In this latest article in my 'Behind the Shot' series, I’ll talk all about one of my most popular shots, 'Nautilus'.

The image features an ice cave we visited in Vatnajökull, the largest glacier in Iceland (and in Europe), on a photography trip in December 2011, while scouting locations for my 'Land of Ice' workshops. It wasn’t easy getting there, and we had to use a modified jeep to traverse the snow. Even when close to the cave it was tricky to spot it, but luckily we had a wonderful guide who had previous knowledge of the whereabouts of this hidden wonder, and he brought us there quickly and safely.

I love shooting ice caves, there are just so many unique things about them. Each cave has its own different shape, texture and patterns. The light strikes each differently, and brings about a unique atmosphere. And this image is especially special in the sense that all these aspects were just perfect: the incredible spiral shape, the colors on the snow outside and the reflections on the ice. But an amazing location isn’t enough: we need to know how to compose, shoot and post-process such an image to get the best result possible.

Composition

The most important aspect here is, naturally, composition. I feel the composition is really what makes this image stand out, so let’s talk about the different compositional elements.

First and foremost, there’s the converging spiral starting on the bottom and going a full circle into itself. This element alone can make a shot, but there’s more. I see the opposing circular shapes as important, too: the dark hole in the ice, 'facing' left, balances the quasi-circular reflection on the left. The lines of the dark path on the bottom-right also lead the eye to the cave's entrance, contributing to the strong feeling of convergence.

Note that I had to choose my position extremely carefully to make sure the reflection was in this shape. Other shots from the same location have other reflections which I liked much less.

Setup

The image was taken with a Canon EOS 5D mark II, and a Samyang 14mm f/2.8 ultra wide lens. I love using the 14mm for ice cave shoots, since you can fit so much into one frame without needing a panorama. This is especially important when shooting HDRs since it’s quite tough building an HDR panorama, and I try to avoid it unless absolutely critical. The 14mm is also good for creating an extreme perspective, because it allows you to get really close to the foreground, while still having enough background in the frame.

It’s very important to mention that this shot is a manual HDR. This means I took several shots of the exact same composition, but at different shutter speeds, to cover all brightness levels in the scene. Ice cave often demand the use of HDR, since their insides are usually much darker than the outsides. This was the case here, and the bright blue sky was quite a few stops brighter than the cave walls. I thus had to take 3 different shots to cover the entire dynamic range. The aperture was set to F14, and the shutter speeds were 1/2sec, 1/6sec and 1/25sec. I used ISO 100 for maximum quality. In the images below you can see the untouched RAW files.

Post-Processing

On to post-processing. The dark art of manual HDR is far too complex to be explained in one short article, but I’ll try to briefly explain how I blended the image. 

To manually combine different exposures, we first have to stack them as layers on a Photoshop project. If we stack the images in descending order of brightness, we need to gradually get rid of the washed-out areas by selectively deleting the very brightest pixels of the brighter layers. Selection of the brightest pixels is done by ctrl-clicking the RGB channel and then multiplying the selection by itself (by ctrl-alt-shift clicking the same channel-mask) numerous times, until the selection is discriminatory enough for our needs.

The images are stacked as a photoshop project layers, in descending order of brightness. The histogram is that of the top layer, since it's all that is seen.

This is the state of things after ctrl-clicking the RGB channel one time, which selects each pixel according to its brightness. The marching ants show us the boundary of the set of pixels which are selected 50% or more. See the histogram now - it only contains the brighter pixels, since the histogram refers only to pixels which are 50% or more selected.

It might seem a bit complicated to understand, but in a nutshell, the selection is now stronger for brighter pixels.

After multiplying the selection by itself by repeatedly ctrl-alt-shift-clicking the RGB channel, the selection is discriminatory enough to distinguish the bright pixles from the dark ones. You'll understand it much more easily if you try for yourself!

We then use the eraser (partial opacity) to delete the selected bright pixels and reveal the darker pixels in the layer underneath it. Try it at home and you’ll see it’s not too complicated.

The stacked layers after a few passes of a low-opacity eraser. The brighter pixels have been affected to a much larger extent, since the selection was much stronger in those pixels. The cave walls remain unchanged, and the washed-out entrance is now somewhat darker and more dynamically-balanced.

After carefully deleting the washed-out pixels in the top layer, I merged the two top layers and did the same: I selected the brighter pixels by ctrl-clicking the channel mask, multiplied the selection several times and erased those pixels on the brighter layer to expose the darker ones underneath.

Merging the 2 upper layers is done by shift-clicking to select them both, then right-clicking on any of them and choosing 'merge layers'.
Again, I achieved a selection favoring the lighter pixels by ctrl-clicking the RGB channel, and then ctrl-alt-shift clicking it a bunch of times to multiply the selection by itself and strengthen the discrimination toward the bright pixels.
The image after another few passes with a low-opacity eraser. Now it looks much more balanced!

It’s critical to avoid over-blending. Doing so usually results in a very ‘HDRy’, unappealing look. The blending only serves to compensate for the huge global contrast in the image, and not to avoid local, desired contrast. In addition, see how I maintained a relatively high brightness level in the sky. Having the sky at the same brightness level as the inside of the cave would seem unnatural, which is highly undesired in my eyes.

After merging the 2 remaining layers, the image is almost ready. To finish up, I boosted the saturation a bit, and then converted the shot to the sRGB color-space (for internet use), applied some sharpening, and I was done. I hope you've enjoyed the outcome! 


Erez Marom is a professional nature photographer and photography guide based in Israel. Every January, Erez guides his Iceland winter photography workshops: 'Land of Ice' in the south and 'Winter Paradise' in the north and west. If you'd like to experience and shoot some of the most fascinating landscapes on earth with Erez as your photography guide, you're welcome to see the workshop webpages for details and participation, and view Erez' Iceland gallery.

You can follow Erez on his facebook page500px and deviantArt galleries.

More articles by Erez Marom: