Quick and Drity Crop Factor Test

[Anton Shavlik]

There's some talk on the forums recently about whether format size is an irreversible factor in what kinds of images you can create. I maintain it's not.

Below are two shots, taken moments apart with the camera in the exact same position, white-balanced to the same point, with these settings:

18mm f4.0 at 4k FF with a motion mount setting of 4 stops ND.
35mm f8.0 at 8k FF with a motion mount setting of 2 stops ND.

Note: these are not necessarily shown in the same order that I listed the settings.

The only post was to scale the 8k to 4k, and use photoshop to type in known lens distortion profiles to correct barrel distortion. The sliders were not tweaked to taste, I just looked up the numbers and typed them in blindly. 4k -> 8k is 2x, while 18 -> 35 is a little less, so the 8k image was zoomed 8% which matched crop factors.

There are certainly tells for which is which resolution, but optically it's hard to spot any sign of 'compression'. I'd like to do another test for depth of field later.

 

[Ilya O.]

I frankly did not get the counter point. What's it? Like wider focal lens introduce less compression or change the perspective?

As for the perspective, it's all the same, irrelevant to focal length, it just depends on the camera to object position. It's just that simple.
If camera stays the same position and you change the lenses from let's say 18 to 100 mm you just change the field of view, not the perspective itself. Object size relations and relative lines and angles would be the same.

We usually in the conventional conversation say perspective changes (wides) or the image is compressed (tele lenses) but we mean actually for the wides that the objects appear to move and increase in size very quick if go to-/outwards the camera, or in case of tele we mean that the image is flat and 2d and the object is moving slow — that's just the effects of field of view narrowing not the perspective change.

 

[Aaron Lochert]

I frankly did not get the counter point. What's it? Like wider focal lens introduce less compression or change the perspective?

That was exactly it. We've been over it time and time again, but it's evident that there's still a huge misunderstanding among the forums. There's a belief that the move to larger formats is motivated by some sort of large format "mojo," which somehow allows for wider fields of view with more compression. Meaning that wide fields of view look more "normal" and less "distorted" because the compression is stronger.

This stems from the belief that the lenses are the ones determining perspective -- this idea that equivalent FoV's turn out more "compressed" because of the longer focal lengths needed when compared to a smaller format.

But we know that the lens does nothing to perspective and neither does the format.

 

[James Falco]

But we know that the lens does nothing to perspective and neither does the format.

What do you mean with:The lens does nothing to perspective?
If that was true the term perspective distortion would not exist.
I don't get the point of this exercise, as the key factor is that the image was altered in photoshop, de-barrelizng it.
Also this trick works only well with subjects far away from camera.

 

[Anton Shavlik]

We usually in the conventional conversation say perspective changes (wides) or the image is compressed (tele lenses) but we mean actually for the wides that the objects appear to move and increase in size very quick if go to-/outwards the camera, or in case of tele we mean that the image is flat and 2d and the object is moving slow — that's just the effects of field of view narrowing not the perspective change.

You're exactly correct but I just want to nit pick here, whether tele or wide the size of the object changes at the same speed if it moves between the same absolute distances. I think you're assuming that on the tele the object is starting at a further distance. Of course you know that but I don't want anyone reading to think anything about perspective changes between lenses.

What do you mean with:The lens does nothing to perspective?
If that was true the term perspective distortion would not exist.
I don't get the point of this exercise, as the key factor is that the image was altered in photoshop, de-barrelizng it.
Also this trick works only well with subjects far away from camera.

Perspective distortion does not exist, that's exactly what I demonstrated. I really really suggest you try for yourself. I'm not a crank, it just drives me crazy that this is such a wide-held misconception.

Barrel distortion isn't an alteration of perspective as it can be corrected to match a lens with no distortion exactly, whereas shooting from different distances produces images which cannot be manipulated in that way, it would be impossible to match them. Barrel distortion is a different projection of the perspective, not a different perspective in and of itself, much like world maps come in different projections.

I will post another test with larger apertures and macro distances soon, and you will see that once again the images match. I'll use better lenses to avoid barrel distortion.

That was exactly it. We've been over it time and time again, but it's evident that there's still a huge misunderstanding among the forums..

I think it's hard to sort out because many people shoot on primes, where it's not obvious that an optical zoom (changing focal length) and a digital zoom (changing format size) have equivalent effects on an image, ignoring depth of field and diffraction. This is compounded with specific lenses having more or less distortion in their projection so that people get a sense that certain focal lengths are to blame, when companies like Zeiss have several excellent wides that cover super35 or more with virtually no distortion. There's also teles (most of them zooms) with barrel distortion, but I'm guessing those with this misconception don't use zooms to begin with.

It's also hard to combat because in certain circles it's treated as common knowledge so you look like a crank for questioning it, much like that guy who claims lenses with more than 7 elements cannot render perspective naturally that made its rounds in camera/lens forums.

 

[Aaron Lochert]

What do you mean with:The lens does nothing to perspective?
If that was true the term perspective distortion would not exist.
I don't get the point of this exercise, as the key factor is that the image was altered in photoshop, de-barrelizng it.
Also this trick works only well with subjects far away from camera.

Perspective distortion is 100% a product of camera-to-subject distance.

If you don't move the camera or the subject, then the subject's relative size to its surroundings remains the same, regardless of focal length.

This is why we usually prefer to dolly-in rather than zoom in. A zoom will do nothing to perspective and just scale up the entire image (which has it's own look and aesthetic). A dolly will change the perspective because the camera is moving.

Anyone confused should watch this:

https://youtu.be/GbESqYkodmQ

Further demonstration:

https://www.youtube.com/watch?v=IoQOaPZz-nI

Make note of the guys' relative size to the woman in front. In the dolly, you'll see that they are the same size as as the distance from the bottom of her chin to the bottom of the splotches on her shit. At the end of the dolly, you'll see they are much smaller, relatively - middle of her shirt to top of her shoulders. In the zoom, you'll see they are the same size relative to the woman at all times, regardless of the focal length.