Distortion in Portraits with Nikon D300?

[anonymous]

I have a question that I haven't seen answered anywhere on the Web. I thought that maybe someone here would have an answer. It's a very practical question in fact, since it clearly affects my photography in the field.

I just bought a Nikon D300 with a 50mm 1.8 lens and an 85mm 1.8. When I did a portrait of my wife in landscape orientation (the usual for the camera), her face seems a bit larger than life, a bit stretched on the horizontal and less long in the vertical axis. When I rotate the camera in portrait orientation, she seems thinner than usual. I tried to keep the same distance for the two situations but it always does this strange effect. Of course, she likes looking thinner, but I was wondering why I can't make her look natural with the normal orientation of the camera. Is there a physical reason to that ? The effect seems worse with the 50mm than the 85mm. I put her face on the left side of the picture for a better composition while in landscape orientation. My distance to the subject is about 1 meter to 1.5 meter and a half with the 50mm, just enough to have only her face in the picture while in landscape orientation.

Thank you very much and please excuse my English, I'm a french native.

[McQ]

Other than a difference in distortion between lenses (which I believe is negligible with both these models), the primary cause of this observed difference ought to be from "perspective distortion." I put this in quotes because in some sense it is not really distortion as one would ordinarily think of it.

The key here is that one would have to get a little closer in portrait orientation to get the face filling the frame-- thereby exaggerating the perspective. This website gives an example of perspective exaggeration using two interactive diagrams on the page about understanding camera lenses and focal length.

This would also be consistent with your observation that it is worse with the 50mm lens. If one were to try a 35mm lens this effect ought to be even worse (the wider the lens the more perspective distortion). 135mm should eliminate this almost entirely. In your case, though, the 50 and 85mm are not normally considered wide angle lenses, but the effect is there with all lenses-- just to a *much* less pronounced degree with telephoto lenses. Hope this helps.

[wings4me]

as an owner of a D300, I totally agree with you.

[TonyBeach]

Other than a difference in distortion between lenses (which I believe is negligible with both these models), the primary cause of this observed difference ought to be from "perspective distortion." I put this in quotes because in some sense it is not really distortion as one would ordinarily think of it.

I believe linear distortion increases as you focus closer. Changing perspective will have some distorting effects though and cannot be eliminated as a consideration -- although not in this instance if we assume the OP only changed the orientation.

The key here is that one would have to get a little closer in portrait orientation to get the face filling the frame-- thereby exaggerating the perspective.

The OP said he was taking both shots from the same distance with the same lens. The focus distance might be a factor in distinguishing the 50mm from the 85mm, but lens design probably is also a factor.

This would also be consistent with your observation that it is worse with the 50mm lens. If one were to try a 35mm lens this effect ought to be even worse (the wider the lens the more perspective distortion). 135mm should eliminate this almost entirely. In your case, though, the 50 and 85mm are not normally considered wide angle lenses, but the effect is there with all lenses-- just to a *much* less pronounced degree with telephoto lenses.

Perspective distortion increases as we move away from a "normal" focal length. The "normal" perspective is achieved with the focal length of the lens that replicates what we normally see; that focal length is format dependent and on DX format it would be approximately 35mm. Therefore, I would argue that on DX format 85mm has more perspective distortion than 50mm has, since both are shrinking the foreground relative to the background (the inverse is true as you start using shorter than "normal" focal lengths).

Linear distortion is also consistent with the OP's observations, as the Nikkor 50/1.8 does have it (specifically it has barrel distortion). Here is a shot taken with the Nikkor 50/1.8 at close range and then cropped twice, once in the original landscape orientation (on the left) and then again after it was rotated to portrait orientation (on the right):



The image is indeed being stretched more along the long axis than the short axis (note the extra space on the bottom of the left crop and on the right side of the right crop), and that would cause the subject to look relatively wider along the long axis and thinner along the short axis. Now I know why it's called "landscape" and "portrait" orientation.

Some lenses actually have pincushion distortion, particularly zooms at their longer range. This thread has stirred my interest in this effect as these considerations could have significant ramifications for my portraiture photography.

I always suspected that shorter focal lengths could be used advantageously. I would go so far as to suggest that for some portrait subjects using a shorter focal length with careful framing could be the preferred approach.

As for the difference any camera model such as the D300 would play in this, that is a canard.

[Bill44]

Many years ago, and maybe still, 105mm on FF was considered the portrait lens. Now by portrait I mean a head and shoulders shot taken in the portrait orientation, not landscape.
So your 85mm lens on the D300 would be nearly the equivalent,
If you wish to only shoot her head the the advice that McQ has given is absolutely correct.

[shreds]

It is an interesting observation and Bill44 is correct in his view on the 'portrait' lens; my choice too.

(Indeed the Nikon f2.8 105mm Micro is probably my favourite lens for a variety of jobs and brilliant IQ at the present time.)

I will dig out my D200 (predecessor to the D300) and compare with the D3 using the same lens and see what the results are like.

In the meantime, as long as we recognise the fact, there is probably little in practical terms we can do about it, other than bear in mind that for flattering portraits, use the portrait format, which ordinarily I would do as a matter of course!

[_GUI_]

Perspective should not change if distance to your wife was exactly the same in both shots. The only difference I find could be between both lens is geometrical distortion, but I don't think is so determinant here to make a difference in the result.

BR

[JonathanC]

Maybe the pixels aren't quite square in the D300?