Talk:Perspective correction

Untitled

There are quite a number of things that can be done in photography to correct for image distortion. Would this be the place to discuss them? Or would it be better to find a photography page and link to this? I don't know how to set up a link.--Jocsboss 03:41, 31 Dec 2004 (UTC)

To some extent it should be discussed here, I believe. To add a link you just type [http://www.link.com Description]. Paranoid 08:19, 31 Dec 2004 (UTC)

I've always felt that perspective correction was just another type of distortion. If you see a low-angle shot of a building with parallel walls, it may look more aesthetically pleasing than if the walls were converging, but in no sense is it "correct" since it's a view that cannot be seen in reality. Lee M 16:14, 2 Jan 2005 (UTC)

First, some distortion is caused by the lenses, not just the position itself. Second, you are right to some extent, but the image that we see in reality is processed by the visual cortext extensively anyway. We don't see reality, we see what we believe it should look like. Two parallel vertical lines that we see may not be really parallel, but our eye usually perceieve them as parallel (even though in the "raw data" that our eye gets they are not even straigt!). So it makes sense from the aesthetical and psychological points of view to modify the images so that straight vertical parallel lines are represented by vertical parallel lines on the photo (unless it's a shot of a skyscraper seen from the ground-level position right next to it, where "incorrect" perspective is intentional).

To summarise: the only way to represent visual data 100% correctly is to create a hologram. All other ways are shortcuts. The traditional perspective is the most efficient one in most cases, so unless there is compelling artistic reason, the images should be perspective corrected. Paranoid 16:48, 2 Jan 2005 (UTC)

arbitrary illogic. the article is mistitled and inaccurate, pulling opinion out of the sky because it suits the preference (taste) of the contributor(s). the defense above admits that there is no possible correction in conventional image distribution, yet clings to the false term perspective "correction". this is an encyclopedia for truth, not mislabeled bias (regardless of the scope of its acceptance). if the article title is to remain, it should be clearly identified as a misnomer in the article text. ideally, the article title (Perspective correction) would become a redirect to "Perspective modification" or the like. SaltyPig 07:24, 23 July 2005 (UTC)

I've removed "Assuming that the view direction of the photographer was parallel to the land surface when the photo was made, the" As if the view direction was parallel to the ground there would be no distortion (assuming taking a picture of a tall building), but instead a whole lot of foreground 154.5.47.15

The term "perspective correction" is actually the proper term. If I, as a photographer, wish to take a photograph of a tall building and want it so that the vertical lines are indeed vertical, then it would require that the film plane (or image sensor) of my camera be parallel to the face of the building as well as centered vertically and horizontally. This is not always possible. Therefore, my image would be, in a sense, a failure that would require correcting; either in a darkroom or computer. Also, as a side note, this type of work has always been called "perspective correction" and, while I realize that this is not justification in and of itself, it certainly lends weight to the argument that the article should remain titled as it is. Jocsboss 21:30, 14 August 2007 (UTC)

I am surprised that the term "restitution" is not in use. From my point of view (Sweden and Swedish terms) the perspective is the picture you get from the point where you are and the photo is taken. When the picture is taken you can not change the perspective. But you can, if you want, restituate vertical lines as vertikal (as we mostly do in architectural images). This can be performed by shift-lenses or by restitutional modifications when enlarging the image (mostly of course nowadays made in digital environment). If you take pictures from the same point, with wide angel lens or tele photo lense, the perspective is still the same. Maybee you can call this article "False correction of perspective"? Xauxa (talk) 18:32, 1 March 2008 (UTC)

I disagee that what the camera sees is distorted. The perspective correction is distortion. As M Lee said, the corrected image is "a view than can never be seen in reality." The "corrected" image would not match what you see elevated to the midline of the building, or what you would see from a telephoto lens. The correction does not correct the image: it just removes the most obvious effect of perspective. Everything else in the image still exhibits perspective. Plus, my brain straightens the unadulterated photo's lines when I view them. Clarkcol (talk) 18:16, 14 March 2008 (UTC)

From the article: "The popularity of amateur photography has made distorted photos made with cheap cameras so familiar that many people do not immediately realise the distortion." You might just as well say that the popularity of home cooking has made people unable to appreciate MSG! Perspective correction is a photo editing technique, and should be described as such. I don't think the issue should be the title, especially if our photography expert Jocsboss claims it is the accepted term. Wikipedia doesn't need any more ridiculously unnecessary neologisms. What is more troublesome to me is pretentious hogwash such as the above quotation. To respond individually,

• Lee M and Clarkcol, I totally agree with both of you. Perspective correction, like all photo editing techniques, alters the image to produce something less natural, however favorable.
• Paranoid, there are serious logical flaws in your argument, as noted by SaltyPig. It is true that the brain processes and interprets information captured by the eye. However, this does not necessitate perspective correction, because the brain processes photographs the same way. Also, the idea that almost all photographs should be perspective-corrected is unrealistic and subject to opinion. Very few photographs are modified in this way, and not because of those lowly amateurs. I find the example used in the article ironically objectionable: the cathedrals of that period were designed to tower into the heavens from a ground-level point of view. Why should photographs present them any differently than they have been seen for hundreds of years?
• Jocsboss, your statement about tall buildings is incorrect, and your view of unedited photographs ("a failure that would require correcting") is an artistic anomaly. Even if you were facing the ten-story building squarely, standing on the roof of the five-story building a few blocks away, optically natural perspective would still exist. If you were to see through the building, this would be evident (Fig. A). You would not have a "perfect" image (B). If you walked a little to the right, you would see sloped horizontal lines (C), not "correct" ones (D).

  _______     ._______.     _______        _______._
 |\_____/|    |       |    |       |\     |       | |
 ||     ||    |       |    |       | |    |       | |
 ||     ||    |       |    |       | |    |       | |
 ||     ||    |       |    |       | |    |       | |
 ||     ||    |       |    |       | |    |       | |
 ||     ||    |       |    |       | |    |       | |
 ||     ||    |       |    |       | |    |       | |
 ||     ||    |       |    |       | |    |       | |
 || ___ ||    |       |    |       | |    |       | |
 |/_____\|    |_______|    |_______|/     |_______|_|

  Fig. A        Fig B        Fig. C         Fig. D

MagnesianPhoenix (talk) 06:19, 28 June 2008 (UTC)