what you are talking about. He talks about different objects in relation to each other, viewed at a straight angle. I have done exactly the same test as he did and I came to the same conclusion.
 

Shashin

Well-known member
I'm not sure what you are talking about. He talks about different objects in relation to each other, viewed at a straight angle. I have done exactly the same test as he did and I came to the same conclusion.
OK. But correlation is not causation. In the last example, he is just changing image size. There is no change in the ratios of object sizes, so what is causing the perception of compression?
 

Jorgen Udvang

Subscriber Member
OK. But correlation is not causation. In the last example, he is just changing image size. There is no change in the ratios of object sizes, so what is causing the perception of compression?
Distance. The longer the distance from the viewer to the first subject, the shorter the relative distance between the first subject and any subject behind it. It's the same effect you see when trekking. When there's still a few kilometres left to the first peak, the second one seems to be right behind it. When you approach the first one, you realise that the second one is much farther away than you imagined because the relative distances between the two and you change.

I used to struggle with this when I was a kid. During Easter Holiday, we used to stay at a forest cabin north of Oslo. There's one peak that can be seen very clearly from that cabin, as can most of the hilltops in front of it. When we started, it always looked as if the peak weren't too far behind the first hill and very close to the last visible one, while in reality, the last visible hilltop wasn't even halfway. Which I only realised after sprinting up hill number one, two and three. But I learned... I think :rolleyes:
 

Shashin

Well-known member
Distance. The longer the distance from the viewer to the first subject, the shorter the relative distance between the first subject and any subject behind it. It's the same effect you see when trekking. When there's still a few kilometres left to the first peak, the second one seems to be right behind it. When you approach the first one, you realise that the second one is much farther away than you imagined because the relative distances between the two and you change.
There is no distance in a photograph...
 

Jorgen Udvang

Subscriber Member
There is no distance in a photograph...
... but there's an illusion of distance. Actually, there's nothing in a photograph other than pixels or ink dots or differently exposed areas on a flat sheet of paper, but the illusion of reality can be great sometimes, although it's really just a kind of hoax :)

... or magic :) :) :)
 

Shashin

Well-known member
... but there's an illusion of distance. Actually, there's nothing in a photograph other than pixels or ink dots or differently exposed areas on a flat sheet of paper, but the illusion of reality can be great sometimes, although it's really just a kind of hoax :)

... or magic :) :) :)
Magic, that's it! (and it is magic)

But the lack of distance is why using linear perspective does not explain compression. And you are absolutely right in thinking photographs are illusions--it is our perception that creates them. The feeling of compression is a perceptual attribute. How do we get the visual information in order to experience that? That is where thinking in terms of angular size actually can solve the problem. It also has a technical name--apparent perspective.
 

MGrayson

Subscriber and Workshop Member
With a picture taken with any (non fisheye) lens, there is a distance form the picture where it will look natural. Stand very close to a wide angle photo and the distortions go away. Get very far from a telephoto shot and the compression looks like the natural compression you see looking down a long street into the distance. This is all pretty straightforward as projective geometry, but unless you're a mathematician or have studied technical drawing, perspective in art, or 3D computer graphics, the language is unfamiliar.

To be clear, it's not a function of what's IN the photograph. If the angle your eye makes with the edges of the photo matches the lens FoV, then the picture will look undistorted. Get closer, it looks compressed, get further, it looks WA distorted.

(Didn't I just tell Ken that I would only post pictures? That lasted 5 minutes... :chug:)

--Matt
 

Shashin

Well-known member
This is all pretty straightforward as projective geometry, but unless you're a mathematician or have studied technical drawing, perspective in art, or 3D computer graphics, the language is unfamiliar.
Or studied photography at a technical or scientific level. This is also well documented in photography and comes under the topic of viewing distance. It is just not normally part of the curriculum for photographers--unless you went to RIT.

It is so nice having a mathematician at GetDPI. I am going to have to look into projection geometry.
 
Last edited:

Robert Campbell

Well-known member
With a picture taken with any (non fisheye) lens, there is a distance form the picture where it will look natural. Stand very close to a wide angle photo and the distortions go away. Get very far from a telephoto shot and the compression looks like the natural compression you see looking down a long street into the distance. This is all pretty straightforward as projective geometry, but unless you're a mathematician or have studied technical drawing, perspective in art, or 3D computer graphics, the language is unfamiliar.


--Matt
Indeed; try taking a full face image with a very wide angle lens. At a 'normal' viewing distance it will seem very distorted. Go close to the image and it will appear unremarkable. (A 'normal' viewing distance for a book or a modest sized image is around 20 inches.)

This is a bit difficult to do on a monitor where you can't get close enough to focus. Try projecting the image onto a large screen or a white wall.

Perspective is also why lenses with focal lengths of around 85 - 90 mm are 'ideal portrait' lenses on 'full frame sensors' or 35mm film. What is really happening is that the photographer must stand some distance from the subject to get a head or head and shoulders to fill the frame, so that enlargement can be done with minimal loss of quality. From the same position, a wide-angle lens will include the subject's body and more; but enlarge just the head and shoulders, and you get the same perspective to the 'ideal' lens when the image sizes are the same and viewed from the same distance. But the enlargement will be grainy and of poor technical quality, which is why this isn't done.
 

Jorgen Udvang

Subscriber Member
Indeed; try taking a full face image with a very wide angle lens. At a 'normal' viewing distance it will seem very distorted. Go close to the image and it will appear unremarkable. (A 'normal' viewing distance for a book or a modest sized image is around 20 inches.)

This is a bit difficult to do on a monitor where you can't get close enough to focus. Try projecting the image onto a large screen or a white wall.

Perspective is also why lenses with focal lengths of around 85 - 90 mm are 'ideal portrait' lenses on 'full frame sensors' or 35mm film. What is really happening is that the photographer must stand some distance from the subject to get a head or head and shoulders to fill the frame, so that enlargement can be done with minimal loss of quality. From the same position, a wide-angle lens will include the subject's body and more; but enlarge just the head and shoulders, and you get the same perspective to the 'ideal' lens when the image sizes are the same and viewed from the same distance. But the enlargement will be grainy and of poor technical quality, which is why this isn't done.
Some lazy PJs I've met, actually did that, using a WA lens for most of their work. The resolution on newsprint back in the day was so bad anyway, that they could crop a lot before there was visible grain. The most extreme example I met was one who carried an F5 with a 35mm only, and did all his work with that lens. He did make a living from it.
 

Shashin

Well-known member
Perspective is also why lenses with focal lengths of around 85 - 90 mm are 'ideal portrait' lenses on 'full frame sensors' or 35mm film. What is really happening is that the photographer must stand some distance from the subject to get a head or head and shoulders to fill the frame, so that enlargement can be done with minimal loss of quality. From the same position, a wide-angle lens will include the subject's body and more; but enlarge just the head and shoulders, and you get the same perspective to the 'ideal' lens when the image sizes are the same and viewed from the same distance. But the enlargement will be grainy and of poor technical quality, which is why this isn't done.
That isn't what is happening. Most portraiture, 8x10s on a wall or on a page in a magazine, are viewed at distanced greater than normal. The short telephoto allows the apparent perspective to appear normal. And this is especially important with human faces. We are far more sensitive to "distortion" (or in this case, projection problems) in facial feature more than any other type of object.

This effect in compression is a problem in the viewing conditions. The shooting condition in and of themselves do not determine if or how much compression we will perceive, even though it is possible to determine that from those conditions. But then the viewing conditions need to be known. But this and other qualities related to the perception of the image, for example, depth of field, are perceptual properties of a viewer of an image, not the optical conditions when the image is taken.

Photographers find this counterintuitive because they are always taught that lenses do this or that, when actually the simple geometric nature of photography just makes it easy to relate some of these shooting conditions to perception of an image: depth of field scales would be the classic example (which is also why so many photographs claim their DoF scales are not "right"). Of course, it is very useful to use this knowledge about focal length and aperture to help photographers visualize the results of their choices, but correlation is not causation. And it is a real pity we don't learn this.
 
Top