Perspective Revisited

[Ray]

Yes, Ray, we know that you prefer the Humpty-Dumpty concept to the Aristotelian one. I think Wayne and Guillermo have expressed (in images) my own feeling at this point very well.

My one remaining suggestion is this: Why don't you coin a new term for your version of "perspective" which differs so greatly from the accepted meaning. How about "Rayperspective" for your version and simple "perspective" for the rest of us.

Eric

Why isn't the burden on you [Ray] to make clear what you are referring to when you talk of perspective? Why is it that you [Ray] own the definition of that word?

Eric and Isaac,
I have no need to coin a new term for perspective. It's already been defined in the Oxford English Dictionary. That's the definition that I've been using all along.

It seems that in this thread, I'm the only one who has provided a definition to make it clear what what I'm referring to. It's no wonder you guys are banging your head against the keyboard. You're without a comprehensive definition. 

To help alleviate your frustration, I'll provide the definition again, from the most authoritative dictionary of the English language. It even provides Americanisms.

PERSPECTIVE

“The art of drawing solid objects on a plane surface so as to give the same impression of relative position, size, or distance, as the actual objects do when viewed from a particular point.”

On the way to the airport yesterday, I asked the taxi driver to stop at a temple we were passing, so I could take a few quick shots.

As I raised the D700 to my eye, with lens set at 14mm, I got a shock. "Oh my Gawd!, I exclaimed. Some of those buildings look so far away. I haven't got the time to walk over and have a closer look. I might miss my flight."

However, when I lowered the camera, I could see with my naked eyes (although still wearing clothes, of course) that some of the temple buildings were not nearly as distant as they appeared through the 14mm lens, so I relaxed and my heart beat slowed down. I wasn't going to miss my flight.

Now the Oxford dictionary refers specifically to drawing in its definition. To include photography and the concept of focal length, we could rephrase the definition as follows, without changing the underlying principle.

The art of depicting and presenting solid objects on a plane surface so as to give the same impression of relative position. size, or distance, as the actual objects do when viewed from a particular point with the naked eye or when viewed through a standard camera lens of focal length approximately equal to the diagonal of the sensor.

In the image below I didn't use an actual 50mm lens. I used a 50mm equivalent, which provided the same impression of distance that I actually saw with my naked eyes.

[Eric Myrvaagnes]

What frustration? If quarrelling about definitions makes your day then go right ahead.

+10!

[theguywitha645d]

To help alleviate your frustration, I'll provide the definition again, from the most authoritative dictionary of the English language. It even provides Americanisms.

OK, I lied, I will make another post. Here is another definition:

From The focal Encyclopedia of Photography:

PERSPECTIVE Persepctive refers to the appearance of depth when a three-dimensional object or scene is represented in a two-dimensional image, such as a photograph, or when the subject is viewed directly.

The difference you are seeing in the examples you are posting is simply the difference in the "correct" viewing distance--the correct viewing distance is different for each image, but you are using the same viewing distance for both and so the perspective is different. Viewing distance is simply changing apparent perspective. While you are seeing a relationship between focal length and perspective, it is an indirect relationship. Viewing distance actually gives the final perspective because you can actually get the same perspective at different distances and with different focal lengths, so those are not determining how we finally view an image--viewing distance also explains why we perceive the crop different from the whole because the whole will have the same perspective as the crop if magnified in relation to the crop. See my post above.

[Ray]

The difference you are seeing in the examples you are posting is simply the difference in the "correct" viewing distance--the correct viewing distance is different for each image, but you are using the same viewing distance for both and so the perspective is different. Viewing distance is simply changing apparent perspective. While you are seeing a relationship between focal length and perspective, it is an indirect relationship. Viewing distance actually gives the final perspective because you can actually get the same perspective at different distances and with different focal lengths, so those are not determining how we finally view an image--viewing distance also explains why we perceive the crop different from the whole because the whole will have the same perspective as the crop if magnified in relation to the crop. See my post above.

Not at all. My approach is rational and scientific. I fully realise the importance of the 'correct' viewing distance and make sure I've got it.

I'm surprised that certain individuals who have commented in this thread, and who appear to have a fairly sound scientific background, are experiencing such difficulty with this concept of perspective.

Surely everyone with at least a basic knowledge of scientific procedure and methodology should understand that in order to determine the effect of any change in a particular parameter, one should at least try to keep everything else the same. The only change, if possible, should be in the parameter under examination, in this case focal length of lens.

If we wish to examine what a change in position has on perspective, then we should change only the shooting position, either up or down, from one side to another, or forwards or backwards, and keep everything else the same, if possible, in order not to confuse the results with other influences, whether such influences be either direct or indirect.

We should use the same camera and lens, and view the resulting photographic images at the same size and from the same distance. We can then appreciate the precise effect that changing the shooting position has on the final result.

Likewise, if we wish to examine whether a change in focal length has any effect on perspective, we should adopt the same procedure of eliminating all variables other than lens focal length, as far as possible. We should use the same camera, but use different focal lengths of lens. We should shoot from the same position and angle, make prints of the same size and view them from the same distance. That's just plain common sense. It's not rocket science.

Having done this, it's clear to me that the choice of focal length of lens affects the sense of perspective in the resulting image, whether on print or screen. How anyone could deny that, beats me, unless he is on hallucinatory drugs.

However, I would agree there is another approach to keeping everything else the same as far as possible. Instead of keeping the print size the same we could keep any object that is common to all images under comparison, the same size. We would then have to scale the print size in inverse proportion to the focal length. In other words, the largest print would be from the shortest focal length, and the smallest print from the shot using the longest focal length.

If the largest print that one decides is practicable is, for example, 24"x36" using a 12.5mm lens (35mm format), then the print size from a shot with a 25mm lens will be 12"x18", a 50mm shot 6"x9", a 100mm shot 3"x4.5", a 200mm shot 1.5"x2.25", and a 400mm shot 0.75"x 1.125" which is the size of a rather small postage stamp.

Now if this is your method of working as a photographer, to scale print size in inverse proportion to focal length, then you are absolutely right to claim that focal length has no bearing on perspective. If we were to examine, for example, the stationary insect on a leaf which fills the small postage-stamp print, using a magnifying glass, we might notice the compression distortion that is  typical of a telephoto lens. When examining the same leaf in the centre of the 12.5mm shot, assuming we are able to find it, we might get the impression that the insect exhibits the same qualities of compression distortion. However, you'd probably need a D800, or better still an IQ180 to carry out such tests, and even then the insect might be so blurred it could be impossible to be sure if the perspective were the same.

But let's not get side-tracked by such practical difficulties. In principle, if the resolution is sufficient to compare the same size objects in different size prints, from the same close-up viewing distance, which is necessary to see small objects, then perspective should be the same, having previously eliminated lens distortions.

However, I have to admit that you guys are the only photographers I've come across who scale print size in inverse proportion to the focal length of lens used, in order to conform with the sacred rule that focal length has no bearing on perspective. 

[fredjeang]

Guys, we all have been fooled.

Ray is a master indeed and shouldn't be underestimated.

I suspect that he knows perfectly what we all were trying to explain. But the thing is that he is playing with us, has the time to because he is retiree, travel and has fun when he reads our posts.

The C.I.A or the Mossad should hire him for certain secret provocative tasks.

He may put all of us on nerves, but we are in fact manipulated.

[Bart_van_der_Wolf]

The difference you are seeing in the examples you are posting is simply the difference in the "correct" viewing distance--the correct viewing distance is different for each image, but you are using the same viewing distance for both and so the perspective is different. Viewing distance is simply changing apparent perspective. While you are seeing a relationship between focal length and perspective, it is an indirect relationship. Viewing distance actually gives the final perspective because you can actually get the same perspective at different distances and with different focal lengths, so those are not determining how we finally view an image--viewing distance also explains why we perceive the crop different from the whole because the whole will have the same perspective as the crop if magnified in relation to the crop. See my post above.

Exactly, couldn't agree more. I've already explained it in another thread with a reference to anamorphic projection distortion, or anamorphosis.

I've also outlined all elements that play a role from a photographic perspective (pun intended).

I don't feel an urge to repeat myself ad nauseam, hence the links for those who are interested and have more time to waste while trying to re-invent the obvious.

Cheers,
Bart

[theguywitha645d]

Having done this, it's clear to me that the choice of focal length of lens affects the sense of perspective in the resulting image, whether on print or screen. How anyone could deny that, beats me, unless he is on hallucinatory drugs.

And you are right, but it is not to do directly with focal length, but the viewing distance--I have made the same error and did research to figure it out. If I take a photograph from the same point with multiple formats and the focal length of each lens is proportional to the format diagonal (or the angle of view is the same), then the resulting perspective is the same. So is it the focal length? The angle of view? Both? Neither?

Viewing distance answers the question simply and easily. In order for the perspective to appear normal, not only is the taking position important, but also the viewing position. This has been well recognized in perspective drawing as shown by the invention of the viewing pinhole and also by creating perspective drawing in reference to where the viewer would stand. But it is also known that perspective is very robust in the fact we can view an image from another position and the perspective does not look unnatural--the reason the viewing pinhole did not last long.

So while you are correct in stating the focal length choice will affect the apparent perspective in the image, you are incorrect to assign it to the focal length per se as the perspective will always be the same if images are viewed by the relative position equal to the taking position. The reason we see focal length choice is we view images from a standard distance, which is relative to the image size, rather than to the taking position. Once you compare images at a standard viewing distance, the change in perspective created by focal length/angle of view becomes apparent.

[Ray]

And you are right, but it is not to do directly with focal length, but the viewing distance--I have made the same error and did research to figure it out. If I take a photograph from the same point with multiple formats and the focal length of each lens is proportional to the format diagonal (or the angle of view is the same), then the resulting perspective is the same. So is it the focal length? The angle of view? Both? Neither?

A focal length which is equal to the diagonal of the sensor is called a 'standard' lens, whatever the format and whatever the actual focal length that is marked on the lens. It produces an image in the viewfinder of a DSLR which depicts very closely the size, and impression of relative distances of the objects viewed, as the naked eye sees them.

I've mentioned a few times it's not the focal length per se that counts regarding changes to perspective but the equivalent focal length in relation to sensor size. It's also helpful to have a standard reference point such as '35mm format equivalent'.

So while you are correct in stating the focal length choice will affect the apparent perspective in the image, you are incorrect to assign it to the focal length per se as the perspective will always be the same if images are viewed by the relative position equal to the taking position. The reason we see focal length choice is we view images from a standard distance, which is relative to the image size, rather than to the taking position. Once you compare images at a standard viewing distance, the change in perspective created by focal length/angle of view becomes apparent.

That sounds very confused to me. Can't make head nor tail of it. It seems to me the correct viewing distance from any print is the distance which enables one to take in the entire composition within a single glance, and yet see all the essential details in the composition, such as very small objects and distant buildings etc.

Such viewing distances tend to be somewhere between 1 & 2 times the diagonal of the print. If the print is large and high resolution it may be necessary to go closer if one wishes to inspect fine texture.

Can you give me some specific examples of correct viewing distances and print sizes for the following two examples. (1) I photograph an entire tree containing a motionless lizard in the centre, using a 24mm lens (35mm format). (2) From the same position I photograph just the lizard and parts of a few surrounding branches, using a 400mm lens (35mm format).

What are the correct viewing distances in relation to print size for these two examples, in your opinion?

[theguywitha645d]

What are the correct viewing distances in relation to print size for these two examples, in your opinion?

Standard viewing distance is defined as equal to the diagonal of the print/display area.

Correct viewing distance is defined as proportional to the image magnification and focal length.

With an image taken with a normal lens, the standard viewing distance and correct viewing distance are the same. The correct viewing distance would be closer for images from wide angle lenses and longer for telephotos compared to the standard viewing distance.

Since you placed two images at the same size next to each other, you have fixed the viewing distance--I will not change my position when looking at one or the other. Because that distance is held constant, you can see the impact the change in focal length or crop has on perspective between the images. But the reason for that is the correct viewing distances for each image is very different. If you scaled one image in relation to the other so that the correct viewing distance would equal the distance I am viewing the images, then the perspective would be the same, but the images would not be the same size.

And to anticipate your question, as you get closer and further away from an image, the apparent perspective changes.

BTW, this is not my opinion. If you look at my first post above, you will see a few citations as well as a link. This problem was solved a long time ago and there is nothing to suggest perspective works otherwise.

[Ray]

Since you placed two images at the same size next to each other, you have fixed the viewing distance--I will not change my position when looking at one or the other. Because that distance is held constant, you can see the impact the change in focal length or crop has on perspective between the images. But the reason for that is the correct viewing distances for each image is very different. If you scaled one image in relation to the other so that the correct viewing distance would equal the distance I am viewing the images, then the perspective would be the same, but the images would not be the same size.

As in my examples in reply #111?, where I wrote:

" If the largest print that one decides is practicable is, for example, 24"x36" using a 12.5mm lens (35mm format), then the print size from a shot with a 25mm lens will be 12"x18", a 50mm shot 6"x9", a 100mm shot 3"x4.5", a 200mm shot 1.5"x2.25", and a 400mm shot 0.75"x 1.125" which is the size of a rather small postage stamp.

Now if this is your method of working as a photographer, to scale print size in inverse proportion to focal length, then you are absolutely right to claim that focal length has no bearing on perspective."

The above comment was made to highlight the absurdity of claiming that focal length, or FoV if you prefer, has no bearing on perspective.

Since it is not normal practice for photographers to scale prints in inverse proportion to the 'effective' focal length of lens used in order to maintain correct perspective; and since it is not normal practice for viewers to increase viewing distance in proportion to the 'effective' focal length of lens used in circumstances where the photographer has deliberately made a largish print of a telephoto shot, as most of them do; and since it often may not be possible to get sufficiently far away from, say, an A2 size print of a telephoto shot because the room is not large enough; and since most people when viewing a photograph wish to see the detail in the photograph, which is the whole purpose of using a telephoto lens in the first instance, I think it is reasonable to conclude that, in practice, changes in 'effective' focal length of lens used does have an impact on the perspective of the resulting image as experienced by a viewer acting sensibly and normally.

BTW, this is not my opinion. If you look at my first post above, you will see a few citations as well as a link. This problem was solved a long time ago and there is nothing to suggest perspective works otherwise.

Indeed! The science of linear perspective was addressed long before the camera and the modern lens was invented. But it's interesting that Leonardo da Vinci anticipated the conflict between the principles of linear perspective and the effect of wide-angle lenses.

I've never argued that changing position does not alter perspective. I'm arguing that the change in the Field of View which results from a change in the 'effective' focal length of lens used, also has a bearing on perspective, in practice.

And to anticipate your question, as you get closer and further away from an image, the apparent perspective changes.

But not necessarily with regard to photographing that image from different distances with different lenses. I can photograph the picture from any distance, using the appropriate lens to capture the whole picture and maintain the same perspective of objects depicted within the image, provided I am viewing the picture 'head on' and not at an oblique angle. The change in perspective experienced by the viewer, as he moves closer to, or further from the picture, is a change in perspective of the picture as an object in relation to its surroundings. The perspective of the objects depicted within the picture is fixed, whatever the viewing distance, and has been determined by the shooting position of the photographer and his choice of lens.

[theguywitha645d]

Ray, the two factors that impact (linear) perspective are the object distance (true perspective) when the photograph is taken and viewing distance (apparent perspective). It is the viewing distance that causes a viewer to see the change in perspective. Apparent perspective can only be seen when viewed. That has nothing to do with the focal length directly--it is a difference between the correct viewing distance which is related to the taking condition and the actual viewing distance. That is the whole story.

If you want to say the choice in focal length changes the apparent perspective in the image, that is fine. It has been common to say that wide-angles give wrong perspective and telephotos compress space. It is a nice rule of thumb. But focal length is not determining perspective--you cannot say 150mm focal length has weak perspective as you would also have to know the format size, on 8x10 it would be wide angle. So are you going to claim angle of view? Again it would also be true only if you know the viewing condition before hand.

So the underlying experience of apparent perspective is really the ratio between the correct viewing distance and the actual viewing distance. Because it is more usual to view something close to the standard viewing distance, it is easy to predict how focal length/angle of view will impact an image generally. But it does not explain that the thumbnail image that seemed to have strong perspective looks rather weak as a 5 foot print on my living room wall.

Anyway, that is my two cents. I really don't have anything to add.

[Eric Myrvaagnes]

I really don't have anything to add.

Nor do I.
Nor does Ray.   

[Rob C]

Nor do I.
Nor does Ray.   

Hello Eric,

It was snowing here in Mallorca when I drew open the curtains this morning; it became a little more pleasant after lunch, so I took myself off to have a coffee and then I decided to invest in a walk. That was a mistake: a threesome asked me to take their picture on a cellphone (belonging, I assumed, to one of them) and that instantly (and obviously) told me that they'd been admiring my own cellpix, and so I said of course, step this way, let's get a shot against that black curtain over the mountains surrounding the Bay. The mistake was in the time that took: before I could make it back to the car I was hit by all the hail that had been lurking within that curtain. Next time, I'll copy Mr Clooney and tell them they are mistaken, that in fact, I'm only me. Oh to be normal!

Rob C

[Eric Myrvaagnes]

Hello Eric,

It was snowing here in Mallorca when I drew open the curtains this morning; it became a little more pleasant after lunch, so I took myself off to have a coffee and then I decided to invest in a walk. That was a mistake: a threesome asked me to take their picture on a cellphone (belonging, I assumed, to one of them) and that instantly (and obviously) told me that they'd been admiring my own cellpix, and so I said of course, step this way, let's get a shot against that black curtain over the mountains surrounding the Bay. The mistake was in the time that took: before I could make it back to the car I was hit by all the hail that had been lurking within that curtain. Next time, I'll copy Mr Clooney and tell them they are mistaken, that in fact, I'm only me. Oh to be normal!

Rob C

Ah, Rob! The pain of Fame! You have my sympathy. I'm glad I'm (mostly) still anonymous.

Eric

[Rob C]

Ah, Rob! The pain of Fame! You have my sympathy. I'm glad I'm (mostly) still anonymous.

Eric

You're kidding - right? Fame has claimed us for its own: have a website and the world gives you fame or infamy, usually the latter, as in infamy, infamy, everone has it in for me!

;-)

Rob C

[Ray]

In summary, I'll make the following comments to see if we can reach an agreement.

Statements such as, "Focal length of lens used, or Field of View chosen for a composition at a specified shooting position, have no bearing on perspective, period", are simply wrong, or at best misleading.

A more correct statement would be, "Choice of focal length of lens (or FoV) at a fixed shooting position, does not affect the perspective in the resulting 2-dimensional image or print provided that precise viewing-distance instructions accompany such resulting prints, in accordance with the principle that viewing distance should vary in proportion to focal length of lens used and in proportion to print size, (in order to correct for the different perspective created through the use of a particular lens in conjunction with a decision to make a particular size of print), and provided that the viewers of the resulting prints comply with such instructions however inconventient and awkward such compliance may be.

Does that statement appear reasonably accurate, or do you think it should be amended? 

I thought of adding: "The author of the print reserves the right to sue the viewer for non-compliance with correct viewing-distance instructions, because such non-compliance may distort the perspective in the print and reflect unfavourably on the artistic intent of the photographer."

But I guess that would be a bit unreasonable. 

[JohnTodd]

How about a pair of statements:

1) Perspective is solely dependent upon the relative positions of observer and subject. (i.e. the current dictionary definition.)

2) The *perception* of perspective is *additionally affected by* field of view and print viewing distance.

[Eric Myrvaagnes]

How about a pair of statements:

1) Perspective is solely dependent upon the relative positions of observer and subject. (i.e. the current dictionary definition.)

2) The *perception* of perspective is *additionally affected by* field of view and print viewing distance.

That does it.
All along Ray has seemed to be trying to equate "perspective" with "perception of perspective."

Eric

[Ray]

How about a pair of statements:

1) Perspective is solely dependent upon the relative positions of observer and subject. (i.e. the current dictionary definition.)

2) The *perception* of perspective is *additionally affected by* field of view and print viewing distance.

The problem I see here is that all types of perspective, whether considered natural, correct, incorrect or distorted, are perceptions. Without perception there is no perspective. The fact that linear perspective obeys the laws of geometry, does not make it less of a perception.

Distant objects appear small. They are perceived as being small. No-one believes they actually are small, unless one is looking at a world of totally unfamiliar objects. Perspective is an illusion, but an illusion which can be described and measured to some degree mathematically.

[JohnTodd]

Ray,

After I posted, I anticipated that you might have more to say about my possibly artificial distinction.

I realised that there are two distinct phases to consider. Geometrical perspective, part 1 of my distinction, is fixed on the camera sensor at the moment of image-making. Short of deliberately distorting the flat image post-shutter-release, the principle of geometrical perspective can be demonstrated with a ruler and does not change.

That flat image can now be cropped in various ways, printed at various sizes and those prints viewed at various distances by various observers (each of whom can bring any number of cultural assumptions to the experience.) In each case, their individual perception of the image may be different, but the *recorded image doesn't change*. The printed pixels do not shuffle about on the paper based on the distance of the viewer from the paper.