Commentary: Digital Manipulation vs. Digital Correction - Towards a Standard?
When someone hears that an image has been digitally manipulated they automatically assume that the image now strays more from the original scene than before the manipulation. But what do they think when they hear that the image has been digitally CORRECTED?
The Problem: Public Perception of Digital "Manipulation"
What's your first thought when you hear that an image you're viewing has been digitally "manipulated"? I'm betting that even the most open-minded of us automatically think they're viewing an image that now strays more from reality than the "original" pristine image file. Perhaps something has been added to the image ("hey...I bet that bear was photographed in a zoo and dropped into that beautiful scene"). Perhaps something has been removed ("...the photographer MUST have removed some twigs and branches - no background is THAT clean!"). But invariably you're thinking the image is less real than before the manipulation.
In another commentary (Reality in the Age of Digital Photography) I argued that there is no a priori reason that digital manipulation is in the direction of reducing the reality of an image. No image recording device - analog or digital - is perfect. The output of any camera is always only a reasonable facsimile of "reality". Alterations to the output (i.e., the image) can just as easily be in the direction of reality re-creation/enhancement as they can be in the direction of reality reduction.
In Search of a Better Term
Whether or not the term "digital manipulation" is being used incorrectly or now carries the connotation of image adjustment ONLY in the direction of reality reduction is pretty much irrelevant. Trying to change world-wide public opinion on the meaning of this term is akin to swimming upstream during a flash flood. Why bother? Instead, I propose we accept the public's understanding of "digital manipulation" and simply introduce a new term which more clearly indicates that an image has been modified to make it a MORE accurate record of the original scene.
Do we actually need a new term? I'm of the opinion that without a positive term for minor digital modificiations we're forcing many "would-like-to-be-honest" photographers underground. Because of their fear of being accused of "manipulating" their images it wouldn't surprise me if many photographers are hesitant to comment on ANY changes they've made to an image, regardless how trivial the modification is. And, in very rare cases, some photographers have probably taken the easy way out and mislead viewers about how pristine a specific image really is. Personally, I've heard the "Aha, you're manipulating that image..." comment during field shoots when non-photographers watch me convert RAW image files to TIFFs (without changing any of the default settings of my RAW converter!). The real irony is that some photographers are very likely keeping quiet about minor image edits when all they've done is adjust the images to make them MORE realistic. So, let's introduce a new term and let the photographers come clean.
But what term or phrase is most appropriate? Here's a few options (and I make no secret about my preferred term):
Digitally Re-touched? I've been seeing this term being used a little more lately. It does imply that the image manipulation is minor, i.e., that any changes in the file are simply minor "touch-ups". But, when I hear the term I immediately think of the process of removing blemishes or skin imperfections from a formal portrait. And, although minor, these touch-ups are invariably in the direction of reality reduction. Which more-or-less puts us back in the same place we started. Which is why I reject proposing this as our new term for digitally massaging an image towards reality.
Digitally Adjusted? I'm not sure I've heard this term used much (or anywhere), so it probably doesn't drag contextual baggage or preconceived notions along with it. And it does imply some change to the image it refers to. And, at least to me, it doesn't necessarily imply a direction to that change with respect to the degree of reality the image represents. But, because of the negative connotations associated with digital "manipulation", I think I'm searching for a term that implies movement toward reality. So, forget this term.
Digitally Corrected? I like this term. It lets the reader know the image has been altered. And, the term "correction" implies that it's an adjustment that is fixing a flaw in the image-capturing process. Which, if we use the term carefully, is exactly what we're trying to say. Which means that we must either define the term or, at the very least, limit the boundaries of its use.
Digital "Correction" - Defining Some Boundaries of Use
I'm no master of the English language or at defining terms. And, in this case, I think a single sentence definition of the term "Digital Correction" (or "Digitally Corrected") would be lacking. So, instead I propose we use the following two guidelines to define the boundaries of the usage of the term:
1. The manipulation should produce an image MORE reflective of reality than the original un-edited image.
All cameras record a particular scene somewhat differently than our eye records it. The colour of an image may be more or less vibrant than the original scene. The white balance (warmness or coolness of the image) may differ from the original scene. The brightness - or "dynamic" - range of an image will undoubtedly be reduced relative to the original scene. And so on. I am comfortable with using the term "digital correction" to refer to any digital manipulation that pushes the output more towards the original scene than the un-edited image. I don't care if the camera produces the change (through the use of various "scene" modes or user-supplied tone curves). I don't care if it occurs during the RAW conversion stage. I don't care if the changes occur in Photoshop. The critical point is that the photographer is intervening in the process with the intent of moving the image more towards reality (as he or she perceived the original scene).
The thoughtful reader (which obviously includes you) is probably developing a critical question right now: "How can we measure, or judge, any image manipulation to assess if it conforms to this standard?" Unfortunately, we can't. The best judge of the faithfulness of the image to the original scene is the person who captured the image. So, we must rely on their honest disclosure (more on this later).
That same thoughtful reader (yep...you) is probably also thinking that we're moving into a little bit of a gray zone here in terms of just how much adjustment falls within the realm of digital "correction" before it becomes digital "manipulation". For instance, how much colour saturation can be added to an image before it crosses the line into digital "manipulation?" Good thought. Which leads me to the next guideline:
2. Digital corrections should be limited to changes in scope approximate to those available using traditional photographic technologies.
Huh? What the heck does that mean? Just this: traditional photographic technologies (i.e., film-based) carried variability with them. This variability is inherent in the process of photography and has never, in itself, produced accusations of image manipulation. And, I suggest that this range of variability is what should guide us in selecting how much digital adjustment is acceptable under the banner of digital "correction". For instance, decades ago nature photographers chose either an Ektachrome or a Kodachrome transparency film. These films differed fairly dramatically in colour saturation, contrast, and colour cast (equivalent to today's concept of white balance). Yet no one questioned the "reality" of the images produced by these films. A more modern example would be to compare the identical scene shot with Kodak's E100G (a film I REALLY like) and Fuji's Velvia (pick any flavour of it). The E100G slide would appear extremely UNDER saturated compared to the Velvia slide. Which is closer to reality? Well, no one really cares (or could possibly know). They may have a strong preference for one image or the other, but they don't get excited about which is more real.
So...where is this leading us? We can mimic any variation that is produced by different film choices in Photoshop. In the E100G vs. Velvia example above, I would argue that you're looking at a range of about -10 to +10 Photoshop units (in the Hue/Saturation dialog box). The same argument could be made for contrast (i.e., slight alterations to tone curves in Photoshop), colour/white balance, and more.
Expanding beyond film choice, keen photographers know that images differ significantly depending on lens used, aperture choice, et cetera. And, any photographer who is competent enough in Photoshop to successfully correct an image knows when they're applying a change that goes beyond what could have been accomplished through film choice, lens choice, aperture choice, and so on. And, that's really what it all comes down to - photographer honesty and full disclosure (more on this in a bit).
Perhaps a few examples will help illustrate how these guidelines could be applied:
Image Cropping: You view an image that you love, but then discover that the photographer used a high resolution camera and the image only represents 40% of the full image. Should you consider this digital manipulation or digital correction? More tightly framing an image when in the field is something that can be easily accomplished with a longer telephoto lens, i.e., by lens choice. So cropping a digital image is somewhat akin to choosing a longer lens. So...I would argue that this is completely legitimate digital correction. In fact, given the price of quality super-telephoto lenses (which can cost many thousands of dollars) it could be argued that allowing image "sub-sampling" (i.e., cropping) removes a barrier to entry to many "would-be-if-they-could-afford-to-be" wildlife photographers! It should be remembered that there is a real finite limit to how much cropping you can do to ANY digital image, regardless of how many pixels the latest and greatest cameras have.
Colour Adjustments: You view an image of a sunset that has brilliantly warm tones and deeper than deep oranges and reds. Later you find out the photographer captured the image as a RAW file and adjusted the white balance toward warmer tones during RAW conversion. And, to make matters worse, they used a LBA warming Photo Filter in Photoshop after RAW conversion. The nerve! Well, a traditional (film) photographer could have just as easily made these adjustments by choosing a warm, highly saturated film and placing a glass warming filter over their camera's lens. So...I would be completely comfortable labeling these digital adjustments as "corrections" and move on.
What about digital modifications to an image that are similar to what could be accomplished in traditional darkroom, such as dodging or burning? I see no reason not to keep these modifications under the banner of digital "correction" as long as the fulfill Guideline #1, i.e., as long as they produce an image more reflective of reality than the original un-edited image.
So, in summary, I believe strongly that in using "digital correction" within the boundaries of the guidelines above we have a useful standard. And we've given the photographers a POSITIVE thing to say about their digital adjustments. So there's absolutely no reason for the photographer to provide anything but full and honest disclosure - right?
The Critical Importance of Full and Honest Disclosure
There is no completely objective way to measure digital alterations to an image and draw a firm line between digital manipulation and digital correction. Ultimately, and unfortunately, it comes down to trusting what the photographer tells us. Even looking at the Edit History of an image file can be misleading - not only can this data be altered (or erased), but even if it's all there it can be very misleading. For instance, in my workflow I commonly bump up a parameter (like saturation) a TON but then cover that change with an image mask. I then reveal only a small percentage of that saturation through selective brushing away of the image mask. The resultant change may end up being only a percentage or two of saturation enhancement (on only a small portion of the image) but this would be almost impossible to interpret from examining the Edit History of an image.
Photographers - please be honest about what you've done to an image. Digitally correcting an image is NOT a crime. In fact, you should be proud that you have the skillset necessary to improve upon your camera's shortcomings. And leave your Edit Histories in your image meta data. If nothing else, at least they can teach us something!
February 6, 2007.
Other Commentaries
DISCLAIMER: The views and opinions expressed in this section are solely those of Brad Hill and do not necessarily represent those of any other group, organization, or corporation. In fact, it is possible they do not represent the views of anyone else on the planet - yet!