Standards for unmanipulated photo declaration

In photo.net's <a

href="http://www.photo.net/photodb/manipulation">standards for

unmanipulated photo declaration</a>, it states that no use of blurring

tools is allowed. Does this mean no noise/grain removal tools (which

act by blurring), also? Also, my current default tool for conversion

of images from raw format, Adobe Camera Raw, has a "smoothing"

parameter which imposes some amount of blur on the photo, and I

believe Canon's default software always performs a similar smoothing

step; must this be turned off?

I think what the standards should read is that no SELECTIVE use of the blurring tool is allowed. For example you could simulate small depth of field by selectively blurring the background and this would constitue unacceptable image manipulation.

 

I don't think this rule was ever intended to prohibit the use of noise/grain removal tools or internal smoothing of images by digital cameras, but I could be wrong here since I didn't write the standards. Noise/grain reduction really wouldn't be necessary for small web images anyway, so for the purposes of display on photo.net the point is moot in most cases.

 

If you follow the obvious INTENT of the standards rather than a literal interpretation of the words I don't think anyone is going to get too upset, so if you only applied grain removal/smoothing to large uniform areas like the sky, nobody is going to revoke your photo.net membership!

Noise removal/reduction is fine. I will edit the document to reflect this.

I hate to quibble over rules, and I can quite happily live with them as they stand. I'm not sure about the following though:

<p>

<em>No use of perspective correction tools is allowed. Any perspective correction must be done pre-exposure through camera movements or PC lenses</em>

<p>

I'll declare an interest here in that I've just written an article on perspective correction for the static content pages (see "what's new" on the home page).

<p>

It's been quite common in the past to correct perspective in the darkroom optically by tilting the paper and/or negative. The intent of the process is to make the image look MORE like the original scene, not less, thus passing the old FoundView test of not being someting "intended to deceive the viewer" or to alter the image in some unnatural way.

<p>

As I said, I can live with strict interpretation of "no perspective correction", though I'm not sure it's necessary. Manipulation to make the image look more like the actual scene rather than less seems to me to be acceptable.

<p>

Of course the problem is that if you allowed perspective shifts, someone could use them to make the image look LESS like reality, but then that's not much different from using a fisheye lens or a T/S lens to enhance perspective distortion rather than correct it.

<p>

Making a universally applicable set of rules which lead to a single goal is a very difficult task!

Bob, it is very difficult, true. My aim was to make the rules as clear and as objective as possible. It proved impossible to achieve this entirely. In writing the rules, the paradigm for non-manipulation that I had in mind was a slide processed through standard chemistry.

 

Dodging and burning, sharpening, cropping, and contrast and color balance adjustments are all manipulations, in fact. When printing a negative, some of these, such as contrast and color balance adjustments are unavoidable, in that one has to commit oneself to some filtration or paper grade. Other manipulations, such as color space transformations, noise removal, and sharpening are almost impossible to avoid when scanning or preparing an image for digital presentation. If I had prohibited these, no image on photo.net would qualify as unmanipulated.

 

Thus, a certain amount of subjectivity crept in from the necessity of allowing these, subject to the proviso that the result be "realistic" and faithful to the original scene. These exceptions make the standards less clear. And all of these allowed manipulations, when taken to extremes, have the same effect as other operations that everybody agrees are not consistent with the idea of an "unmanipulated photograph", such as cloning. Try extreme sharpening some time. Or slamming the saturation slider too far over when balancing the colors. Where is the point when "acceptable" manipulations become unacceptable? How does somebody know if he has crossed it? On what basis might a moderator protest that our standards have been violated.

 

Perspective correction might have been included in the same class of allowed manipulations as dodging and burning, but since it is not a frequently-used technique, it seemed better in the interest of clarity to exclude it.

 

All those people who are so adamant that images on photo.net should not be digitally "manipulated" should try their hand some time at defining what they mean in some sort of rigorous way.

 

P.S. I don't think perspective correction was allowed under the FoundView rules.

Re perspective correction -- I sometimes use perspective correction tools to correct minor amounts of barrel distortion or chromatic aberration from the lens, without significantly changing the overall view. Surely that should be allowed?

This is going to get "interesting". David, no perspective correction is not allowed, even if it is to correct lens defects. If you would like to label your images as "unmanipulated", you should use a lens that doesn't require digital correction of chromatic aberrations, etc.

 

(I'm really not trying to be a jerk; but I think we are all about to learn the lesson that one man's "manipulation" is another man's "correction".)

Use a camera, film, and a wet darkroom. Straight scan into computer with only corrections allowed are to reproduce the look of the original wet print. Anything else is Digital/manipulated.

 

I can live with that definition.

<blockquote><i>Use a camera, film, and a wet darkroom. Straight scan into computer with only corrections allowed are to reproduce the look of the original wet print. Anything else is Digital/manipulated.<br><br>I can live with that definition.</i></blockquote>

 

I can't. Or rather, I don't care much about it on photo.net, where all it seems to mean is a little bit of information displayed in the photo details, but I would be very annoyed if, say, I wanted to enter a contest that had that as its rules. It would force my photographs, which are taken with a digital camera but usually with the intent of replicating the scene as I saw it just as I would with a film camera, to compete with images that are pure digital manipulation and should be judged on very different grounds.

A couple years ago, I was going to petition the "League of NH Craftsmen" for membership and picked up the paperwork and had an initial interview. They were very clear that any photography was to be film based, produced in a wet darkroom, and not done with any "trickery" that could not be duplicated 10 times in a row.

I'm being a bit of a Devil's advocate here, but if perspective correction is manipulation, why is color correction, gamma correction, digital "dodging and burning", sharpening and grain removal NOT manipulation.

 

One definition which would include this is that the SHAPES of objects in the image may not be changed, added to or subtracted from. Thus you can't remove or add objects (i.e. you can't clone out telephone wires or add an eagle in the sky) or change their shape (correct perspective). I believe FoundView had some wording like this (i.e. the concept of "shapes" or "objects").

 

I think it boils down to the fact that you do have to establish a fairly arbitrary set of rules. THIS is OK, THAT is NOT. You can then argue forever about exactly what the rules should be of course!

 

The simple rule would be that the image as presented should be a true representation of what was in front of the camera at the time of exposure with no attempt to deceive the viewer. That's clearly the SPIRIT of what the rules are trying to establish. There's clearly no harm or deception intended if the photographer corrects barrel distortion or perspective. I don't think many, if any, photographers would actually count that as a "manipulated" image. Manipulated images are those which try to present a scene different from the one in front of the camera at the time of exposure. It also seems very arbitrary that you can correct perspective with a T/S lens, but you can't do the same thing after the image is taken, even though the end result is identical. Thus you have two identical images, one of which is OK, the other of which isn't, even though you could never prove which was the valid image!

 

However such a simple rule is so open to interpretation that instead of discussing what the exact rules should be, we'd end up discussing whether individual images were consistant with the simple goal as stated above or not. You really can't win this game.

 

However as long as everyone plays by the same rules and the rules are somewhat reasonable, then we're OK, whatever the rules actually are.

Bob, I sympathize with your point of view, but your formulation makes the whole thing too subjective and wishy-washy. (This is inevitable to a degree, as I argued above.)

 

Your formulation is basically the FoundView one: don't alter, add, or subtract from the shapes in the image. This is a fairly objective standard, but it doesn't work out very well because of all the well-meaning manipulations that people want to let in that alter, add, or subtract from the shapes, and all the spawn-of-the-devil manipulations that they want to rule out that don't alter, add, or subtract from the shapes. This led to the lame "don't deceive the viewer" escape clause, which is what ended up being invoked for all the hard cases, and made their standard not a real standard at all.

 

Did Ansel Adams deceive the viewer when he burned the clouds into the sky in "Moonrise"? He was trying to communicate the elements of the scene to which he was responding. The clouds weren't important, but they were there. Does someone who creates a stitched panorama deceive the viewer? What if the same moving person appears on both sides of the stitch?

 

Does someone who applies perspective correction deceive the viewer?

As I indicated, I'm pretty sure that perspective correction was one of the things that was prohibited under FoundView. I no longer have the document in front of me, but it was in the FAQ section of the site. No scene ever looked to the human eye like the result of perspective correction. There is nowhere you could stand that would make the scene look like that. The sides of a tall building may be subjectively vertical, but anybody at the scene would not see the sides of the tall building as parallel. Yes, with a PC lens or view camera movements, you can distort the perspective so that the photograph will look better (if you will, correspond to what we imagine we saw), but that is a pre-exposure manipulation, which for some reason is beyond criticism.

I didn't answer your question. Why are gamma correction, etc, not manipulation? As I said above, they are manipulation. If we were dealing with slides, I would not allow them. However, slides are not the only photographic process: there are also negatives and, lately, image files from sensors. The negatives have to be printed and scanned. The slides have to be scanned. And the camera files probably have to be sharpened and down-sampled, and to go through some color space transformations (or contrast and color balance adjustments, which amount to the same thing as a color space change.) All of these steps involve, explicitly or implicitly, decisions about gamma, color balance, resolution, etc. "Manipulation" is part of the process, and the only thing we can say is: don't choose parameters that take the result too far from the original scene -- whatever that means.

 

What about dodging and burning? That is a better example for your argument, because nobody can argue that dodging and burning is a necessary step. If I were being perfectly consistent, I should rule out dodging and burning too, but this manipulation has been blessed by so many decades of tradition, that I could never get away with it. When I was doing darkroom work, cropping was definitely beyond the pale with my teachers. We had to print on filed-out negative carriers to prove we hadn't cropped. Even in that environment, dodging and burning were considered fine.

<i>

No scene ever looked to the human eye like the result of perspective correction.

</i>

<p>

Not true: perspective correction (such as done to a tall building with a view camera/pc lens or the after-exposure equivalent) is no different than using a wider lens, leveling the camera and cropping the foreground. So if using wide lenses and cropping (which are obviously allowed) isn't manipulation, then neither is pre- or post-exposure pc correction.

<p>

Plus, part of the correction Bob is talking about DOES make the result more like the human eye: correction for barrel, pincushion

or fisheye distortion.

Bob is right and Brian is right, and we can't have a "wishy-washy" mix and we can't cover all cases with any set of rules Brian or anyone else would write. So here is the solution...

<p>

DISTINGUISH BETWEEN THE ACTION OF MANIPULATING & THE INTENT OF THIS MANIPULATION. (And ask users to be specific on both.)

<p>

"I think it boils down to the fact that you do have to establish a fairly arbitrary set of rules. THIS is OK, THAT is NOT. You can then argue forever about exactly what the rules should be of course! The simple rule would be that the image as presented should be a true representation of what was in front of the camera at the time of exposure with no attempt to deceive the viewer. That's clearly the SPIRIT of what the rules are trying to establish." - Bob A.

<p>

Exactly ! The SPIRIT is the INTENTION. With ANY manipulation - i.e digital or not - you are either:

<p>

a) TRYING TO BRING BACK INTO A PROPER JPEG WHAT YOUR PICTURE LOOKS LIKE - AN ATTEMPT AT REALISM

<p>

or b) TRYING TO BRING BACK YOUR PICTURE TO RESEMBLE MORE WHAT YOU SAW THAN YOUR PICTURE OF IT DOES - AN ATTEMPT AT HYPER-REALISM, if I may use this word in this context.

<p>

or c) TRYING TO IMPROVE A BIT ON THE REAL THING YOU PHOTOGRAPHED - an attempt at an AESTHETICAL IMPROVEMENT UPON REALITY OR UPON THE REALITY AS IT WAS PHOTOGRAPHED

<p>

or d) TRYING TO CREATE A NEW AND DIFFERENT IMAGE THAT DOES NOT LOOK LIKE THE REAL THING AT ALL - IRREALISTIC or SURREALISTIC approach.

<p>

These are your 5 categories - 4 in fact, but d) actually branches out again in 2 very different currents.

<p>

Conclusion: Brian's set of rules is exactly as valid as any, and does not cover ALL possible cases because that would be impossible. CONCEPTS do cover (in theory at least) all aspects of specific contingencies, whereas rules like this set of rules CAN NOT be perfect. So, let's take this set for what it is and live with it. And let's maybe create another set of 4 checkboxes, which would be the a), b), c), and d) above. A user asked to declare manipulated images the way Brian conceived it is just fine. But then the user would just need to chose between a, b, c, and d, as a way to declare what was the intention he had with his manipulations.

<p>

For example, perspective correction cases enter the category Manipulated > b) . Etc. That way, no further confusions.

<p>

All you need is to ask the author to check one of the 4 cases, a, b, c, or d for each upload. AND ALSO to declare "Manipulated" or not, based on the rules you just published.

Mark, unless I'm mistaken, the human head does not come with an interchangable wide-angle eye, or tilts, shifts, and swings. I stand by my statement that almost no PC corrected image looks like what a human eye would see at the scene.

 

All you are saying is that there are pre-exposure manipulations that can produce the same effect as the post-exposure manipulations, and that since the pre-exposure ones are not considered manipulations the post-exposure ones shouldn't be either.

If we are going to open that gate, then are you going to admit, as unmanipulated, photographs produced with Photoshop operations if they only do what a "creative star filter" or smearing vasoline on the lens would have done pre-exposure? Are we going to say that you don't have to check the manipulated box if you can describe some Rube Goldberg camera and/or setup that would have produced the same image on the negative or sensor as the post-exposure operation produced?

 

That might be a one criterion, but it isn't the one I chose. I've tried to define manipulation in terms of the most "straight-forward" possible transfer from the recorded image to what is presented. To the extent that the processing requires some selection of parameters such as gamma, color space, etc, then these should fall within a range that can be regarded as "true" to the original scene.

 

My definition would actually define as "manipulated" transformations intended to make an image look exactly like what the human eye would have seen. For example, suppose I took the picture with a telephoto lens, but I want to present it with "normal" perspective, which is theoretically the perspective that a human eye would have perceived. This is possible with sofware, but it would be a manipulation by my definitions. So be it, since it was not what was captured.

 

All this isn't a question of whether the image is good, or even whether it is allowed on photo.net. It is a question of standards for what post-exposure operations make an image "manipulated".

I am sorry but your post is not imo an answer to mine in any way I can understand. I am not saying that you are wrong or discussing your set of rules - AT ALL.

<p>

I even wrote this:

<p>

"Perspective correction cases enter the category Manipulated > b)"

<p>

What more can I say ? Then you tell me that perspective corrections are manipulation to you. Yes, agreed !

<p>

My post goes a bit further than that, if you are interested enough to read it properly...

<p>

What I am proposing is NOT to change anything AT ALL to your set of rules. BUT rather to consider ADDING separately a sub-menu for people to specify what was the INTENT of their manipulations.

<p>

What for ? Well, to allow a more precise targetting thanks to a conceptual frame. The idea on my mind is this:

<p>

I have just uploaded a picture of a bottle that is considered manipulated by your set of rules. I have absolutely no problem with that by the way. On another hand, I have uploaded, as you know, the wildest photo-montages really HEAVILY manipulated in PS. You and I both know that these are 2 absolutely different kinds of images.

<p>

So, what I'm thinking is: fine, both types are manipulated, but not the same way and not with the same intention. You stated in another thread today that you want people to judge each image based on its own criteria: well, so do I. So, you categories have to be specific enough to make a distinction between shots like this bottle and heavily manipulated PS composites - for example.

<p>

My post above is a very simple way to categorize images in a finer way, in order to get in the same category later images that somehow really belong together. That's all it is. If you want to ignore it or presume it's stupid, fine by me, but those 4 categories of intent are a standard worldwide in artistic categorization of "photo-based" images - for what ever that means...

<p>

If you read what Bas and Doug Burgess stated in Bas's thread, you will see that people are already showing that they do not want their say "pretty real photos" to be categorized together with wild PS montages. I can understand them. I personally don't care what will be the picture next to my wild PS montage. It can be the most conventional black & white, and it can be Anna's work - same thing to me. My pictures are what they are no matter in which category they are. BUT as you know, many people do care... So I'm just offering you a tool to make further distinctions, that's all.

<p>

Your position seems to be to teach people a lesson, Brian. Well, you don't have to teach it to me, because I am exactly of your opinion as far as all the digital issues are concerned. But you had already 2 interesting comments in this other thread. Lex saying "Careful what you wish for" and William G, saying that you should not have used the word ghetto... Basically, my view is that you should read with a bit more respect the various views that people express in these forums, and also work on your PR - if I may...

<p>

Of course people are going to make noise, like Bob here, or Doug and Bas there... But you did the right thing... The only thing is: what's the next step ? Will there be categories ? As long as it is just a checkbox issue, no big deal, but the day "manipulated" images of ALL sorts will be separated from the "non-manipulated" ones, you're going to have a real rebellion here... The tool above is just meant to prevent that - nothing more. And again, I fully agree with your set of rules as they are as perfect as they could possibly be. I just feel you need another vertical set of criteria based on INTENTS so that all the scopes of works are covered and that the future categories (if any) will make sense.

<i>

Mark, unless I'm mistaken, the human head does not come with an interchangable wide-angle eye, or tilts, shifts, and swings.

</i>

<p>

Tilts and swings are irrelevent. All we're talking about is shifts, first of all.

<p>

And shift are nothing but crops, which you've already said are OK before or after exposure.

<p>

The human eye works just like a rectilinear camera lens. Single point perspective. If you point it up, the lines will converge. If you keep it straight ahead, they will not.

<p>

I stand by my statement that almost no PC corrected image looks like what a human eye would see at the scene.

</i>

<p>

Stand by it if you want, but it's a statement so vague as to be meaningless.

<i>

All you are saying is that there are pre-exposure manipulations that can produce the same effect as the post-exposure manipulations, and that since the pre-exposure ones are not considered manipulations the post-exposure ones shouldn't be either.

</i>

<p>

No, I'm saying I don't consider using a wide angle lens manipulation.

<p>

If a wide angle lens is manipulation, than I'd have to ask, what focal length represents an unmanipulated image. The only possible answer to that question is "the lens that covers the same angle the print will subtend when it's being viewed" because this gives technically correct single-point perspective. But of course that means a print will become manipulated if it's held at the wrong viewing distance! It also means a photo can become manipulated or unmanipulated on the basis of post-exposure cropping, which violates your statement that cropping is OK. Ergo, there is no correct focal length. Ergo, wide angles are not manipulation. Ergo wide angles plus cropping are not manipulation.

Mark, you are missing the point. Of course it is not "manipulation" to use a wide-angle lens.

 

However, it requires a transformation of an image that was not captured with a wide-angle lens to make it look like it was. For example, you can use the Transform function of Photoshop. Such as a transformation is a manipulation by our standards.

 

If you want to declare an image that looks like a wide-angle lens image as unmanipulated on photo.net, then you should shoot it with a wide-angle lens.

 

Hypothetical counterfactuals don't count. You can say, well I could have used a wide-angle lens, and this transformation only makes it look I did. So its not manipulation. Where does that stop? I could have gotten up early and driven in the dark and taken that picture when the sun was rising and there were red clouds in the sky, so cloning in a sun-rise isn't manipulation. Only, you didn't, and it is.

My apologies, Brian... I didn't understand why your post seemed so far away from an answer to mine... But now I do... There was another Mark, with K, involved in this thread, and your reply was to him and not to me...

<p>

I'm Marc with a C - I must try to remember that, even if many people don't...:-) Sorry again for that misunderstanding.

You'll never get a set of self consistant rules.

 

For example it's not OK to clone out a telegraph pole that's in the center of an image, but it is OK to crop out a telephone pole at the edge of an image. How can you rationalize that?

 

You can dodge and burn, but you can't dodge and burn to the extent that it alters the image by removing something (e.g. a cloud). What if it 99% removes the cloud? or 98%, or 90%?

 

You can distort the image with a T/S lens when you take the shot, but you can't distort the image in exactly the same way with PhotoShop.

 

I could go on, but I won't. This is NOT a criticism of Brian's standards, which are, I think, a very good attempt at a workable set of rules. What I am trying to point out is that it is very, very, very difficult to define this problem with a rigid set of rules, because you'll always be able to find exceptions.

 

Defining a manipulated image is like defining pornography (or art!). It's hard to define, but you know it when you see it - and it's different for different people. FoundView appears to have thrown in the towel since theie website seems to have gone.

 

Brian's rules are as good as any. It's not like this is a contest or anything and there is no prize for winning!

Refusal to include stitching and skews will severely compromise your

ability to get people to participate in this proposal. Please

recognize that if you dont' get most people to accept where the lines

are, then the whole thing will collapse. Compromise on this one . .

for the sake of the success of the venture.

 

I think I can find a copy of 'foundview'. I don't think skew or PC

was mentioned specifically.

 

It's not a wishy washy thing. Is this the equivalent of what a camera

can do? If, yes, then it's OK. .

... I have to disgree with your objection, and to agree with Bob. What people need to understand is that 1) a manipulation is not a crime 2) that stiching involves cloning and involves the addition of several images that are NOT take as 1 single shot. Based on what Brian wrote above, if you want to call a panoramic a "non-manipulated" image, then you have to shoot it with a panoramic camera.

<p>

Leave this set of rules as they are, simply because no set CAN be perfect. What matters is now imo what difference this checkbox will actually make. Will there be categories, separate viewing of different categories, etc. Regards.

Brian, it's hard to debate with you because you keep changing your argument! Is it about being faithful to what's on film, or the original scene? You seem to want to have it both ways when it suits your argument.

<p>

Nontheless, I don't really care. I don't do architecture, so I could care less, in fact. I'm going home.

<p>

<i>

if you dont' get most people to accept where the lines are, then the whole thing will collapse.

</i>

<p>

Never happen. Nearly everyone is always going to have a problem with one rule or another. The best you can do is convince people that the rules as a set are better than no rules at all, which I guess they are.