perspective w/ 35mm lenses?

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

  1. PrincePete01

    Nostrobino Guest

    Get straightened out on the flat plane vs. spherical business first, Jeremy.
    There is no point in continuing with this until you've got that
    misconception corrected.
     
    Nostrobino, Aug 4, 2004
    1. Advertisements

  2. PrincePete01

    Nostrobino Guest

    Well, those I would say are all the same plane, if you're talking about the
    film plane. You're just moving it around. I'll agree that in
    three-dimensional space, anything can be moved through three planes. :)

    That would not be the ordinary understanding of "camera position," I think.
    Most people would take the term to mean "where the camera is," nothing more.

    Not a real view camera, no. I did have a press camera about 50 years ago and
    played with that, but of course its movements were more limited. It was
    still fun, though.
     
    Nostrobino, Aug 4, 2004
    1. Advertisements

  3. PrincePete01

    Jeremy Nixon Guest

    No, they won't; nor will the stretching of a person's head in the corner of
    a wide-angle shot. That's what you call "distortion".
    And how is that three-dimensional world converted into a two-dimensional
    representation? You conveniently skipped that part, which is what I'm
    talking about. You need to use a projection to do that.
    I'm sorry -- where was it that you got the mistaken impression that I ever
    said anything about the "field of best focus"?

    Oh, I know what it was -- I mentioned "focusing" your camera to 10 feet.
    That was to indicate the distance to your imaginary subject. I'm not
    talking about focus at all, here. Okay, nevermind my other (short) reply
    where I still thought you were talking about the image plane being flat.

    The image as it comes into the camera is spherical. This is due to the
    fact that the lens is in one place, and the world is all around it. The
    spherical image is projected onto a flat image plane. This process
    distorts the image, no matter what you do. The lens uses (intentional)
    distortion to "correct" it in some way, using a specific projection to
    flatten the image.

    When you photograph a flat surface, with no perspective at all, the lens
    distortion serves, for all practical purposes, to cancel out the projection
    distortion, so it all looks right, edge to edge. With a three-dimensional
    object, this can't happen. It is going to be distorted in some way, because
    that is inherent to making it flat.
     
    Jeremy Nixon, Aug 4, 2004
  4. PrincePete01

    Jeremy Nixon Guest

    If you try taking some shots with a view camera you'll understand why (to
    us, at least) the only sensible definition of "camera position" is where
    the image plane is, not any other part of the physical camera. After all,
    it is quite possible (and indeed routine) to point the lens in a different
    direction without moving the image plane at all.
     
    Jeremy Nixon, Aug 4, 2004
  5. PrincePete01

    Nostrobino Guest

    There isn't any "now" about it, Dave. I have never said, hinted, suggested
    or implied that enlarging an image changes its perspective. Quite the
    contrary.

    What I have said is that enlarging the center of a wide-angle image to
    "prove" its perspective is the same as that of a long lens doesn't prove any
    such thing.

    Perspective (and again, I have said this many times) is a quality of the
    picture as a whole.

    This, by the way, is what's wrong with your earlier example of using many
    pictures to make one. You can do that, all right, and it's been done many
    times, but it never really duplicates the perspective of a wide-angle shot.


    The field of view is reduced and the image that remains is enlarged.

    Necessarily, yes. But not of the whole original image, of course.

    Ah, but it DOES change the perspective, as I have repeatedly point out.

    No shot you can take with a long lens, or any number of shots pasted
    together, can accurately reproduce the perspective of, say, a 17mm or 20mm
    lens. You can paste together a lot of shots that will give you the same
    angular coverage, but the perspective won't be right. Try it and see. This
    is really the proof that perspective is necessarily a quality of the picture
    IN ITS ENTIRETY.

    I've already suggested a method much easier than doing all that pasting.
    Just take an ultra-wide-angle shot, then enlarge a corner of it and see if
    you can duplicate that perspective with a much longer lens. I can tell you
    right now you won't be able to, at least not with a conventional camera. To
    establish perspective, both shots have to contain two or more solid objects
    in the same placement.


    Any time. Let me know how you make out with the paste-ups, and/or when
    you've given up and finally admitted I'm correct in this.
     
    Nostrobino, Aug 4, 2004
  6. PrincePete01

    brian Guest


    I know - I read the thread - but I just couldn't help myself!

    Brian
    www.caldwellphotographic.com
     
    brian, Aug 4, 2004
  7. PrincePete01

    Nostrobino Guest

    Yes, they will. It's a well-known fact, Jeremy.


    It all disappears when viewed from the proper position. This is well known.

    A "projection" can mean many different things. For example, the Mercator
    projection used in most world maps. That is definitely a distortion, a
    necessary one to get a sphere imaged on a flat surface. But a camera lens
    doesn't do anything like that.

    What do you see when you look out the window? Anything spherical? No. You
    see various solid objects arranged in three-dimensional space. Now imagine
    the image of those objects, just as you see them out the window, being
    transferred to film or a CCD, or some other imaging sensor. That's all a
    rectilinear lens does.

    Yes. You said:
    If you focus your lens to 10 feet, what are you photographing? It's not a
    flat plane 10 feet in front of the camera, parallel to the image plane. It
    is the inside surface of a sphere 10 feet from the lens in all directions,
    <<<

    And that is absolutely WRONG. It IS "a flat plane 10 feet in front of the
    camera," and definitely not "the inside of a sphere 10 feet from the lens in
    all directions."

    No, it is not. There's nothing spherical about it, if we're talking about
    rectilinear lenses.


    So what? It's still not the inside of a sphere.


    There is no distortion (other than within tolerances) in any rectilinear
    lens.

    ASK people, Jeremy. Don't just make this stuff up out of your own head.
    You're very badly confused about this.
     
    Nostrobino, Aug 4, 2004
  8. PrincePete01

    Nostrobino Guest

    Look out the window, Jeremy. See? No sphere.
     
    Nostrobino, Aug 4, 2004
  9. PrincePete01

    brian Guest

    Simply cropping a picture cannot move (much less REMOVE) vanishing
    points. Perhaps you don't understand the simple fact that vanishing
    points can lie *outside* the boundaries of an image. For example if I
    draw or photograph a building in such a way that the vertical lines
    are kept parallel, then the vertical vanishing point is *infinitely*
    far away.

    Field of view, NO! It has nothing to do with perspective.
    Does this mean that you agree with me that you can't make vanishing
    points move simply by cropping a picture?
    Absolutely YES. By all means, do the deed ASAP! Utterly stubborn
    stupidity has only limited entertainment value.

    Brian
    www.caldwellphotographic.com
     
    brian, Aug 4, 2004
  10. PrincePete01

    Jeremy Nixon Guest

    "Rectilinear" refers to the image it produces, not the one it sees.
    Please, someone, tell me if I'm wrong.
     
    Jeremy Nixon, Aug 4, 2004
  11. perspective.

    So you apparently missed the message from that image...

    SNIP
    Wrong, at least two things happened. When you (or the lens) look(s) at
    a disc (that's parallel to the film/sensor surface) at an angle, it
    will look like an ellipse from a distance. However, the rectilinear
    lens projection on a flat (film/sensor) surface distorts the ellipse
    back to a quasi circle. A lot has happened.
    No that's not his point, but an observation that the spheres, unlike
    the discs, still look like a sphere from an angle. The author's point
    is that the projection on a flat plane, stretches/distorts the
    film/sensor image. That understanding can be exploited if one
    manipulates the angle of incidence on the film/sensor.
    Not true, e.g. take two normal CDs or DVDs and hold them at different
    distances so that they don't cover each other, you'll be able and tell
    which one is closer than the other, aka perspective.
    The point you are missing is that subjects at the same distance from
    the lens, will be imaged with the same magnification. That applies to
    all subjects, effectively forming the inside of a sphere. Due to
    differences in subject distance the magnification/perspective changes,
    flat objects will not look flat (depending on orientation towards the
    line-of-vision), and due to the incident angle on the non-spherical
    film/sensor there will be projection distortion.

    Bart
     
    Bart van der Wolf, Aug 5, 2004
  12. Read the message thread again:

    In message <>,

    It sure sounds like someone on "our side" said exactly that, or
    perhaps you consider brian to be on "your side"?

    For that matter, read my first post to this thread again:

    In message <>,

    Hmmm... so much for "neither you nor others on your side".
    Since you seem so eager to lump me in with "you've all", please
    point out where I've said that camera position is all that matters.
    Others may have, but I have not.

    Nevertheless, it's not the camera position that matters, but rather
    the positions of each sensing point on the image plane relative to
    the scene being imaged. We can represent each sensing element
    ("sensels" in the digital case, "grain sites" in the case of film I
    suppose) by a set of three cartesian coordinates (X,Y,Z). If we
    change the direction the camera is pointing, the coordinates for
    these sensing elements change. It's far easier for folks to say
    "if you change the direction you're pointing the camera, you change
    perspective", but in reality all you're doing is changing the
    positions of the sensing elements relative to the scene being imaged.

    When you get right down to it, position actually does seem to be all
    that matters. Perhaps I should have said that at the beginning after
    all!
    The first and third of your three things are actually two parts of
    the same thing, namely position of the sensing elements relative to
    the scene being imaged.

    Let's move on to focal length, which has been and still is the only
    one that _I_ dispute.
    No, the difference in perspective is caused by a change in position
    of the sensing elements relative to the scene being imaged. The
    fact that using a different focal length lens caused you to have to
    recompose the shot does not mean that the focal length lens caused
    the change in perspective. It's the recomposition of the shot that
    changed the perspective.

    Stating that the lens is what "causes" the change in perspective above
    is incorrect. At that point you might as well say that you "caused"
    the change in perspective since you put the lens on the camera, or
    that the lens manufacturer "caused" the change in perspective since
    they made the lens in the first place. After all, if that lens
    didn't exist, you wouldn't have had to recompose the shot, thus
    changing the perspective, right?

    You could go all the way back to "primum mobile" at that point and
    blame the whole thing on God if it weren't for those darn bolts of
    lightning he tosses about from time to time. ;)
    Change the size of your sensor to encompass a larger portion of the
    image circle projected inside the camera. Does this make a normal
    lens turn into a wide angle lens? This, incidentally, is why "wide
    angle", "normal" and "telephoto" are basically meaningless terms in
    the grand scheme of things. For a medium format camera, a 50mm focal
    length might be "wide angle". It certainly is not "wide angle" for a
    35mm SLR format, nor is it "wide angle" for a DSLR with a 1.6x crop
    factor or greater.

    For the purposes of our discussion, I am going to interpret "wide
    angle" to mean "short focal length" and "long lens","telephoto", and
    the like to mean "long focal length".

    All you're seeing with what you call a "wide angle perspective" is an
    increased field of view. If you put a longer focal length lens on
    the camera (without moving the camera in any way), your "wide angle
    perspective" is still there; it's just been moved off of the
    surface of your sensor/film. Put a bigger sensor or larger format
    film in there, and voila, there's the bigger field of view, complete
    with your "wide angle perspective". The focal length of the lens has
    not changed.
    You're talking field of view at this point, not perspective.

    Be careful about not caring what I do, because you may just be
    surprised. Since your experiment was originally a thought
    experiment, I'm not limited by finances.

    So here we go...

    I can get an identical field of view at 55mm if I replace the sensor
    in the back of the camera with one that is right around three times
    as large (55/18 as large) in each dimension, assuming I don't get
    vignetting. Since vignetting is a result of lens design and
    manufacture and not necessarily of focal length, we will assume my
    lenses are vignette free for this gedankenexperiment.

    I now have the same field of view at 55mm with the 3x sensor as I did
    at 18mm with the original sensor. The CD will be in the bottom corner
    of the image in both cases, with the 55mm version being 3x as large in
    all dimensions but retaining the same aspect ratio as in the 18mm
    version. If the sensor elements on the 3x sensor are also 3x as
    large so the total number of "megapixels" is the same in both cases,
    the resulting 55mm version of the image will darn near exactly
    overlay on top of the 18mm version when I view them in Photoshop
    without my having to scale either one.

    Wow, that was easy. :)
    Increasing the size of one's sensor/film while keeping the same
    focal length seems to work quite nicely... and it's "field of view"
    that you're talking about, not "perspective". The two terms are not
    interchangeable, though they often appear near each other in text.
    Now you're comparing two objects, located at completely different
    positions relative to the sensing elements in the image plane. This
    falls under "different positions affect perspective" not "focal length
    affects perspective".
    No, the difference in how the two objects appear is the result of
    their respective positions relative to the sensor elements in the
    image plane. It's the equivalent of moving the camera, except
    CDs weigh less and don't normally require a tripod. ;)

    BJJB
     
    BillyJoeJimBob, Aug 5, 2004
  13. You're wrong.

    The world is not spherical; it's a 3D environment. There is no
    spherical image in front of the lens; there's no image at all there,
    just the 3D subject. The process of mapping that 3D environment to a
    2D image surface is one of projection, but projection does not
    necessarily imply distortion.

    For lenses, we *define* a distortion-free lens as one that has the same
    imaging geometry as a pinhole. Straight lines remain straight, but
    they may converge unless the subject is flat and the image plane is
    parallel to the subject plane.

    Your definition of "distortion" is useless because all lenses and even a
    pinhole distort images according to you. If all lenses are classified
    as distorted, then this provides no information about the lens.

    Dave
     
    Dave Martindale, Aug 5, 2004
  14. PrincePete01

    DSphotog Guest

    My conclusion is that what you really are is a very obtuse TROLL. All be it,
    a fairly well educated troll with a decent command of the language. And a
    really successful one at that.

    I do find it interesting that you will tell others that something is "It's a
    well-known fact, Jeremy" or "This is well known", but have so much trouble
    when anyone tells you the same thing.

    I'm done playing. Have a nice life.

    Best Regards,
    Dave
     
    DSphotog, Aug 5, 2004
  15. PrincePete01

    Nostrobino Guest

    No, trolls don't do what I've done.

    Are you implying that what I told Jeremy is a well-known fact is NOT a
    well-known fact?


    Aye, there's the rub. There are well-known facts, and then there are bits of
    nonsense accepted by many people as fact, which latter are still nonsense no
    matter how many people believe them. For example, there are probably even
    more people who believe all the Bermuda Triangle silliness than there are
    who believe that "perspective is determined by camera position alone" and
    "focal length has nothing to do with it."

    The man who can think for himself can often sort these things out and tell
    the difference between them. The man who cannot, will go on thinking
    whatever he thinks most other people think.

    I'm disappointed you're not going to even try to show me how the perspective
    of a wide-angle lens can be duplicated with a long lens, which (if as you
    claim perspective is determined by camera position alone) should be easy for
    you. I have not the foggiest idea myself how you would do it, and was
    looking forward to this instructive explanation.

    But I understand. As the saying goes, "You can talk the talk, but can you
    walk the walk?"

    Best regards to you.
     
    Nostrobino, Aug 5, 2004
  16. PrincePete01

    DSphotog Guest

    Care to post your home address?
     
    DSphotog, Aug 5, 2004
  17. PrincePete01

    Guest Guest

    it has been explained several times in this thread.
     
    Guest, Aug 5, 2004
  18. PrincePete01

    Nostrobino Guest

    Then I stand corrected, and thank you for the correction. My only excuse is
    the sheer number of posts in this topic, the number of posters, and the fact
    that in general you all seem to have been taking the same position. I
    recognize that there have been differences between you, but they tend to all
    run together after a while.


    Well, in a sense everything related to this has to do with position. The
    position of the camera, the direction it is pointing, and the focal length
    ( = field of view) all have to do with the positions of things.

    But of course that isn't absolutely correct either. If, for example, one of
    those right-angle first-surface mirror devices were mounted on the lens, the
    film plane (or corresponding sensor) would be rotated 90 degrees
    horizontally relative to the field of view. Sure, this is a quibble and
    perhaps an unfair one, but after all we are getting pretty fussy about
    details here.

    But that is only enabled by the focal-length change.

    The point has been and remains that you CANNOT duplicate the perspective of
    a wide-angle lens with a long lens.

    Your objection then is with the verb "cause"?

    I don't think I used it. I said "perspective is determined by," not
    "caused." I don't think I'd have used "cause" in that way.

    Not very much, you couldn't, except of course with a dSLR using 35mm SLR
    lenses. That may be what you have in mind, I understand that, but still you
    could only increase the sensor size to an effective diagonal of 43 mm or so.
    Lens makers (for 35mm cameras) don't give you much extra in the way of
    coverage.

    Well, "telephoto" is actually very specific in its meaning, though it is now
    commonly misused to mean any long-focus lens.

    In fact a 50mm lens is an extreme wide-angle on a 6x6cm camera. I used
    precisely this example myself, somewhere way back in this thread.

    Then it's not "still there."

    Not really, no. Put a "larger format in there" and what you get is a
    circular image on it that doesn't fill the format, unless of course the lens
    you're using was actually designed for the larger format.

    These things are all necessarily relative. With most digital cameras today,
    there are so many sensor sizes that actual focal lengths are essentially
    meaningless and are generally not even mentioned. If e.g. someone says his
    Minolta S414 has a 35-140mm lens, what he really means is that that's what
    it is in 35mm camera equivalence. I have one and I don't have know offhand
    what its actual focal length range is. I'd have to look. Similarly, I know
    my X-series Minoltas have "37-111mm" lenses and my F300s have "38-114mm"
    lenses. Again, what the actual focal length ranges are I don't remember and
    don't really have any reason to remember. My 7i and 7Hi cameras have
    identical lenses, and in this case I do happen to remember that the lens is
    7.2-50.8mm, but the important thing is that its 35mm equivalence is 28-200mm
    (and the zoom control ring is so marked, actually). The same lens exactly on
    my Dimage 5 is a "35-250mm" lens because of that camera's slightly smaller
    CCD.

    All this makes sense with digital cameras since most people are familiar
    with 35mm camera focal lengths, and would be hopelessly confused by true
    f.l. numbers if those were all they were given.

    However, all this has no direct bearing on wide-angle perspective, which is
    really a function of field of view. It's convenient to use focal lengths
    instead of angular measurement, simply because f.l. is the measure that
    people are most familiar with. More people will understand what you mean by
    "28mm lens" than would if you said "75-degree lens," and so on.

    They are largely inseparable.

    Well, okay. We both know that wouldn't work in the real world, but I'll go
    along with you.

    Or close enough. Yes.


    Fair enough.

    Uh . . . what was easy? Did I miss something?

    Both. They are largely inseparable (at least insofar as three-dimensionality
    is concerned), but when I'm speaking of perspective it's perspective I mean,
    not field of view.

    I agree, they are not interchangeable. Field of view means the included
    angle "seen" by the lens and film/sensor, nothing more. It has definite
    implications for perspective but is not necessarily a determinant of
    perspective; e.g., it is possible to take a photograph that has no
    perspective whatever, but it still has a field of view.

    Perspective is the visible relationship between solid objects within the
    field of view, i.e. the apparent size, shapes, angles, convergences etc. of
    those objects. If there is no three-dimensionality (e.g. a photo of blank
    sky), there can be no perspective, though there is still a field of view.

    We are agreed on that I think, yes?

    The different positions that are VISIBLE only become so because of the
    change of focal length.

    Now if you say "the difference in perspective was there all along, you just
    couldn't see it," I agree with you. But we are, after all, speaking of
    photographs. It is only the content of the photograph that matters as long
    as we are speaking of the perspective of that photograph.

    <chuckle>

    Well, BillyJoeJimBob, I will say this for you in all sincerity: You appear
    to be the only one here who has a mind worth engaging, and I very much
    appreciate it. :)

    N.
     
    Nostrobino, Aug 5, 2004
  19. PrincePete01

    Nostrobino Guest

    Well, I wouldn't call that distortion. And it's a real circle, not a quasi
    circle.

    I agree that if you look at a disc from an angle it will appear as an
    ellipse. It will do so with a camera lens also, except in the special case
    (such as in that figure) where the lens axis is perpendicular to the
    subject. As long as that is the case, the lens is rectilinear, and the film
    plane is also perpendicular to the lens axis, a disc will necessarily be
    represented on the film as a circle.

    But geometry REQUIRES that the circle in the subject be represented as a
    circle on the film, as long as those conditions are met (lens axis
    perpendicular to the subject). There isn't any optical distortion involved;
    quite the contrary.


    I agree. (It's understood that they're on the same plane and not at eye
    level.)


    I still don't know where your "sphere" is coming from. In the example shown,
    all the discs are as you say "imaged with the same magnification." They're
    on a flat surface. There isn't any sphere.

    In the case of that figure, the differences in distance from disc to lens
    are exactly canceled out by relatively equal differences from lens to image
    plane. There isn't any perspective at all as far as the discs are concerned
    because they're (effectively) not solid objects, they're two-dimensional.
    The ping-pong balls show perspective because they ARE solid objects.

    Sorry, you've lost me there again. I just don't understand why you think
    there's anything spherical involved.

    N.
     
    Nostrobino, Aug 5, 2004
  20. PrincePete01

    Nostrobino Guest

    You and Bart, I should have said. He has some interesting ideas also, though
    I have great difficulty following them.
     
    Nostrobino, Aug 5, 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.