Pages: 1 2 3 [4] 5 6   Go Down

Author Topic: when is wide too wide?  (Read 82035 times)

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: when is wide too wide?
« Reply #60 on: May 17, 2013, 09:11:19 am »

I only accept facts that make sense, Rob. JJJ has made a very categorical statement as follows:
Quote
Cropping has zero effect on perspective, nor does print/sensor size or viewing distance and the biggest myth of all, focal length.
So let's examine this statement from a rational and logical perspective. First, let's take the  phrase, "Cropping has zero effect on perspective".

The question that immediately springs to my mind here is, "The cropping of what precisely has zero effect on the perspective of what precisely?"

I can find numerous examples where cropping really does have zero effect on perspective, just as I can also find numerous examples where cropping has a very clear and obvious effect on perspective. I'll mention a couple.

As I sit at my computer cropping an image that I'm processing, I'm quite certain that the cropping of that image on the monitor has no bearing whatsoever on the perspective of the elements in another photographic print that is hanging on the wall. The two are obviously unrelated.

Okay! That's silly you're probably thinking. Obviously JJJ is referring to the cropping of a specific image taken with a specific lens, and he's referring to the perspective of specific elements in that image.

So let's talk about a specific image and try to imagine how the perspective of elements in that image might change after cropping, in accordance with the following definition of perspective which you may or may not agree with. If you don't agree with this definition, then that could be the cause of the confusion.

The definition I'm using is as follows: "The appearance of objects, buildings, etc, relative to each other, as determined by their distance from the viewer, or the effects of this distance on their appearance", and I'm applying this definition to  photographic images.

A key concept in this definition is, 'appearance of objects relative to each other'.  Now surely it's not difficult to appreciate the consequences of removing objects within a scene by cropping them out so they no longer exist. Do I have to spell it out?

An object in a photographic scene may appear small and distant relative to another object in the foreground which is larger and therefore appears closer. If one removes the larger object (or objects) in the foreground through cropping, then clearly one has broken a relationship and therefore changed the perspective in at least two ways. (1) One or more of the elements in the image has no perspective at all because it no longer exists. (2) The remaining elements in the image have a changed perspective due to a change in the relationship with other objects in the image that no longer exist.

How anyone could argue with this obvious fact beats me. But I do understand what JJJ and others are trying to say in their muddled way. I believe they are trying to say that the perspective of elements in two images of the same scene with the same FoV, will only be influenced by the position from which the shot was taken. By what process that equality of FoV was achieved is irrelevant to the perspective of the elements in that image. The equality of FoV could have been achieved by using different focal lengths of lenses with different formats of cameras, or it could have been achieved by different degrees of cropping in post-processing, using the same focal length of lens on different formats of cameras.

The main point here is that different focal lengths of lenses may have been required, or different degrees of cropping may have been required to achieve that equal perspective, therefore it is not correct to claim that cropping or focal length has no bearing on perspective.

What is correct to say is that changing position will always affect perspective to some degree, whereas cropping and/or changing focal length of lens may not always change perspective, but sometimes it might.

An example of achieving equal perspective using different focal lengths of lens would be using a 50mm lens on a Canon 7D and an 80mm lens on a Canon 5D2. Provided the shooting position is the same, and provided one does not perform additional and different cropping during post-processing, the perspective will always be the same in the images from both cameras. The image from the 50mm lens on the 7D receives more in-camera cropping than does the image from the 80mm lens on the 5D2, and is therefore said to be equivalent to an 80mm lens on the 5D2. However, a 50mm lens on a 7D is not equivalent to a 14mm lens on a 5D2, and without significant cropping of the 14mm shot in post-processing the 14mm lens will produce an image with a very different perspective, even after correction for volume anamorphosis.

I don't think I can make this much clearer, but if anyone thinks I am presenting a fallacious, bogus or muddled argument, then please point out the flaws in my logic.   ;)

OH. MY. GOD!
My jaw just hit the floor so hard, I now have broken teeth.

Logged
Tradition is the Backbone of the Spinele

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: when is wide too wide?
« Reply #61 on: May 17, 2013, 09:53:31 am »

Really long discussion about perspective here, what's it about?

If you shoot from the same position and crop to the same field of view the perspective will be the same regardless of focal length. If you move to another position, then perspective will change. Otherwise it will not. Oh well, pointing the camera in a slightly different direction will also change how the final picture appears, but with fixed camera position and direction and "perfect" rectilinear lenses you can take a 14mm lens and crop it down to produce the exact image of the 500mm lens, at a reduced resolution of course. As soon as you move and/or change camera direction you get a different view. Clear?

This article is quite good: http://en.wikipedia.org/wiki/Perspective_distortion_%28photography%29
« Last Edit: May 17, 2013, 09:57:16 am by torger »
Logged

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: when is wide too wide?
« Reply #62 on: May 17, 2013, 10:20:15 am »

Already said that torger.
But it appears that there is another universe with alternative physics, but interestingly you can still post to LuLa.  ;D
Logged
Tradition is the Backbone of the Spinele

RFPhotography

  • Guest
Re: when is wide too wide?
« Reply #63 on: May 17, 2013, 11:00:50 am »

But can you prove this, Bob? Anyone can repeat a mantra. Merely repeating something without explanation or examples doesn't make it right.

I'm quite certain that the wide-angle shots that I take, result in images which display elements with a different perspective than the same elements in the narrower-angle shots that I take from the same position.

I can also demonstrate that such differences in perspective, as displayed in the wider-angle shots, can be altered through cropping and made the same as in the narrower-angle shots, thus proving that cropping can change the perspective in a photographic print or image.

Perhaps your definition of perspective is different from mine. What's your definition of perspective, Bob?

The one I'm using is the following, which is similar to the definition I've found in all the dictionaries that I've checked.

You say you can disprove the laws of physics?  If you can, that's probably worth a Nobel Prize.  So do it.  Demonstrate it.

Do the same test that others have talked about here.  I've done it in the past and I've seen others do it.  There are plenty of examples on the web doing it too.  Take a shot with a 50mm lens and a 200mm lens from the same spot.  Now crop the 50mm shot so it has the same AOV as the 200mm shot.  The spacial relationship of elements within the shot will be the same.  That's perspective.  It's got nothing to do with what gets cropped out or any other nonsense. 
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: when is wide too wide?
« Reply #64 on: May 17, 2013, 11:04:19 pm »

Really long discussion about perspective here, what's it about?

If you shoot from the same position and crop to the same field of view the perspective will be the same regardless of focal length. If you move to another position, then perspective will change. Otherwise it will not. Oh well, pointing the camera in a slightly different direction will also change how the final picture appears, but with fixed camera position and direction and "perfect" rectilinear lenses you can take a 14mm lens and crop it down to produce the exact image of the 500mm lens, at a reduced resolution of course. As soon as you move and/or change camera direction you get a different view. Clear?

This article is quite good: http://en.wikipedia.org/wiki/Perspective_distortion_%28photography%29

Torger,
What you've written above seems generally correct to me, with a couple of caveats. For example, if the focal length comparisons are too extreme, it might no be possible to identify the elements in the lower resolution image which are claimed to have the same perspective.

There is also a basic confusion in the above statement between the perspective of elements in the real scene as opposed to the perspective of elements in the photographic image, as implied by the following statement, "If you move to another position, then perspective will change. Otherwise it will not." I presume what is meant here is that the perspective of elements as seen in the real scene do not change, unless one changes position. That makes complete sense to me.

However, surely you guys have taken enough photographs by now to realise that the taking of a photograph is a process of selection and exclusion. The perspective of elements in a photograph can be made to look quite different to the perspective of the same elements in the real scene as seen with the naked eye, through a process of lens selection, cropping, print size and viewing distance. Haven't you noticed?

I haven't checked the Wikipedia article to see if there are any other confused statements. I'm too busy trying to knock some sense into certain posters on LL.  ;)

If cropping has zero effect on perspective, then why is it necessary to crop the 14mm shot to to demonstrate that the perspective is the same as in another shot taken with a longer lens?
Are you guys really not able to see the absurdity of your arguments? Ah! Perhaps 'Alice in Wonderland' is your favourite book. C'mon! Admit it!  ;D

The reason for cropping a wide-angle shot in these circumstances is to demonstrate that position always has an effect on perspective, not to demonstrate that cropping or changing focal length never has an affect on perspective. Can't you see the difference? Blimey! ;D
Logged

RFPhotography

  • Guest
Re: when is wide too wide?
« Reply #65 on: May 17, 2013, 11:45:12 pm »

Ray, what you're referring to is the idea of apparent perspective distortion.  The apparent contraction of distances by longer lenses and the apparent lengthening of distances by short lenses.  The key there is the word apparent.  Do the tests I and others have mentioned.  If you disprove the laws of optical physics, I think the entire membership of this board with support your nomination for a Nobel.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: when is wide too wide?
« Reply #66 on: May 18, 2013, 01:39:22 am »

You say you can disprove the laws of physics?  If you can, that's probably worth a Nobel Prize.  So do it.  Demonstrate it.

I've never said any such thing, Bob. If you think I'm implying that I can disprove the laws of physics, then please specify which laws you are referring to. All 32,769 of them?  ;)

Quote
The spacial relationship of elements within the shot will be the same.  That's perspective.  It's got nothing to do with what gets cropped out or any other nonsense.
 

Absolutely true! Maybe now we're getting somewhere. The perspective, or spatial relationship, of elements within a shot have nothing to do with elements that are outside of that shot. This is an important principle in understanding the perspective effect of a wider-angle lens.

If it's true that the perspective of elements within any photographic image can only be assessed in relation to those elements that are present in the image, and I maintain this is true, then introducing new and different elements into the image, by using a wide-angle lens from the same position, changes the spatial relationship of those same elements in the uncropped, wide-angle image. New spatial relationships are formed.

Quote
Ray, what you're referring to is the idea of apparent perspective distortion.  The apparent contraction of distances by longer lenses and the apparent lengthening of distances by short lenses.  The key there is the word apparent.

Bob, all photographic images are appearances. Everything about them is 'apparent' and a distortion to some degree. It's inherent in the system. Perhaps the most obvious distortion is the fact they are 2-dimensional representations, whereas reality is 3-dimensional.

There is a common distortion called Volume Anamorphosis which wide-angle shots frequently exhibit in the corners. DXO tools can fix it.

But never mind. If you understand that perspective is a spatial relationship between the specific elements of the scene that have been captured, then it should not be difficult to understand that by introducing new elements into the scene, by using a wide-angle lens, you will also unavoidably introduce new spatial relationships. The distant mountain peak not only has a spatial relationship with other parts of the mountain, glaciers, waterfalls and trees etc, at the foot of the mountain, but the whole mountain might have a spatial relationship with the face of your girlfriend that appears in the foreground of the wide-angle shot. That spacial relationship is destroyed if you crop out the face of your girlfriend. The spatial relationships of the mountain would then change and as a consequence the mountain would appear closer on the same size print. What could be easier!  ;D

Logged

RFPhotography

  • Guest
Re: when is wide too wide?
« Reply #67 on: May 18, 2013, 06:59:53 am »

Ray, the number of elements within an image is irrelevant. You're wrong.  Period.  Full stop.  And I'm done with this nonsense.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: when is wide too wide?
« Reply #68 on: May 19, 2013, 08:41:16 pm »

Ray, the number of elements within an image is irrelevant. You're wrong.  Period.  Full stop.  And I'm done with this nonsense.

LOL!  No elements, no picture. Goodness gracious me!

I'm talking about perspective as it appears in a photographic image consisting of numerous elements. I understand completely that perspective as a theoretical geometric model, as applied to a real but static scene, as viewed by the naked eye, is changed only by the position of the viewer. Even moving one's eyeballs to one side is a change in position.

However, the camera can do things which the unaided human eye cannot. You must have noticed.

It can crop and enlarge, or diminish the field of view according to choice of lens. When the image is processed, further cropping may be applied, and when the image is printed the size can vary according to purpose.

Such changes do not alter the laws of physics or the rules of geometry. They merely change the circumstances within which the same laws apply. In a photographic image the eye perceives a distortion of realty. It cannot be otherwise. All photographs without exception are distortions of reality. Only the degree of distortion varies. It never occurred to me that some folks might not be aware of this fact.

A part of this distortion is a change of perspective resulting from practices such as cropping, enlargement of FoV due to use of a wide-angle lens, and physical size of image or print after processing.
Logged

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: when is wide too wide?
« Reply #69 on: May 20, 2013, 05:21:46 am »

Here you go Ray.....




Logged
Tradition is the Backbone of the Spinele

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Re: when is wide too wide?
« Reply #70 on: May 20, 2013, 05:40:54 am »

I have been following this thread with interest and finding it very informative. Ray's reply
#68 makes sense to me but then again it is a "concept" that is difficult to get a proper handle on. :-\

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: when is wide too wide?
« Reply #71 on: May 20, 2013, 10:13:28 am »

Thanks, Stamper. I'm glad at least someone reading this thread is able to think outside of the box.  :)
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: when is wide too wide?
« Reply #72 on: May 20, 2013, 10:21:59 am »

Logged
== If you do what you did, you'll get what you got. ==

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: when is wide too wide?
« Reply #73 on: May 20, 2013, 11:45:41 pm »

Why does this all look like a deja vu ?

Here's why:
http://www.luminous-landscape.com/forum/index.php?topic=61388.0
http://www.luminous-landscape.com/forum/index.php?topic=67945.0

Cheers,
Bart

Hi Bart,
There are many topics that are raised again and again on LL, some of which are raised far more frequently than the topic of 'perspective'. One such topic would be Expose To The Right, or ETTR.
You've provided a couple of links to previous discussions on perspective but have missed out perhaps the longest thread that I link below.
http://www.luminous-landscape.com/forum/index.php?topic=61535.120

The reason why I raise this issue again is because I am genuinely amazed that so many apparently experienced photographers seem to be in such a state of denial about this issue and present the most illogical of arguments in order to maintain their position.

Thinking about the reasons for this, I wonder if it's because the thought that any scene as represented in a photograph might not be real, or that it is a distortion of reality, is too uncomfortable. Perhaps people like to kid themselves that the camera really does never lie. And whilst such people might admit when pressed, perhaps grudgingly, that the camera does or can lie, they perhaps kid themselves that it lies only a little bit, or lies only sometimes.

The main argument presented by those who maintain that focal length of lens, or degree of cropping, has no bearing on the perspective of elements, as seen in a photographic image, is as follows.

If one takes two shots from exactly the same position, using widely different focal lengths of lens, resulting in two images with an apparently different perspective, and one crops the wider shot in post processing so it has the same Field of View as the other shot, then both shots will be identical in terms of perspective, although probably not in terms of resolution.

Now to claim that this experiment proves that neither cropping nor focal length of lens changes perspective is where the absurdity comes in, or the Alice in Wonderland aspect.

Let's examine this absurdity, but first the facts which I hope we agree upon.

(1) Two images can 'appear' to exhibit a different perspective when taken with different focal lengths of lens.

(2) All photographs are 'appearances', not to be confused with reality.

(3) The image taken with the wide-angle lens clearly and unambiguously gives the impression that certain objects, because of their small size in relation to other objects or elements in the image, appear very distant.

(4) The image taken with the longer focal length gives the impression that those same elements that appear so distant in the wide-angle shot are now much closer to the viewer. The perspective from the position of the viewer is therefore different because it appears different, and all photographs are appearances.

(5) If one crops away the large objects in the foreground of the wide-angle shot that are responsible for creating the impression that elements in the background are so distant, lo and behold!, surprise! surprise!, the perspective in the wide-angle shot now appears the same, provided one views same size prints from the same distance.

Now let's examine what this experiment really does demonstrate, in my opinion.

(1) If we have to crop the wide-angle shot in order for the image to have the same perspective as in the narrower-angle shot, then that clearly demonstrates that cropping affects perspective, surely.

If cropping doesn't affect perspective, as JJJ claimed it doesn't earlier in this thread, then why is it necessary to crop the wide-angle shot? Ah! Perhaps JJJ thinks that the different perspective, as seen in the wide-angle shot before cropping, is merely an illusion, an aberration, a distortion, a trompe l'oeil, and that such illusions of perspective can always be magically dispelled through the act of cropping. Wow! Are we now back into 'Alice in Wonderland'?  ;D

What this experiment really demonstrates is the relationship between cropping and focal length of lens, and the fact that any lens used with a given format of camera can be effectively converted to a longer focal length through cropping, but unfortunately with consequent loss of resolution.

The experiment also demonstrates that in order to get photographic images to display the same perspective when the shots are taken from the same position, it is necessary to either use lenses which have the same focal length, or create the same equivalent focal length through cropping in post processing.

But what happens when we want the perspective that only a wide-angle lens can provide, and all we have is, say, a normal lens? No problem, I can hear Bart and JJJ declare. Just take a number of shots and stitch.

Really! I've never succeeded in taking shots for stitching a wide panorama without moving the position of the lens between shots. What's the technique here, Bart? On the one hand you're claiming that only position determines perspective, and that only a change in position results in a change in perspective, not a change in focal length, and to demonstrate this principle you are changing the position of the lens in order to create the equivalent of a single wide-angle shot which has been taken from one precise and exact position.

Something fishy going on here!  ;D
Logged

jjj

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4728
    • http://www.futtfuttfuttphotography.com
Re: when is wide too wide?
« Reply #74 on: May 21, 2013, 07:25:59 am »

But what happens when we want the perspective that only a wide-angle lens can provide, and all we have is, say, a normal lens? No problem, I can hear Bart and JJJ declare. Just take a number of shots and stitch.
Really! I've never succeeded in taking shots for stitching a wide panorama without moving the position of the lens between shots. What's the technique here, Bart? On the one hand you're claiming that only position determines perspective, and that only a change in position results in a change in perspective, not a change in focal length, and to demonstrate this principle you are changing the position of the lens in order to create the equivalent of a single wide-angle shot which has been taken from one precise and exact position.

Something fishy going on here!  ;D

Only in your mind Ray.  :P
There are a couple of things wrong with this particular bit of your batty rambling essay. No-one suggested using a narrower lens to replicate the wide field of view and slightly changing angle of camera is not noticeably changing camera distance/position relative to subject, which is the bit that actually alters perspective and which you still do not understand. All you are doing in that case is making a small sensor capture the scene as if it were a much bigger sensor with a wider angle lens.
And finally here's a flickr pool dedicated to doing making wideangle shots from narrower lenses.

The reason why I raise this issue again is because I am genuinely amazed that so many apparently experienced photographers seem to be in such a state of denial about this issue and present the most illogical of arguments in order to maintain their position.
Rather than proposing a lot of bizarre ideas that seem to rewrite physics, go and take some photos that demonstrate exactly what you mean and prove us idiot photographers and all the physicists wrong.
« Last Edit: May 21, 2013, 07:33:13 am by jjj »
Logged
Tradition is the Backbone of the Spinele

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: when is wide too wide?
« Reply #75 on: May 21, 2013, 11:00:41 am »

There are a couple of things wrong with this particular bit of your batty rambling essay. No-one suggested using a narrower lens to replicate the wide field of view and slightly changing angle of camera is not noticeably changing camera distance/position relative to subject, which is the bit that actually alters perspective and which you still do not understand. All you are doing in that case is making a small sensor capture the scene as if it were a much bigger sensor with a wider angle lens.

There seems to be some confusion in your above statement, JJJ. I have never asserted that changing position does not change perspective. Where did you get that idea? I understand quite well that any change in the position of the camera lens causes a change in perspective, whatever the focal length of lens. What I'm arguing against is the absurdity of the logic in the claim that only a change in position of the camera lens can result in a change of perspective in a photographic image, and that a change of focal length of lens, and/or a change in effective focal length of lens due to cropping, can have no effect on the perspective or the spatial relationships of elements within the image.

The proof that is frequently used to demonstrate that focal length has no bearing on perspective is the experiment of either cropping the wide-angle shot to the same FoV as the telephoto shot taken from the same position, demonstrating that perspective then becomes the same, which I agree it does, or stitching a number of shots with a telephoto lens to duplicate the perspective that one would get with a single wide-angle shot.

What I would like you to answer and explain to me is how one can achieve the same perspective, as in a single shot from a wide-angle lens, by taking a series of shots from different positions, whilst simultaneously maintaining that any change in position must result in a change in perspective.

Or do you perhaps believe that it is not possible to duplicate the perspective in a single wide-angle shot by stitching a number of shots taken from different positions? If you believe it is true that this is not possible, then by what sort of logic can you claim that focal length of lens has no bearing on perspective?

Alternatively, do you believe that turning one's head from the left to the right when viewing a real scene, does not constitute a change in position and therefore does not result in a change in perspective?

Logged

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: when is wide too wide?
« Reply #76 on: May 21, 2013, 11:07:08 am »

Ray,

To take the discussion further you need to post images that proves your point. Here's an experiment for you:

1. Mount the camera on a tripod, point it at some area with geometric objects, like a playground or similar
2. Shoot two photographs with one wide angle and one longer focal length (or use a zoom). Do not move the camera or change its direction.
3. Make sure that you apply distortion correction in your raw converter so the lenses render as close as possible to perfect rectilinear
4. Crop the wide angle so it has the exact same FOV as the longer focal length
5. Compare.

If I understand you correctly you claim that there will be some more difference apart from lower resolution. I have not been able to understand what difference that would be from all your long posts. Can you describe in a short way what you mean the difference between cropping a wider and using a longer focal length would be?

I can give you a lead: the experiment above will show that there is no difference to be found.
« Last Edit: May 21, 2013, 11:15:04 am by torger »
Logged

RFPhotography

  • Guest
Re: when is wide too wide?
« Reply #77 on: May 21, 2013, 11:20:54 am »

That's about the 5th or 6th time someone has suggested the exact same test.
Logged

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18087
  • When everyone thinks the same, nobody thinks
    • My website
Re: when is wide too wide?
« Reply #78 on: May 21, 2013, 11:35:34 am »

...
4. Crop the wide angle so it has the exact same FOV as the longer focal length
...

If I am allowed to interpret Ray's position:

If you CROP to achieve "the EXACT same FOV as the longer focal length," than you are effectively comparing "the longer focal length" with "the longer focal length," which, as a tautology, then proves nothing.

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: when is wide too wide?
« Reply #79 on: May 21, 2013, 12:06:16 pm »

If I am allowed to interpret Ray's position:

If you CROP to achieve "the EXACT same FOV as the longer focal length," than you are effectively comparing "the longer focal length" with "the longer focal length," which, as a tautology, then proves nothing.

Exactly. That's why I don't really understand what Ray wants to point out.

That you actually use different focal lenghts to be able to stand closer or farther away from a subject and therefore render it with different perspectives (wider or more compressed) is obvious, but then we change position to change perspective, focal length is just for cropping, ie identical (assuming you increase f-number proportionally to achieve the same DoF too). It seems to me that Ray tries to convince us that a longer focal length does something different than cropping in terms of perspective.

Cropping from the center is of course important to achieve the same effect, if cropping off-center one will instead simulate a shifted lens, and shift is as we know often used for perspective control in architecture photography.
« Last Edit: May 21, 2013, 12:09:22 pm by torger »
Logged
Pages: 1 2 3 [4] 5 6   Go Up