How Was This Picture Made #04: The Answer

This article is the answer to my “How was this picture made?” post from a couple weeks ago. First, an apology – I intended to have this article published several days ago, but my winter classes already have taken their toll on my spare time (and sleep). That said, I hope the answer is worth the wait!

In all, people did quite a good job of guessing how this image was taken. No one got every part of my question correct, but many of you were close. I especially would like to congratulate our reader Rob, who somehow managed to find the Google Maps location that I took the image. Well done! Others of you got four out of the five questions correct, so I must extend another congratulations.

Without further ado, then, here are the answers to my five questions:

Is this a single photograph?

Yes, it is a single image. Many people took the shadows as evidence that it was a composite – which is very clever thinking – but the image is just a single frame. Other readers suggested that I used an off-camera flash to illuminate the subject, which also is not the case. What happened instead (as I will mention in the next answer) is that I lifted the shadows fairly dramatically on the horse; if you do this, something is going to light your subject. In this particular case, due to the off-camera pattern of clouds, the light from the left-hand side of the sky was brightest; that’s why the horse appears lit from the left, even though it should be a silhouette. As much as it looks like a trick, it wasn’t!

A couple of readers wisely pointed out that the fringes of the horse’s legs still look backlit, which shows that the main source of light is indeed from behind. An astute Photoshopper certainly could have added these highlights for a more realistic effect, but I am far from an expert at compositing! Indeed, no parts of this image were cloned or removed, aside from a couple of dust specks and distracting bits of snow.

Also, the horse does indeed cast a shadow; the light is just so low that it is difficult to see. Look, though, at the grass closest to my camera. You can see a few backlit sparkles on either side of the horse, but none directly in the center of the image. Again, a professional photo manipulator easily could simulate this effect, but it was entirely natural in this case.

What “unusual” post-processing did I do for this image?

This is the fun part. To process this image as best as possible, I needed to brighten it significantly. For comparison, here is the original RAW file (with Lightroom’s “camera standard” rendering):

Quite a big difference! If the photograph hadn’t been so dark, though, the sun would have lost all of its detail – underexposing was my only option. The tricky part was to make the image look natural in post-processing. (I was helped by the high dynamic range of my Nikon D800e, although I am convinced that most any modern camera would be able to make something of this shot.)

Although I experimented for quite a while on this RAW photograph, I wasn’t getting anywhere. My edits for the land were ruining the sky, and vice versa. Even the local adjustments in Lightroom weren’t doing the job, since they are fairly limited in functionality (for instance, the Hue/Saturation/Lightness sliders are conspicuously absent from the local adjustment options). Since this image needed opposite post-processing adjustments for almost every slider, my best option was to combine two separate edits in Photoshop.

Here is my initial Lightroom edit for the ground:

More than anything, I increased the overall brightness of the image, then vastly decreased its highlights. The result is far better than the RAW file, but I don’t like the washed-out purple color of the sky. As much time as I spent with local adjustments, I couldn’t get the sky to be the nice blue color that I wanted. This was accomplished in a few seconds with a virtual copy:

With this separate copy, though, the colors in the ground are far too cool (and the clarity is too high, although that can be fixed by a local adjustment). To reach the final image, I brought both copies into Photoshop as layers, and I used a simple gradient mask to blend the two shots together. I centered the gradient along the horizon, making for the most natural transition. If I had wanted, I could have been more elaborate with my blending; in this case, there simply was no need for it.

Here is the result of that blend:

It’s starting to look much better! I then exported the resulting image back to Lightroom, where I made a few final edits. I cloned out a few spots of snow, and I minimized the small amount of lens flare that had appeared. More than anything, I just changed around some local contrast and saturation, primarily trying to draw attention to my subject. This part was no different than for most of my images; the hard work is already done. Perhaps the most dramatic adjustment I made was to remove the purple cast of the right-hand mountain; I simply moved the “tint” slider to the left with a local brush. A screenshot shows some of my edits:

Perhaps a more skilled photographer could do the same edits without Photoshop; I hold no delusions that my method is the only way to process such an image. Personally, though, this was the quickest way to reach my end goal. I rarely blend multiple copies of an image in post-processing, but it can be an invaluable technique when it is needed.

Did I use any filters while I was in the field?

This question tripped up most readers, although that certainly was not my intent. I did use a filter for this photograph – specifically, a soft-edged graduated filter with a -2.5 stop rating. This helped me keep the sky and water manageable, as you can see in the original photograph, although it may not have been essential.

Which settings did I use for the photo? In particular, what was my exposure compensation (using matrix metering)?

I asked this question because I wanted to point out a particular feature of the matrix meter in most cameras, and one which may not be obvious at first. This photograph was vastly underexposed – an exposure compensation of -2 and 2/3. (I was in aperture-priority mode.) Most commenters in the original article were correct, but others said just the opposite: they believed I had a significant positive exposure compensation. Why is this?

If you have ever shot into the sun, or another bright light source, you may realize that your meter underexposes the image heavily – it tries too hard to preserve highlight detail, and you may need a high exposure compensation as a result (assuming that you don’t blow out any important highlights). In this photograph, though, the sun was not yet bright enough for my camera’s meter to care; as far as I could tell, it didn’t even take the sun into consideration.

This is an interesting scenario if you are shooting into a bright light, particularly one which changes in intensity. At first, the meter may not yet recognize the bright object at all; once it does, your reading may change suddenly and significantly. If you want definition in your highlights – as I did for this photograph – you will need a vastly shorter exposure than the camera recommends.

Aside from exposure compensation, my other settings depended entirely upon the brightness of the scene (and the wide depth of field that I needed): f/11, 1/25 second, and ISO 100. My D800e was on a tripod with the Sigma 50mm f/1.4 A lens; as one commenter mentioned, the lens quality is the reason that this image has minimal flare (although I corrected it slightly in Lightroom). I took several photographs to ensure that the horse would be sharp at 1/25 second, but I didn’t need to worry – that horse was a perfect subject!

And just for fun – was this photograph taken at sunrise or sunset?

I was very impressed that most people guessed that the photo was taken at sunrise – I didn’t think that this question would be so easy! Also, I will give credit to the readers who said that this was the midnight sun, and thus neither sunset nor sunrise. Indeed, I took the photo right around Iceland’s summer solstice, when the sun barely dips below the horizon. However, Iceland is mostly below the Arctic circle, and at my elevation the sun actually did set – though only for an hour or so. Technically, yes, this image is from sunrise; at the same time, it was taken at two in the morning!

I hope that this exercise was fun and informative – I certainly enjoyed writing it. If you have any questions about my technique, please do not hesitate to ask in the comments.