HDR gets Old School
I’m adding another blog to my roll. Drew Gardner (UK Photographer) gave me an idea that is so utterly simple that it has me wondering "Why didn’t I think of that?" Simple. Brilliant. Alternative to HDR.
High dynamic range imaging (HDR) has been around for a while. Film and digital sensors are incapable of capturing detail in every tonal range of a scene. Imagine you are taking a photo of a sunrise over a lake high in the mountains. As the sun rises, it casts light in the sky and on the peaks of the mountain. The lake reflects some of that light. The foreground (the surrounding forest) is still dark. If you meter for the lit mountain peaks, no detail will be captured in the foreground and the trees will just be dark shadows. Alternatively, correctly metering on the dark trees will blow out the sky and the mountains, and the warm tones cast by the rising sun will be lost, defeating the entire purpose of waking up early for sunrise. Good example:
There are a few solutions. The first would be to use a gradient filter on the lens that darkens half of the frame. Spin the dark half to cover the sky and mountains, and the light half over the trees. You can get a couple stops of exposure out of the trees while keeping the sky dramatic. I don’t own such a filter, so the only option for me is to set my camera on a tripod and take several photos, one exposed for the mountains, the other for the trees and the third somewhere in between. I can open all three in Photoshop, merge them into an HDR image, and have the magical mathematical formulas even out the image so that everything is correctly exposed. The video below shows a before and after shot:
My main beef with HDR is that the images come out looking fake. It doesn’t replicate the image that the human eye processes at the original scene, but artificially boosts levels to give the next best representation. More importantly (for me), this can’t be used for action shots. Since multiple frames have to be captured, the scene can’t change in any way. The clouds can’t even move. Anyway, on to Drew’s idea.
He picks one frame that he likes from a sequence of action shots and creates 2 duplicates. To one, he boosts the exposure. To the other, he drops the exposure. The overexposed image is added as a layer below the original. Using the eraser tool with a suitable opacity setting, he erases dark areas on the original to reveal some of the more detailed exposure beneath. He repeats this with the underexposed image, erasing the hot spots on the picture. The results are believable. While my mind is telling me that some of his shots aren’t photographically possible, my eye doesn’t think "that looks fake." It really reminds me of my days in the darkroom, burning and dodging areas of the negative as I exposed it to the paper. Simple. Effective. Check out his post on the subject, with samples.
As I find time, I’m going to dig up some of my photos to try this on. Keep your eye out for the results.
