You are currently viewing this site as a Guest

Getting started with HDR Photography

HDR, or High Dynamic Range photography is something relatively new to the digital photography world. One limitation that the digital sensor currently suffers over film is that its dynamic range is less than that of film. By that I mean the full range of light values that can be recorded on a single shot, from the darkest to the brightest in that scene is lower for digital than film. Anything over or under that range is forever lost (those are your blinkies that you see when you chimp your pictures). For "prosumer" digital cameras today, this range is about eight stops of light. Film cameras are considered to have even more dynamic range, however some studies have shown that the gap is closing. However, there is still a need to attempt to put more apparent dynamic range in digital images in order to show detail in darker darks and brighter brights. HDR to the rescue! Click here to read more.

The basic idea behind HDR photography is that the photographer takes several pictures of the same scene (usually on a tripod), but each picture is exposed slightly higher or lower than the previous. For instance, I might want to take five pictures inside of a church, one at "normal" exposure (that is, what the meter in your camera averaged the entire scene to be at 18% gray), then take four more images, one at -2 stops, then -1, then +1, and then finally +2. The order does not matter as the software will figure out what needs to be done with the images in post processing. So what you now have is an image exposed at the middle, an image exposed for the darker parts of the scene (the +2), and an image exposed for the brighter parts of the scene (the -2). And you also have two in-between images (the -1 and +1). You then feed these images to some software that analyzes the images and takes the parts of each image that are exposed nearest the middle and sandwiches them all together to form a full image. You could of course do this with even more images, say seven, or nine. One important thing to remember is that the scene should be the same for each exposure, otherwise the software will become confused. That's why it's important to use a tripod. In fact, it's also a good idea to shoot on manual, and adjust your exposure using only your shutter speed so as to not affect the depth of field. Of course, you could also do HDR with fewer images (three). And there is a method that even allows you to do it with one image.


Doing HDR with one base image

This is a very technique for when you have already left the scene and you are doing your post processing and you realize that you have too much clipping in the brights and the darks. Or you may be photographing a scene with moving objects that won't allow you to get three or five unchanged photos at different exposure values. Then what you do is take your base image in your RAW processor and export one jpg at the normal exposure. Then export a jpg at -1, then export another at -2 (making sure to use different filenames, of course). Then export one at +1 and another at +2. Now we have five images, each "exposed" one stop apart. Supposedly the quality of this method isn't as good as if you had done it live, in the camera, but I have had great success with it myself.


Software

Now that you have your exposures, what do you do with them? I use a program called photomatrix by hdrsoft (http://hdrsoft.com/) This software is relatively easy to use and has quite a lot of features. You simply tell the software what images to use and the software processes them. It then presents you with a window in which you can tweak the overall effect and you can usually come up with a pretty nice picture. Another alternative to photomatrix is CS3 itself. Under the CS3 menu, File->Automate->Merge to HDR... you can also create HDR images. The dialogs there are pretty easy to follow, and you can usually end up with pretty similar images to that of photomatrix. Myself, I use photomatrix. It's the software I started with and I'll probably keep using it.


How about some samples?

Sure. Here is a series of pictures that I took inside of a church one day (I actually took five images, but I'm only showing three here). As you can see, no single image can capture the entire dynamic range of this scene. Either the stained glass windows are blown out, or the inside of the church is too dark. And taking the middle road just gives you an overall lousy picture, in my opinion. But in less than two minutes, and just using the default settings, I was able to produce the final image, which as you can see is pretty much perfect exposure throughout the entire picture, and is quite pleasing to look at. I would probably give this final image a very small contrast boost in photoshop and sharpening, and it would be done. Finally, I do have a sample image of what would have come out of CS3's HDR converter. This image has had some saturation adjustments already, and as you can see, it's still not quite as vivid as the photomatrix version. I know that with a few more minutes of tweaking, I could get this image to look pretty close to the photomatrix version. If all I had was CS3, it would be a viable option.