If you’re like me, you’ve probably seen images like the one above before and chalked it up to a neat photoshop effect. However, a lot more than that separates Trey Ratcliff’s take on the Capitol Columns at the National Arboretum from my attempts. The image was produced through a process called High Dynamic Range Imaging. Originally created as a technique to create realistic objects in computer graphics, the technique essentially creates images containing more visual information than a standard, single-exposure photo. The result contains more detail in the dark and bright areas, more like how we actually see. Wikipedia explains how it works this way:
Information stored in high dynamic range images usually corresponds to the physical values of luminance or radiance that can be observed in the real world. This is different from traditional digital images, which represent colors that should appear on a monitor or a paper print. Therefore, HDR image formats are often called “scene-referred”, in contrast to traditional digital images, which are “device-referred” or “output-referred”.
This article explains some of the concepts in more detail. HDR photographs can be taken from real life by digitally combining several photos taken at different exposure settings, and the resulting photo will contain details from the lightest and darkest portions of the pictures as well as more color. The technique seems to becoming increasingly popular, and a recent tutorial published by Popular Science explains the software needed to create the eye-popping images is freely available. I also noticed a couple HDR photos have popped up as DCist photos of the day this spring.
I stumbled across the technique looking for photos of Ballston, of all places. A local resident and Flickr user sduffy had uploaded the photo to the right and several other particularly well done images.
District resident Jon Ross has created several images of D.C., including this view up 15th Street:
Another D.C. Flickr user experimenting with HDR is sunyata, who has created a set of some of his favorites. His style is a bit more subtle, as seen in this version of the fountain at Meridian Hill Park:
What’s your favorite HDR image?
> Popular Science: High-Dynamic-Range Photography: A Guide
> “The Future of Digital Imaging – High Dynamic Range Photography”
> Flickr HDR pool
Nice post, Rob! I too stumbled upon HDR photography on Flickr a while back. I even gave it a try myself:
this is fascinating stuff, thanks.
what kind of camera do you have?
My initial reaction was one of wonder, but as its popularity has grown, I find it is overdone; people make HDR images when single-exposure shots would be more compelling. In the above examples, the waterfall and aerial scenes are the ones where the HDR method gives you a better sense of what it is like being there.
As film/digital is both a tool for documentation and for artistic representation (and there’s some artistic choice no matter what), it’s tough to criticize it, but from my primary interest in architectural documentation (for lectures on extant architecture, for example, I raid flickr almost exclusively), I find people enthralled with the capability and splurging in that regard, rather than restraining themselves because of a guiding idea of appropriate use. (Of course we all realize that that’s a relative term).
I like to use HDR in a lighter, more subtle way that simply corrects the deficiencies of the camera sensor and produces a natural effect. For instance, here are some HDR shots of the DC Katrina Cottage: http://cnudc.org/Images/KCDC.jpg. I rarely push the HDR parameters to get intensely surreal effects like Trey Ratcliff’s, although it can be a lot of fun.
Here are some poignant photos by Ryan Southen of the David Whitney building in Detroit, which has been vacant since 1999. HDR shots are listed with auto exposure to show the comparison.
1a. Lobby & elevators – auto exposure
1b. Lobby & elevators – HDR
2a. Whitney lobby – auto exposure
2b. Whitney lobby – HDR
3. A different shot of the Whitney lobby – HDR
Comments are closed.