HDR imaging using layers





Dartmoor sunset

I like bright sunlight and high contrast - they often make for striking images. Sometimes though the result can be a disappointment because the shadow areas are completely black. Recovering the detail in shadows using curves is can be a solution. But if you have a really large area of deep shadow then this may not be satisfactory, and in that case you can use multiple exposures. I used this technique to create the image above. This is shot is combined from 3 frames taken on a tripod.

The first frame is exposed for the sky at 1/200s at ISO100, f16. Only the sun's disc is burnt out - since it's small in the frame this is not an issue. The sky looks nice overall with a well saturated blue. The other two thirds of the frame is so dark that I was never going to get much detail out of that.

The second frame is exposed for 1/50s ( two stops more ). The foreground looks OK now but the rocks on the left are still very dark. The third frame is 1/15s giving me a total 4 stops difference across the frames. I can see the details in the rocks but I have a lot of flare. At the time I thought that was unavoidable, but with hindsight I could have shielded the lens from the sun with my hand - I wasn't going to use the sky from that frame anyway.

To combine the frames I first overlayed the dark frame as a layer onto the middle one. I want a smooth transition so I erased the dark frame in the lower part of the image using a large-radius eraser and 7% opacity. Gradually the foreground came through.

The third image is so much brighter than the first that the rocks won't look right against that sky. They are still the shadow area of the image and we need to maintain that. So I adjusted the levels of the third image so that most of the tones are dark, but kept the highlights to reveal the details of the rocks. The highlights are set at values around 128.

Then I overlayed the combined image from the first two frames onto the third, made a selection around the rocks using the magic wand tool, feathered 3 pixels, and erased over the rocks to reveal the third layer. This is a sharp transition and needs some touching-up along the boundary. The whole process took a few hours but it would have been much easier if I didn't have so much flare to deal with.

I'm pleased with the result because I feel it captures the scene as a saw it on that Dartmoor hilltop. Our eyes can adjust for different levels of brightness but a camera can't do that in a single frame, and this can lead us either to shoot silhouettes or else not bother with the shot at all.

Some people may prefer to use the Photoshop 32-bit HDR function to combine the frames. I found a couple of problems with this. One is that small camera movements between frames cause a loss of sharpness over the whole picture ( I heard that CS3 is better in this respect ). This can be managed better using layers since you only have the transition areas to worry about. The second problem is that the 32-bit file needs to be converted to16-bit using one tone curve. So when you lift the shadows the whole image looks washed-out. I could then rework the 16-bit file using selections, but then I am getting back to where I was before using my layers method, and I have lost data in the conversion from 32-bit.


iainwest.com for photography