perspective w/ 35mm lenses?

Discussion in 'Digital Photography' started by PrincePete01, Jul 16, 2004.

  1. PrincePete01

    Jeremy Nixon Guest

    Except that I just plain, straight-up, don't see that.
    That doesn't make sense. If it's a spherical projection, surely it can be
    modeled as a sphere?
    They are both distorted in some way; the imaginary spherical image is also
    "flat" on its curved surface, so that must be distorted as well. It will,
    however, preserve what you can see if viewed from its center. You say that
    a flat print will do that, too, if viewed from its optical origin, and I
    don't disbelieve you, but I cannot see that, and I've just been trying.
    But that's all I was saying. It's clearly not good in a general sense to
    describe something as "distortion-free" when that term already has another
    meaning in the field, though.
    Outside this rather theoretical discussion? Not much, I'm sure. :)
    Jeremy Nixon, Aug 5, 2004
    1. Advertisements

  2. PrincePete01

    Nostrobino Guest

    It doesn't, it's not and it can't. :)

    Dave has essentially retracted that "spherical projection" statement.

    [ . . . ]

    Jeremy, the problem here is that you're trying to see the world around you
    as the inside surface of a sphere, which it simply is not. It's a
    three-dimensional world, made up of three-dimensional objects, and all a
    rectilinear lens does is create a two-dimensional representation of those
    objects, whatever is in front of it that it's able to view.
    Nostrobino, Aug 5, 2004
    1. Advertisements

  3. PrincePete01

    Jeremy Nixon Guest

    Retracted? He never made it in the first place.

    Anyway, you're still the only person in the entire world who thinks that
    focal length affects perspective.
    And how does it do that? If it's not a spherical projection, what is it?
    You can't just say it makes a 2D representation of a 3D world; there has
    to be a method of doing that. Anyway, I'm not asking you, I'm basically
    done with you.
    Jeremy Nixon, Aug 5, 2004
  4. PrincePete01

    Jeremy Nixon Guest

    Okay, but, what I want to know, then, is how is that *not* a rectilinear
    projection, a method of mapping a sphere to a plane, in the mathematical
    Jeremy Nixon, Aug 5, 2004
  5. It isn't what I want, it's what I have observed. Simply asserting that
    this observation translates into some psychological defect on my part is
    unworthy of a genuine seeker after truth.
    It is clear that you have thought very deeply about this. However, the
    argument has arisen almost totally because you insist in describing a
    well-known phenomenon, which most respondents understand very well, in
    terms of a word that means something else.
    The error has been pointed out ad nauseam. You insist in simply
    repeating your thoughts, most of which are quite correct, but not
    accepting that you are using the wrong word to encompass it.
    I agree that you are clearly not in the business of caring what others
    think. Unfortunately, you insist on carrying this over into the meaning
    of photographic terms, which, of course, you cannot make mean what you
    want them to mean by any amount of argument.
    David Littlewood, Aug 5, 2004
  6. So, you ^have^ shifted your ground. It's not focal length which matters
    (you now say) but focal length in relation to film format. You did of
    course start out by saying that focal length (unqualified). A quote from
    you on (I think) 2 August):

    Clearly, focal length/sensor size affects the appearance of the picture.
    We call that "field of view". It also causes some issues with the
    appearance of peripheral objects. We call that "the wide angle effect":
    it is of course a distortion caused by the extreme geometry.

    Rather than go over all the issues, may I suggest you read the
    comprehensive entry in The Focal Encyclopedia of Photography under
    "Perspective". One pertinent quote on linear perspective* is:

    "Since short focal length wide angle lenses tend to be used with the
    camera relatively close to the subject and long focal length telephoto
    lenses tend to be used with the camera at relatively large distances,
    strong perspective is often associated with wide angle lenses and weak
    perspective is similarly associated with telephoto lenses, but it is the
    camera position and not the focal length or type of lens that produces
    the abnormal linear perspective."

    [*Yes, there are other elements of perspective - aerial (haze), colour,
    and possible depth of field - which give information about relative
    distances. These were omitted in my opening statement - which I regret
    to say sparked this whole thing off - because the were not relevant in
    the context of the then discussion.]
    All very well, but it still isn't PERSPECTIVE.
    David Littlewood, Aug 5, 2004
  7. One sometimes feels under an obligation to avoid others being misled by
    particularly obsessive errors. This is not unlike the issue with Preddy,
    i.e. why we sometimes have to respond, in case a newcomer gets confused.
    To be fair, though, your arguments are more logical and your sin rather
    more modest (being simply an error over the meaning of a word). It was,
    after all, my post to which your response started this whole wild goose
    Petulance? It was mostly (I am only human after all) aimed at the
    others, not least Brian himself, who I suspect knows more about optics
    than the rest of us added together (he is or was a lens designer). Your
    remarks seemed contemptuous. Ironic, really.
    You keep insisting we disagree with your logic. What we really disagree
    with is your vocabulary. You are having the wrong argument. I suspect
    you will go on doing so, but that won't make you right.
    David Littlewood, Aug 5, 2004
  8. PrincePete01

    mcgyverjones Guest

    I am not talking about movements here at all.
    A view camera with the lens and film planes parallel adheres to this
    definition of a conventional camera.
    No, distortion

    As has already been gone over here (why am I bothering) the outer edges of a
    wide angle shot are simply gathering light from a more extreme angle and
    projecting them on a flat surface.

    But this is not perspective, by the very definition of the word as it
    applies to photography.

    mcgyverjones, Aug 5, 2004
  9. Play a first person shooter sometime and put your face about half
    your monitor width away from the screen. See if you get a bizarre
    3-D-like effect since the projected field of view in the game now
    matches your eye's field of view of the screen itself. Do it long
    enough and you can get dizziness and motion sickness if you're
    susceptible to such things. ;)

    BillyJoeJimBob, Aug 6, 2004
  10. It is a method of mapping a sphere centered on the lens pupil to the
    plane, but it is also a method of mapping *any 3D subject at all, with
    no restriction* to the plane. You seem to want to talk about it only in
    relationship to the sphere, *but there is no sphere* in all but very
    unusual cases.

    The difference is that a sphere is an essentially 2D surface; two
    coordinates define the location on the surface. Most photographic
    subjects are fully 3D; you need 3 variables to describe where the ray
    starts from. The sphere->flat mapping is a *subset* of all possible
    3D->flat mappings that project through a point.

    Apple juice is a drink, but a drink is not apple juice (except in
    special cases).

    Dave Martindale, Aug 6, 2004
  11. Another example of this: visit an IMAX theatre or IMAX DOME (formerly
    OMNIMAX). In both of these theatres the field of view is very wide;
    90-100 degrees for middle-of-house seats in IMAX and nearly 180 degrees
    in the dome. In films made specifically for one format, the wide-angle
    shots use a camera lens whose field of view is roughly the same as that
    of the audience member sitting in the middle seat, so the perspective is
    quite natural (and certainly capable of causing motion sickness).

    The difference between the two systems is that true IMAX films use
    rectilinear camera and projector lenses, and a nearly flat screen. IMAX
    DOME uses fisheye lenses on camera and projector and a dome screen.

    (Unfortunately, these days most IMAX films are not shot for one theatre
    type exclusively; they're designed to show in both and are a

    Dave Martindale, Aug 6, 2004
  12. PrincePete01

    Jeremy Nixon Guest

    There is no *actual* sphere. What I'm talking about is the image; it's an
    imaginary sphere used to visualize the process of creating the image, and
    despite what you're saying -- none of which contradicts what I'm saying --
    that model is still working for me.

    A rectilinear projection is a method of mapping any 3D subject *as seen
    from a single point* onto a plane. The center point, the lens, can see
    in all directions, but can only see a single point in each direction,
    nothing past that point. In order to create the image, all of the points
    it can see, one per direction, are mapped onto a single, flat surface.
    Before the projection into flatness occurs, that surface is a sphere --
    one point in each direction, with distance removed to make it flat. The
    field of view is simply what section of the sphere the lens is "cutting
    out" to project onto the plane. The section it "cuts out" is not flat,
    it is a section of a sphere, and thus must undergo some projection in
    order to become flat, normally rectilinear projection.
    Jeremy Nixon, Aug 6, 2004
  13. The *image* is what is formed on the focal plane by the converging rays
    from the lens. "Image" is a very well-defined term in optics and
    photography. You are creating an imaginary sphere centered around the
    lens entrance pupil in object (subject) space and projecting the 3D
    world onto that sphere before reprojecting it through the lens onto the
    focal plane. That sphere *is not an image*, and calling it that will
    just confuse everyone.
    Well, in fact what you're describing is a pinhole, not a lens, because a
    real lens captures a ray bundle and has a limited depth of field as a
    But that's only in your mind. There is no sphere in reality, and no
    mapping onto it. If you must map the 3D world onto a 2D surface, why
    not just use a large flat plane in front of the camera? That will work
    for everything but fisheye lenses. Or map the world onto a square box
    with the lens entrance pupil at the center; that's computationally
    easier in some respects.

    Both the sphere and square box are used for things like reflection maps
    and environment maps in computer graphics, where there is a need to
    obtain a 2D projection of a 3D world. The choice between sphere and
    box is an essentially arbitrary one; just pick the one that's easiest
    to code or which provides the best performance.

    However, in optics and photography, there is no real need to project the
    world onto any intermediate surface. We can project the 3D world
    through the lens directly onto the film plane (yes, even with
    "distorted" lenses like a fisheye). So why add the extra step of
    mapping to a sphere? This loses actual depth information.
    Well, this may be a useful model to you, but it's an arbitrary one, with
    no physical reality. Most people don't seem to find it useful, and
    certainly not necessary. It's just silly to suggest that it's a
    fundamental part of the subject->image mapping.
    This is all true *if* you assume this intermediate spherical
    representation of the world. But there's no need to assume that.

    Dave Martindale, Aug 6, 2004
  14. Let's say that "artistic perspective" is inseparable from field of view.
    But "photographic perspective" or "optical perspective" is a different
    thing, and does not change with field of view if relative subject and
    camera (actually lens entrance pupil) remain the same. Everyone else is
    talking about photographic/optical perspective. Why not just recognize
    Yes, I've heard that, in the context of art. I've also heard it in the
    context of photography, where the author was referring to the fact that
    if you shoot with a wide angle lens and look at the print from too far
    away, so the apparent visual angle is varied, you get a particular sort
    of distortion. I just recognize those as two different meanings of
    perspective from the one used in optics.
    If you think that, you haven't been reading very carefully. I claim, as
    others have claimed, that there is no difference between a wide-angle
    and telephoto image *over the field of view covered by both lenses when
    that field of view is enlarged to the same size*. That's all that is
    meant by saying that the photographic perspective is the same, and it is

    You just insist on saying that perspective means something else, that a
    wide-angle shot looks different because it's wider. Well, that's
    obviously true, but the difference is not in the photographic
    perspective. The difference is in the FOV, and in the way the image
    looks when viewed from some fixed distance (rather than the correct
    distance for the FOV). You're correct about the details, except that
    this is not what perspective means in photography.
    It is not related to *how much angle of view* is contained by the
    These things happen when viewing a 3D scene with your eye; with no
    camera involved. When you record the scene with a camera, and view the
    print from the correct distance, it usually *faithfully* records this
    effect, so the image is not distorted. Both WA and tele lenses
    record *the same thing your eye saw* (ignoring DOF issues).

    If there's a difference in angles or convergence with distance between
    what your eye saw in the scene and what you see looking at a print, that
    means you're looking at the print from the wrong eyepoint or the wrong
    distance, which is certainly not the fault of the lens. You are
    introducing the distortion in that case; it's not the camera or the

    (The obvious exception to this is lenses with deliberate distortion like
    a fisheye).
    It's a lot easier to continue to use the existing commonly-accepted
    definition of perspective, than for everyone to change to suit your
    perspective on the matter. Most of us can deal with multiple different
    meanings of the word in different contexts without getting confused
    and without insisting that the world change to accomodate us.

    Dave Martindale, Aug 6, 2004
  15. PrincePete01

    Nostrobino Guest

    Sure he did, in his reply to you yesterday. You kept speaking of spherical
    projection and he agreed:
    It is a spherical projection. But there is no sphere involved

    You don't remember that?

    When I questioned this, he essentially retracted it:
    I'm being generous. It might be called a spherical projection in the
    sense that, if you represent the position of the object relative to the
    camera in spherical polar coordinates (two angles and a distance), then
    the position of the image point corresponding to the object is
    determined by the two angles (but not the distance).

    But you're right, it's not really a projection of a sphere onto a flat
    surface, and not a spherical projection in the cartography sense.


    Actually, no, I'm far from "the only person in the entire world who thinks"
    that. Anyone who can see what's in front of him (and whose reasoning power
    has not been destroyed by reading nonsense) can see that a photo taken with
    a wide-angle lens has an entirely different perspective from a photo taken
    with a long lens from the same position. Unfortunately, you appear to be one
    of those so afflicted; as I recall you insisted repeatedly that there is "no
    such thing as a telephoto look," etc.

    Go in peace then, brother, and enjoy the sphericity of your spherical world.
    Nostrobino, Aug 6, 2004
  16. PrincePete01

    Nostrobino Guest

    First you suggest it's composition and now it's distortion? What distortion?
    Of what do you think the distortion consists?

    The "more extreme angle" of objects around the periphery are what gives a
    wide-angle shot its characteristic perspective. Of what do you think
    PERSPECTIVE consists?

    Go ahead and define the word as you believe it applies to photography. Do
    you imagine that it has some entirely different meaning in photography than
    it does, say, in perspective drawing? Why would it?
    Nostrobino, Aug 6, 2004
  17. PrincePete01

    Nostrobino Guest

    No, I have not. I would cheerfully do so if logic, reason, common sense and
    good deportment required it, but they do not.

    "On the same camera" was of course presumed to be understood. Here's what I
    said on August 3:
    Perspective depends on three things: camera (or viewer) position, included
    angle of view (or focal length), and the direction the camera (or viewer's
    gaze) is pointed.

    In all of my discussion of focal lengths, when they were mentioned
    specifically (e.g., when I suggested comparing shots taken with 20mm and
    200mm lenses), it would have been obvious to nearly any reader that 35mm
    SLRs were presumed.

    Yes. Among other things, they provide the picture with its perspective.

    Obviously this is incorrect and even self-contradictory, isn't it?

    The statement itself speaks of "abnormal linear perspective."

    "abnormal linear" what?


    Once again, what is it that's "abnormal" and "linear"?


    And this results from using what kinds of lenses?

    "wide angle lenses and . . . telephoto lenses."

    And use of these lenses results in what, again?

    "abnormal linear perspective."

    Well, I'm certainly glad we got that straightened out at last.

    Now to go on from that enlightened observation and then say that "it is the
    camera position and not the focal length or type of lens that produces the
    abnormal linear perspective" is clearly contradictory. If this were true,
    then it would have to be possible to change THE CAMERA POSITION to one in
    which it DOES NOT "produce the abnormal linear perspective." But you cannot.
    A wide-angle lens will still produce the same "abnormal linear perspective"
    or "strong perspective" (as the article calls it) wherever you place the
    camera, provided there are objects in the field of view to demonstrate it;
    contrariwise, a telephoto lens will still produce its own kind of
    perspective regardless of camera position.

    Correct, they are not, and you need not apologize for it since those things
    have nothing whatever to do with perspective as we are using the term here.
    We are speaking only of geometric perspective.

    Your own citation from the Focal Encyclopedia of Photography says that it
    is, Dave. Your encyclopedia article says that "strong perspective is often
    associated with wide angle lenses and weak perspective is similarly
    associated with telephoto lenses, . . ." Clearly they are using
    "perspective" to mean PERSPECTIVE OF THE PICTURE AS A WHOLE, which is what
    I've said repeatedly is precisely the reason it's not valid to take some
    small part from the center of the picture and say that establishes the
    Nostrobino, Aug 6, 2004
  18. PrincePete01

    Jeremy Nixon Guest

    Of course.
    The reason I find it useful is that it models the distortion that happens
    from flattening the image. In fact, it models it rather perfectly. Yeah,
    I know we already talked about that word "distortion", but I don't know
    exactly what else to call it.
    Jeremy Nixon, Aug 6, 2004
  19. PrincePete01

    Nostrobino Guest

    I have been and still am using the word exactly as your article from the
    Focal Encyclopedia of Photography uses it, to mean perspective as it affects
    the entire picture. That is absolutely clear from the excerpt you
    posted--for which I am grateful, by the way.

    Again: "strong perspective," "weak perspective," "abnormal linear
    perspective"--these are exactly the terms used by the encyclopedia article
    you posted, and clearly they refer to the perspective of the entire picture.

    If your objection is to my use of "wide-angle perspective" (where your Focal
    Encyclopedia uses "strong perspective" or "abnormal linear perspective" in
    connection with wide-angle lenses), I have no great objection to replacing
    my term with either of those used by the encyclopedia. However, I do think
    that because this is a perspective associated with, and only with,
    wide-angle lenses, "wide-angle perspective" seems a more apt term.

    I can only again refer you to your own posted encyclopedia excerpt, which
    takes exactly the same meaning that I have.
    Nostrobino, Aug 6, 2004
  20. [snip]
    This is sheer sophistry. It doesn't impress.
    The "abnormal linear perspective" comes from viewing the print from a
    position inconsistent with the position (relative to the subject) from
    which the picture was taken. This is of course perfectly clear from the
    piece I quoted, at least to those who subscribe to the normal meaning of
    It's only "inconsistent" from the point of view (I almost said
    perspective) of someone who uses words differently from the great
    majority of the photographic community.
    Whew, at least we avoid another wild goose chase!
    Who's Dave?
    As a piece of inaccurate selective quotation, that should win a prize -
    you omitted the point that the piece then goes on to state that "it is
    camera position and not the focal length or type of lens that produces
    the abnormal linear perspective". This of course flatly refutes your
    principle argument.

    I suggest you read the piece cited yourself. If the overwhelming
    majority of those here, and an authoritative reference source, all
    disagree with you, can't you at least have some doubt.

    I guess not.

    You weren't one of O J Simpson's defence team were you?
    David Littlewood, Aug 6, 2004
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.