• 
• 
• 

Tuesday, January 5, 2010

A Brief Guide To HDR Photography

High-dynamic-range photography has caught on like wildfire, and now with DSLRs capable of creating HDR images automatically, it’s time to revisit this hot trend

This Article Features Photo Zoom

If you go to the beach at sunset and see a brilliant sunset, what do you see? You clearly see the detail in the sand, in the sky and can enjoy the sunset. You’ll never hear someone ranting that there are too many silhouettes and that it’s hard to see. However, if you try to take a photo, the scene will be mostly silhouettes. Depending on the metering mode of your camera, it will probably meter for the brightest object in the scene, the sun. Everything else will be underexposed. Many will be happy with this because it darkens the clouds and creates an aesthetically pleasing scene. The problem is that you aren’t capturing the scene as you remember it. You’re capturing it as aesthetically pleasing as possible with the limitations of your camera.

This is all because of something called dynamic range. In photography, dynamic range is the luminance values from darkest to brightest. For a camera, this is a subset of the scene’s dynamic range that can be captured without blowing the highlights or reducing shadow detail to noise. Let’s talk numbers. A contrast ratio is worked out very easily. It’s 2 to the power of the bit-depth. So an 8-bit image is 28, which gives a contrast ratio of 256:1. These numbers mean that, as soon as you take a photo with any camera, you instantly lose an incredible amount of detail from that scene.

For years, there have been ways for photographers to get around this. We redirect light or create our own to compensate. Another way is to use HDR.

A Solution: HDR—High Dynamic Range
The concept is simple. Take three or more bracketed photos, merge them together to create an HDR image and then tone-map them to recompress the data into a useable 8- or 16-bit image. Going back to the numbers, a typical sunny day has a contrast ratio of 100,000:1. A 16-bit TIFF file has a contrast ratio of 65,536:1. So quite a bit of data is lost from that sunny day. A 32-bit HDR image has a contrast ratio of 4,294,967,296:1. It’s clear that a 32-bit HDR image can hold an incredible amount of data. Numbers aside, what this means is that you can bring back detail in the shadows and highlights. The highlight detail isn’t clipped, causing overexposed areas, and the shadow detail isn’t turned into noise. This is all a bit complex-sounding, so I’ll use an example. If you were to use your DSLR inside your living room, generally, the outside would be blown out, losing detail in the highlights. Using an HDR technique, you’d take three bracketed photos instead and merge them. Once merged, you’d compress that image back down to something you can see and you’d hopefully have no blown-out highlights. You’d be able to see inside and outside perfectly, something along the lines of how you really saw that scene.