Acting Locally

When I was getting started with the digital darkroom, roughly 15 years ago, there weren’t as many tools available as there are today.  Photoshop was the 900-pound gorilla and it was widely–if not quite universally–regarded as the only “serious” software package for photographic editing/enhancement.  Almost literally all of the tutorials and editing tips at the time were concocted and outlined with Photoshop in mind and so, of course, I purchased a copy of the Mother of All Editing Programs and jumped in with both feet. And I floundered around for about six months before I had an epiphany, of sorts; the rest, as they say, is history.  (The chronology of my digital darkroom experience is, at least arguably, an interesting one, but I’ll save it for another, later post–maybe.)  To this day, Photoshop is, hands down, the least intuitive piece of software that I’ve ever used.  When considering that statement keep in mind that I’ve used a number of advanced statistical packages going back deep into the DOS era.  Photoshop was significantly more indecipherable than any of them.

Otter Cliffs Sunrise, Acadia National Park, Maine

The process of using Photoshop, in the beginning, was so opaque that it’s difficult to convey.  Typically, when using software, the stumbling block that needs to be overcome is how to accomplish a specific goal that has already been identified.  How difficult this is tends to be a function of how complex the software is (i.e. how many things it’s designed to do) and how intuitive the interface is (among other things).  So, for instance, if I’m firing up a statistical package, I might want to carry out what is known as a discriminant analysis using a particular data set.  How do I go about carrying out this particular known task?  There’s a very specific way of doing so–I just have to figure out what it is (probably through some combination of checking through menu items, trial and error and accessing a Help file).  But postprocessing a photograph with Photoshop?  That’s an immeasurably more complex, fuzzier thing altogether.

Au Sable Point at Sunset, Pictured Rocks National Lakeshore, Michigan

The first problem–as compared to the statistical example outlined above–is simply determining what the task itself is.  How should I edit this photograph?  It’s not always so obvious, particularly when you’re new to the game.  Is there a color cast that you feel should be tweaked or removed entirely?  (By the way, if there is…it’s better removed in RAW conversion, assuming you’re shooting RAW, by means of a white balance adjustment.)  And, hey, the image looks pretty flat.  I guess it needs a saturation boost.  Or does it?  Perhaps a contrast adjustment would take care of the problem.  In short, you need to figure out what you want to do before you go about figuring out how to do it.

Coneflower Morning, Nachusa Grasslands Preserve, Illinois

Then there’s this little realization–there are multiple ways to carry out just about any kind of editing adjustment you care to apply in Photoshop.  There are  an innumerable number of techniques at your disposal, utilizing a variety of specific Photoshop tools and a dizzying accompaniment of blending modes, masks and plug-ins.  When I was first starting out, I began to create a Word document that listed different editing techniques as I ran across them, as a reference that I could consult.  I more or less stopped adding to the document after about five years, as I became sufficiently facile to remember/recognize virtually everything I felt I needed.  The document was well over 100 pages in length when I stopped updating it, in part because there were so many different approaches to accomplishing the same basic task.

Sunrise, Foothills Parkway, Tennessee

All of this–and other things, which I’m mercifully keeping to myself–in the early days of my digital darkroom experience meant that simply accomplishing ostensibly very basic actions with Photoshop were considered a triumph.  As a result, there was little recognition of what a blunt instrument Photoshop postprocessing could be.  The emphasis, naturally enough, was on carrying out global adjustments–making the entire image brighter or darker, for example, or adding contrast throughout.  But, in reality, it’s seldom necessary to carry out this sort of adjustment to a decent photograph.  In fact, it’s not only frequently unnecessary, it’s often a bad idea.  The vast, vast majority of helpful postprocessing work is accomplished with a far subtler, more deft, touch.  Truly enhancing adjustments are typically carried out in targeted fashion, via the use of layers, selections and masks.  This is what makes Photoshop such a potentially powerful tool for image enhancement (and the value of these tools is what made up the substance of the the aforementioned epiphany I had, six-odd months after first getting my feet wet with Photoshop).

Totem Pole, Monument Valley Navajo Tribal Park, Arizona

So let me illustrate the point with a broadly accessible example.  Consider the below image of a scene at Cannon Beach near sunset.  It’s essentially unoptimized and illustrates a common issue with many grand landscape scenes (and, not coincidentally, something that bedeviled all of the images accompanying this post prior to optimization)–the yawning chasm of a luminosity difference between the sky and most of the rest of the frame. The issue here is how to better balance the light and dark parts of the image without stripping it of its contrast.  A standard global contrast tweak–using a curves or levels adjustment–will strengthen the contrast, but will actually make the already too-dark tones even darker.  A reverse global contrast tweak using one of the standard techniques will provide better balance but will strip out the contrast. In essence, the goal is to accomplish something that’s difficult if not impossible to carry out with a global adjustment.  One way to accomplish the task is to create a selection of the sky and apply a contrast adjustment to that portion of the image in isolation.  Then, reverse the process and do the same thing with the other portion of the image.  (Frankly, there are a number of different techniques that can be used to successfully carry out the task.  The important point is that the one thing that all of these approaches have in common is that they involve local, rather than global, adjustments.) Making a contrast adjustment to the sky alone tightens things up–and makes colors appear deeper without actually touching saturation.  The non-sky portion of the image is almost a mirror image; an isolated contrast adjustment here lightens this part of the image without doing any damage to the sky.  The following version of the image shows the final version.  Note how contrast is enhanced, with each portion of the image getting what it needs in terms of its luminosity adjustment.

Sunset, Cannon Beach, Oregon

Obviously this is a somewhat exaggerated example, but it’s being used to clearly demonstrate a point–the power of local, rather than global, adjustments.  Often times, the appropriate postprocessing enhancements are made on a much smaller scale than what I’ve shown here.  Regardless, the postprocessing capacity of optimizing well-executed images is frequently realized by making changes below the global level.