The Curious Case of Lens Compression

I remember when I started on my journey to become a photographer, I came across the phrase “lens compression” – but it wasn’t explained.

With an engineer’s background, I readily grasped such concepts as the field of view is wider for lenses that are shorter– because if you take a paper tube, and look through it, the longer the tube, the less field of view you can see at the end. Gotcha.

But I still couldn’t find a good example of this strange phenomenon that people kept talking about – lens compression.

Lens compression is what happens when you take a picture with a longer lens – it apparently compresses the foreground and the background, and makes stuff appear closer together than in real life. With wide-angle lenses, the opposite effect occurs.

For example, let’s take a look at two pictures:

The first photograph was taken with a 24-70 lens, set at 52mm. You can see how far away the guy in the back appears.

The second photograph was taken with a 70-200 lens, set at 130mm. In order to have the front model be the same size, I had to back up quite a bit.

Look at how much closer the guy in the back appears to be to the front model. Notice he is also larger (which some might think is a function of him being closer, but he actually did not move).

I remember a conversation I was listening to with Don Giannatti – I know, I quote him quite a bit, but he was here recently, and he talks a lot (which is good). Anyway, he was looking at a picture of someone, and reverse engineering the photo.

“Well, we can see a lot of the ears, so obviously this was taken with a longer lens,”   Don said.

Whoa! Eureka moment. As you back up your viewpoint, yet magnify what you see, your subject stays the same size; however, the edges of your field of view become closer and closer to parallel.

So, if you imagine two rays shooting out from our eyeballs, in the long-lens scenario, the rays are staying fairly close together from our head, past the first model, and to the second model. Ergo, he looks similar in size to the front model. With the smaller lens, the rays from our eyeballs fan out, so the back model occupies a smaller angle of view compared to the first long-lens example, so he appears smaller. With the lens, the rays are the left and right side of the lens view.

So, extending the principle, if we got a very long lens, and backed up very far, the model in the back would appear to be the same size as the model in the front. The amount of compression is related to the relative distance between the front and back model, and the viewer (the camera, in this case). The same principle applies when using distance ratios for lighting.

The same principle works when shooting portraits. Observe the following four portraits:

The first was shot at 24mm, then 50, 70, 200 mm. I moved physically backwards to keep the model the same (approximate) size.

You’ll notice that in the first picture, her face is, for lack of a better term, bulbously distorted.   Maybe there is a culture somewhere in the world where this is attractive, but in most Western cultures, I believe that flatter faces are viewed as more attractive. The compression of the lens allows for a flatter (front to back) face. If her ears were fully visible, you would be able to see more of the ears in the last photo than the first photo.

As an interesting aside, you’ll notice that her waist appears smaller in the first photo than the last one. Hmm. Perhaps that’s a way to differentiate your portrait services – “We can make your waist smaller!   Of course, your nose looks huge, but hey! Everybody will be impressed with your waist size, they might not notice, eh?”   (Maybe that marketing approach needs work.)

I’m not sure if this is related (leave a comment if you have any thoughts), but when I shoot with my 70-200 set at 70mm, with the camera rotated to portrait orientation, my eyes will “snap” into focus and both eyes see the same thing at the same resolution. What puzzles me is that I was always told that 50mm is the same “focal length” as our eyes – but with the 50mm mounted on the camera, what I see through the viewfinder is smaller than what I see with my open eye. At first I assumed it was because the viewfinder itself is magnifying – but if that’s the case, then what’s through the viewfinder is larger than life — which makes the 70mm even more unlikely to match what I see with my open eye. Bizarre. I’d love an explanation of this.

This Post Has 19 Comments

  1. Here we go again. It’s not “Lens Compression”… it’s purely a factor of the working distance. Long lens, long distance… short lens, shorter working distance. If you were to shoot a far away subject with a short lens, and then blow the image up to to same subject size as one made from the same spot with a long lens, the apparent compression in the image would be identical. Of course, the enlarged image would show quality loss due to the enlargement.

    The apparent distortion in the close-up portrait is simply because you are too close to the subject… the more flattering result is because you are further away, regardless of the lens used.

  2. You’re wrong Bob. For a succinct example of how angle of view (compression) works, simply look to some “vertigo” scenes in movies. They simultaneously move closer to the subject while zooming out (or vice versa), and you get a very familiar contrazoom/dolly zoom effect:

    http://www.youtube.com/watch?v=Y48R6-iIYHs

    Compression in action. Also, think about those pictures of a huge sun setting or moon rising over the hill. Those are taken with superlong telephoto lenses to “flatten” the scene.

  3. Oh dear. I find these concepts so hard to grasp! It is amazing information though and as I reread it, I think I’m starting to understand…
    I have a friend who is an engineer and understands these things automatically. I have always envied people’s ability to grasp such concepts with apparent ease. I come from a creative background and the technical side of photography is a mindblowingly difficult challenge for me! But the visual examples you included are a great help and it makes perfect sense!
    Thank you for opening my mind Bill!

  4. Bob isn’t wrong at all. With the “vertigo” scenes mentioned by Mahmoud, the effect is produced by the dolly moving towards/away from the subject, thus changing perspective. All the zoom does is keep the subject the same size, which is what makes it seem so unnatural.

    When photographing sunsets, the Sun appears about the same size wherever you go. However, if you want a foreground element as a silhouette, e.g. a tree, then you can move closer and further away from the tree to make it bigger and smaller, while the Sun remains the same apparent size.

    For example, say you want to photograph the sun with a tree which is close to you. You want the tree to fill the frame so you zoom in, say, 2x. This is great but the sun is only 2x bigger than normal.
    You then back up several hundred metres and photograph the same tree with the sun behind it, but because the tree is so far away you have to zoom in much more to get it to fill the frame. This time you have to zoom, say, 20x. The tree looks more or less the same but the sun is now 20x bigger than normal, compared to 2x in the previous shot, so the two pictures look completely different.

  5. Coincidentally I just noticed the 70mm thing about 15 minutes before I found this article. I was playing with a superzoom “holiday” lens and I noticed that when I kept both eyes open I could zoom to where the left and right eye see essentially the same thing. When I took the camera away from my eye it was set to 70mm. Btw this was with a Canon XTi and a Tamron 18-270

    1. Andrew, the Canon XTi has a cropped sensor so at 50mm on an APS-C (cropped) sensor it appears 1.6x larger, therefore appearing 80mm… 100mm = 160mm and 45mm would be around 70mm

  6. The 70mm vs 50mm thing… There is a difference between focal length and equal size.

    I believe classic 50mm is the same focal length. When you talk about seeing the same thing it is a question about more than just focal length. Probably mostly magnification.

    When we talk about the same focal length as the naked human eye you have to compare elements in the image to each other. I.e. two elements are placed at different distances. If they look like 1:2 with the eye, they should do the same through the viewfinder. Then you have the human focal length, but not necessarily field of view or magnification…

    Or compare with prints. Make one small print and one huge print. They still both have the same focal length. But one apears much bigger than the other. 😉

    Or actually you can take any print with any lens. Go back to the original scene. At some distance it appears the same as the original. Magnification!

  7. AFAIK 50mm is about the diagonal size of a full frame, while 35mm is about the diagonal size of APS-C. The focal length is considered “normal” when it’s about the same as the diagonal size as the film (or sensor) format. This has actually nothing to with seeing the same thing through your viewfinder and your naked eye, as it varies from camera to camera, depending on the magnification of the viewfinder (try your 70mm on a pentamirror vs. pentaprism).

  8. … oh, sorry, I almost forgot: thanks a lot. Never tried this out… I knew this theoretically, but personally, I think, the pictures you provide are somewhat of an eye-opener.

  9. May I have permission to use these example photos in a presentation about Lens Compression?

  10. Bob, did you not see the first example? The person in the background is twice the size that he was in the first shot while the primary model is only slightly larger, If you resized the first shot to make the male model the same size that he is in the second, you’d only see the female model’s abdomen (not to mention the building in the background which is about 4x as large).

  11. Matt, Bob is correct. The first two example pics were taken from two different distances (a requirement to make the foreground model appear roughly the same size in the frame). Bob’s hypothetical example states that the two pictures need to be taken from the same spot.

    It is a confusing topic, and one can utilize the lens compression effect successfully while remaining mistaken about the nature of said effect. If you have a basic understanding of optical physics, the following PDF very succinctly explains the phenomena correctly as a function of the distance to the subject:

    http://scubageek.com/articles/compression.pdf

  12. Hi All. This site is amazing. And thanks for your comments on compression. I thought I knew what it was and I did for the most part but this solidified it. Another article mentioned you can visualize the same effect by looking through, say, a toilet paper roll vs. a paper towel roll if you maintain a distance that keeps your subject with the same composition with the two.

    Jesse, if you’re not already familiar with the concept, which you probably are, you should test it with your own shots of possible familiar subjects to your audience. Just a thought as the relevance may have more impact.

  13. “please note that all focal lengths are in 35mm equivalent. That means If you’re not using a film camera, D700, D3, 5D MkII, or 1Ds… you need to multiply these numbers by roughly 1.5x. A 50mm lens on a Nikon D90, for example, would look the same as a 75mm lens would look on a film camera.” Found this as I was researching this topic online and remembered your question. (Source:
    http://www.seanmolin.com/blog/?p=69)

  14. Great to see the side-by-side comparisons. It’s really helpful to be able to visualise this as I’m just starting to use a telephoto now and have yet to use it for portraits.

    Though, I must admit I’m very confused now over “Lens Compression” from all the comments arguing both sides. If this isn’t lens compression, then what is? :/

  15. Bob is completely off. He seems to be talking about distortion in his compression argument as well, two completely different things. If you look up compression examples on the web you’ll see the exact opposite of what he’s saying. He sounds as if he has very little understanding of camera lenses and what they do.

  16. Just found this while searching for examples of how moving away while zooming in changes perspective and depth of field. Nice photos! I was wondering though… did you leave the aperture the same or did you let the camera adjust it for you? The telephoto pictures have a shallower DOF but I thought that the increased distance between you and the subject would cancel that out. If you could tell me whether the aperture stayed the same I’d be closer to solving this mystery 🙂

    @Sarah: Bob isn’t completely off. The change in size between the girl and the guy is indeed due to the photographer changing his position in addition to zooming in.

Leave a Reply

Close Menu