Lenses for photographing people...

Started Apr 8, 2013 | Discussions thread
OniMirage
Contributing MemberPosts: 990
Like?
Re: Lenses for photographing people...
In reply to clengman, Apr 12, 2013

clengman wrote:

OniMirage wrote:

clengman wrote:

OniMirage wrote:

clengman wrote:

OniMirage wrote:

clengman wrote:

OniMirage wrote:

clengman wrote:

OniMirage wrote:

clengman wrote:


Yup. I feel for you. It can be exhausting to be so, so wrong. Sorry if your brain hurts.

I like this link from the other thread,

http://www.journalofvision.org/content/12/5/8.full

I like it, too. But you didn't really read it did you? It has a lot to say about the differing perspective effects that result from varying subject distance, and how it's important to match the viewing distance to the perceived subject distance in order to create a "normal" viewing experience.

I like this quote from that article:

"The apparent expansions and compressions in depth are often called                perspective distortion, as if these effects are due to a distortion in the physical projection from the scene to the film plane. The effects occur, however, when the projections are geometrically correct."

This one too:

"If the viewer's eye is positioned at the picture's COP (center of projection), the image cast by the picture onto the retina matches the image that would be cast by the original scene."

It's similar to what I was trying to explain to you here:

"If you took the image captured by the camera and printed it and displayed it so that the print covered the same real angle of view in your field of vision as the angle of view represented in the print, it would look almost identical to your perception of the subject when you put your face all up in your subject's face. (I say almost identical because there are aspects of your perception resulting from parallax and binocular vision that obviously wouldn't be present.)"

and here:

"If you could freeze your entire field of vision, print it out, hang it on the wall and look at it, you would see that it looks very much like a photo taken with a very wide angle lens. Only the central few degrees of the image would be sharp, but the "distortion" would be the same."

For that same reason you ignore that they agreed with the affects wider angles give and the fact that 50mm and above tends to be more pleasing/ideal/correct

I didn't ignore anything. There's nothing I have said that contradicts anything in that article. I never said that wide angle shots don't often appear "distorted" (I don't like to use "distorted" here. I prefer to think of wide angle shots as having an unusual perspective.) I never said that wide focal lengths are ideal for head and shoulders portraits.

I had two quibbles with your first post that I responded to. One stemmed from your poor grasp of symbolic math. Secondly I disagree completely with your wording "It has only to do with rendering the subject accurately without distortion." As I've been trying to explain (and the article backs me up) 1. A rectilinear wide lens renders a scene accurately, 2. "Distortion" is subjective and  3. Perspective effects result from subject distance and angle of view.

"Perspective distortion" is not an inaccurate rendering of the scene and it's not due to any aberrant properties of wide lenses. The perspective effects that you notice are a result of a completely accurate rendering of a scene using a) a closer than normal subject distance or b) a wider than usual angle of view or both. I tried to point out that you can see the same perspective effects without a camera if you put your nose really close to an object, or if you really pay attention to the appearance of large objects at the periphery of your visual field.

If you had instead written "It has only to do with rendering the subject in a pleasing fashion with perspective that corresponds with that of comfortable interpersonal standing distance." (Or something like that...) I would have had nothing to argue about. (Except I still would have wondered why you thought the ideal portrait focal lengths were less than 50mm.)

I feel like this is getting repetitive. I can't really think of any better way to explain this or I would try. If you still don't understand, I guess I'm sorry I wasn't able to make it more clear.

How about I just leave this here then, it confirms without words, it animates it for you and you can come to your own conclusion.

Fine, it's a good illustration of how perspective changes as the camera to subject distance changes. It also confirms that you still don't really even understand what we're arguing about and that you completely missed the point of the article you linked to.

We are only even arguing because you had an issue with the words used and the fact I used a number with a symbol which you disagreed with.

I was just pointing out that the words you used were incorrect.

Now, back to the original argument. In this animation, perspective effects are only changed by changing camera to subject distance, and the only effect of changing focal length in this example is to change the angle of view.

In other terms (camera-subject distance in first frame) < (camera-subject distance in last frame) AND (angle of view in first frame) > (angle of view in last frame). These two facts along with fact that the viewing distance (or output magnification or however you want to think about it) is constant for the whole animation, account entirely for one's perception that in the first frame the cube appears to be stretched and in the last frame it appears to be compressed front to back.

Which can make a persons face appear distorted.

Yes, out of context, it will be perceived as being distorted.

A short focal length rectilinear lens does not distort. It creates a correct rectilinear projection of whatever scene you point it at.

The shorter focal lengths don't look less accurate to represent the model at all, nope.

The shot taken at the closest subject distance appears distorted because it's displayed at a much smaller angle of view than what is represented in the photo.

If that photo was displayed so that the frame diagonal occupied about 97 degrees in your field of vision (This is the diagonal angle of view of a rectilinear 19mm lens on a 35mm camera.) it would look completely normal. Another way to say it is you can print the photo at life size and view from the same distance as the original focus distance (about 10 inches or so). You could also print it at 3x life size and view it at about 30 inches. This is what they call "viewing from the center of projection" in the article linked to above.

My original point was and has always been that wide angle lenses don't distort. They create an accurate projection of subject.

Changing the subject distance changes the perspective.

Viewing the photo from a point other than the center of projection can lead to a sense (subtle or not so subtle) that the subject is distorted.

I think it's interesting on a couple levels. From a practical standpoint for photographers, it gives a concrete basis for the use of display size, viewing distance, subject distance etc to play around with this perceived distortion.

For me, I have kind of a layman's interest in the neural basis of cognition and perception. I love Steve Pinker's books and other writers of his ilk. Basically, I just find this sh*t fascinating.

This is all well and good but the point of the op was lenses for people with examples provided at varying though mostly long focal lengths almost all of which would be considered head shots. For this reason portraits would be the likely end result. I have many times in the past suggested any lens could be used for anything I will always say this but this is very specific with very specific examples provided by the op. In the interest of explaining what the affects were and answering why people suggest there is an "ideal lens" that is why I responded the way I did. Your not going to get people do stand off center to view something unless there was some way you could force them to do it and had the space needed to recreate an accurate viewing angle and viewing distance. In that way what we perceive as correct is based on what we perceive as normal and normally people don't pay attention to or notice what distance does to how an object is viewed.

Quick clarification, my "97 degree diagonal angle of view" doesn't refer to standing off-center. I mean standing on center and at a distance so that most of your visual field is occupied by the photo. I'd say that's easy to accomplish in a gallery setting, or  maybe for a billboard-sized ad.

One implication is that if you wanted to display, for instance, a much larger than life-size portrait in a gallery and you know that the viewing distance will be somewhat close, and you want the photo to appear to have "normal" perspective, the photo should be taken with a wide angle lens. You can even calculate to some degree given the desired display size, viewing distance and framing you want for the final output, what subject distance and focal length you would need to use when taking the picture to make the perspective in the final output appear "normal."

In these cases it isn't so much the image that is taken but the perspective the artist wants to create.

Apart from that, I agree, if you want to print the normal 5x7 to 11x14 prints for display on the wall in your home, focal lengths longer than 50mm (35mm equivalent) will probably produce the most "normal-looking" faces.

I agree.

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow