Posted by: kerryl29 | October 10, 2011

Tutorial: One Way to Solve an Exposure Problem

In my writings on this blog, I rarely deal in depth with technical problems. This isn’t because such matters are unimportant; rather, it’s because, frankly, there are other topics I prefer to write about. But I recently had some questions–from people in the real world, no less–about dealing with a particular exposure conundrum and was encouraged to share my solution here. So, I’m “playing against type,” so to speak.

If you find yourself shooting waterfalls and/or cascades that are located under substantial forest canopies, you’ve got yourself an embedded issue with exposure, even if you’re shooting in even light–overcast or deep shade–conditions. The problem can be summed up in two words: dynamic range. Consider the following image, from Roaring Fork in the Smokies:

Roaring Fork, Great Smoky Mountains National Park, Tennessee SINGLE EXPOSURE, MANUALLY SPOT METERED

No post processing was done on the above image, except to convert the RAW file to a usable format.  The image was exposed by spot metering on the brightest highlight–the lowest cascade–and opening up roughly 1 1/2 stops.  A particular (read: slow) shutter speed was desired to render the water in a particular manner, so the ISO was lowered to a minimum.  A polarizing filter was used and a relatively small aperture was utilized.  The key was to utilize as much light as possible without blowing out the highlights.   The final shooting data:  ISO 200; f/16; 2/5 sec.

No problem, right?  Well, there is a problem:  the image is, on the whole too dark.  Depending on your monitor brightness settings it may not look all that bad but, trust me, it’s way too dark.  I knew when I clicked the shutter that it was too dark.  And in print, the problem is overwhelming.

You’re probably wondering why, if I knew it was too dark when I took the picture, I didn’t adjust the exposure to brighten things up.  That’s where the pesky dynamic range problem comes in.  The image is, approximately, one full stop underexposed in the mid-tones.  Take a look at the histogram of the above image:

Luminosity Histogram, base exposure

As you can see, the shadows aren’t blocked up and the highlights aren’t blown–that’s great.  But look at where the bulk of the tones are located–they’re way off on the left-hand side of the histogram.  The shadows and the mid-tones are nearly indistinguishable from one another.   This is a problem.

Through the power of RAW conversion, here’s what adding one additional stop–adding the equivalent of one EV–looks like:

Roaring Fork, +1 Stop Exposure

As you can see, the mid-tones look a lot better.  But check out the highlights–yup, as implied earlier, they’re blown to smithereens.  In truth, even a 1/2 stop of added exposure results in fully blown highlights and even a 1/3 stop increase would have blown the blue channel.  The exposure I selected in the original image above–the “base exposure”–was the brightest that could be achieved without blowing the highlights.

This problem is a fairly common one, at least if you shoot in the kinds of places I frequently find myself.  So what to do?  There are several ways to go about trying to deal with the problem.  The most obvious is to simply selectively dodge what should be the mid-tone areas of the original image.  There is detail there, after all, to be teased out.  The problems resulting from this approach are multi-faceted: creating a selection is problematic and potentially very time-consuming, and producing a sufficient dodging effect means all kinds of potential issues with feathering and revealing noise.  Dealing with all of this can be done, but doing it well is extremely difficult and will take a lot of time.

A second option is to take multiple exposures–three, let’s say: one optimizing the highlights, one optimizing the shadows and a third optimizing the mid-tones.  Those exposures can then be blended manually in post-processing or can be combined automatically through some sort of high dynamic range (HDR) program.  I’ve done both over the years and each can be a viable option.  But both are potentially plagued with problems, the biggest of which is subject movement.  In the image being used here for display purposes, the subject itself–the water–is moving.  This can be a real problem, particularly for a manual blend.  Most HDR programs have a feature designed to deal with moving water.  But if any other part of the subject moves–foliage, for instance (the rhododendron leaves in the sample image)–both approaches may be unusable.   Unless it’s dead calm this is a problem because even the best-initiated HDR sequence takes time to execute.  Besides, a fairly slow shutter speed–to account for the desired effect of the water’s appearance–is required here, which is going to further lengthen the process of taking multiple exposures.

There is a third option that I began to experiment with some years ago and it is, I believe, the best compromise–creating multiple exposures from the same RAW file and blending them via  some HDR-like technique.  This approach essentially eliminates all of the problems outlined above.  Subject movement from multiple exposures is a non-issue since the multiple exposures are created from a single image.  The process of creating the combined image is a simple one because it’s entirely automated.  Properly executed, an HDR (or HDR-like) approach will take the best parts of each exposure and seamlessly blend them.  The resulting file can then be optimized in Photoshop (or some other image-editing program).

I use Photomatix for my HDR work.  I’ve played with other programs and they all have their advantages and disadvantages, but I’ve been using Photomatix for roughly five years now and have developed a workflow that yields excellent results for the realistic look I’m trying to achieve that I can’t easily duplicate with any other software, so I’ve stuck with it.

The specific approach I use for producing blended files with Photomatix depends on the dynamic range of the image in question.  For the type of image we’re working with here, I’m really not trying to extend the dynamic range in a technical sense.  What I’m trying to do is to neutralize, for lack of a better term, the distribution of tonal values.  Strictly speaking, I want the mid-tones to appear as mid-tones, near the center of the luminosity distribution, not have them left-shifted as they are in the histogram above.

To create the kind of exposure I’m looking for in this case, I ultimately produced four exposures for the purposes of use with Photomatix.  One is the baseline exposure you saw at the top of this piece.  A second was the “+1” exposure above.  The others are a +2 exposure (to better optimize the shadow detail on the bottom half of the rocks) and a -1 exposure (to better optimize the highlights).  This process can be handled using any RAW conversion software.  I saved all four files as 16-bit TIFFs and selected them in Photomatix to be combined in the blended exposure.

In this case, given what I’m trying to do, I found that a Fusion Blend (highlights and shadows) approach is the best for the subject matter.  (For other subject matter one or more of the tonemapping options may be best.)  Here’s what an unedited version of the blended file looks like:

Roaring Fork Fusion Blend Highlights & Shadows, unedited

Contrast that with the luminosity of the original image.  The overall tone is similar to the +1 exposure, but this time, Photomatix has ensured that the highlights aren’t blown.  Here’s the histogram for the above blended image:

Lumiinsoity Histogram, blended exposure (unedited)

Again, contrast this histogram with that of the original image and note the dramatic difference.

Luminosity Histogram, base exposure

Below is the final, edited version of the file (I’ve added a bit of contrast curve in Photoshop) and a slight saturation mask boost:

Roaring Fork, Great Smoky Mountains National Park, Tennessee (final version)

And here’s the histogram for the final, edited image:

Luminosity Histogram, Final (edited) image

Now that’s what we’re looking for!  This final image is dominated by its left-center mid-tones, just as the real life scene is.

If you have an image that’s like the one I was working on here–detail throughout, but intentionally underexposed to avoid blowing the highlights–give this approach a try.  And, by all means, feel free to inquire if you have any questions.



  1. Hi Kerry,

    The final version looks natural. The interplay between the water and the moss covered boulders really comes through. The boulders fascinate me. In fact, I think the water plays a supporting role in this image.

    At first I wondered why you didn’t simply work with the curves but then that probably wouldn’t have worked with color. I’m so used to thinking in B&W and never worry about color shifts or color accuracy. I can’t imagine this image in B&W. You got those mid-tone greens to come alive and that really makes this image work.

    It’s refreshing to see someone use HDR to achieve a natural result.


    • Thanks, John. Just to clarify something…

      The approach I used here is, technically, not HDR. In fact, it’s not entirely a technical distinction, since HDR requires an expansion in dynamic range beyond what can be achieved with a single exposure. What I did here is simply optimize a single exposure’s range; it’s effectively a fairly subtle redistribution of tonal ranges from a single exposure. I guess I would refer to this as “HDR-like,” because I did use an HDR program (Photomatix) to achieve my goal…though I didn’t use one of the tonemapping options, in this instance, and I didn’t (as I mentioned above) rely on multiple exposures to arrive at the end result.

      When using exposure blending–manual or automated–or full-blown HDR itself, my goal is always to achieve a “photo-realistic” result. The rather surrealistic look that is the end-goal of many who use HDR isn’t my cup of tea at all. I resist any temptation, however, to condemn it outright as I occasionally see others do. It is, in my judgment a valid form of art; it’s simply not one I care for.

  2. Hi Kerry!

    I am in the same corner as you with the HDR outlook in so far as I do not go for the fake or cartoon looks. I am also not saying that there is anything wrong with it if that is what you do. I just don’t go for it myself. I have only just started using this single image to make my extra exposures technique. I think you achieved great results up there and it gives me the confidence to maybe cut back on my bracketing. Good article to shed some light on this as a very good technique.

  3. Thanks, Terry. Much appreciated.

  4. […] long after posting a blog entry detailing an editing technique I’ve developed for dealing with certain types of images, it occurred to me that this approach […]

  5. I appreciate this detailed and literate explanation of your technique for optimizing an image.

    Steve Schwartzman

    • My pleasure. If I can be of any help with the specifics of implementing the technique, please don’t hesitate to drop me a line.

  6. Thank you for visiting my blog. You have many wonderful images to view here. You make me want to get out more! Happy travels 🙂

  7. […] Most Helpful – It’s debatable, but based on some feedback, I’ll lean in favor of One Way to Solve an Exposure Problem. […]

  8. This is very helpful! Thank you for writing this tutorial.

    • My pleasure!

  9. […] it’s own way, this is a less extreme version of the kind of image that I’ve blogged about before:  lots and lots of dynamic range, with the need to underexpose the scene as a whole to hold detail […]

  10. […] render more detail. There are a number of ways of going about this including (but not limited to) a technique that I outlined on my blog a few years ago that I’ve found to be very effective.   Whichever approach you choose, simply […]

  11. […] render more detail. There are a number of ways of going about this including (but not limited to) a technique that I outlined a few years ago that I’ve found to be very effective.  Whichever approach you choose, simply be […]

  12. […] than the craft.  It’s not that there haven’t been craft-related posts–there certainly have.  But I haven’t spent a great deal of time discussing technique.  That’s not to […]

  13. […] for a classic dynamic range problem.  Fortunately, my camera‘s sensor could handle it and a postprocessing technique that I’ve refined over the years provided significant downstream […]

  14. […] sequence of shots.  The question was, would my camera sensor’s dynamic range, along with my customized post processing technique for such situations, hold up?  I decided to find out.  I positioned my tripod very, very low to […]

Please feel free to comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.


%d bloggers like this: