 | |  |
|  |
There are a lot of photographic topics that test anyone's comprehension, and a few of these are unfortunately kind of fundamental to successful digital photography. One, and perhaps the most important one, is the concept of dynamic range.
In simple terms, dynamic range is the ratio of light value between the lightest and darkest parts of a scene. Real life presents an enormous dynamic range, and because the human eye adjusts to the brightness of the thing it's looking at, human vision has a large dynamic range. Film and digital imaging processes have a much smaller dynamic range (about 5 f-stops for slide film), and this "shrinking" of the range can create a bunch of really ugly shots with highlights blown to pure white and shadows totally opaque. Most of the time, there's little to be done about the dynamic range limitations of your chosen camera. Sometimes there's more to be done than you think.
Output media, including CRTs and prints, also have dynamic range limitations. That means that even if you were to be able to capture the entire dynamic range of a scene, you couldn't reproduce it on a print. Keep this in mind for a later discussion!
When a scene has a big light-value difference between the bright and dark areas, almost every photographic device will end up either burning the light areas to featureless white, dumping the dark areas into featureless black, or a little of both. This problem is very likely to occur in winter photography, because snow creates a very light highlight area, and the angle of the sun can throw sections of the scene into a shadow that turns completely opaque. Nearly everyone who's taken a picture in winter has a nice collection of silhouettes; dark shapes on snow that show no detail at all.
Dealing with "problem dynamic range" starts with dividing the problem up into two dimensions. First, does the camera have the DR to capture the scene as it is? Second, does the output media have the DR to record the scene as captured?
Let's start with the camera side. Most digital cameras will record more DR than slide film, and many will record more than print film. A few, like the Fuji S3 and S5, are designed to record considerably more range than film. In this situation, your primary question will be how to expose the shot correctly so that you have a chance of capturing what you "see".
If there's a particular thing you want to capture to the exclusion of everything else, just pick a more-or-less neutral area on it and spot meter on that. Let shadows go dark, highlights explode, etc. Most of the time you won't like the result much, though, but the digital camera may be able to help you out.
The "approved" way to capture a lot of dynamic range is with multiple exposures, usually two. One exposure is correct for the highlights and will leave the shadow areas very dark, and the other will be exposed for the shadows and leave the highlights burned out. You then blend the two images together.
What? Blend two images? There are three steps involved:
- Paste the two images into a common digital file as two different layers, with the lighter one (the one where you've exposed for the dark shadow areas) on the top.
- Adjust the image positions as necessary to get them perfectly aligned.
- Combine the images to take the light areas from the lower layer and the dark areas from the higher one.
Of course, these steps might not seem all that easy. The first one is pretty much mechanical, but numbers two and three could certainly be rather involved. Getting the two image layers in alignment is a delicate task, and it's impossible if elements of the picture moved between the two exposures. Combining the two layers might also seem a bit complicated. Fortunately there are some tools that can help in the combining step, but first let's look at how you might use this form of dynamic range expansion manually.
Suppose you have a picture of an animal like a moose, standing in snow. Take the shot under normal matrix metering conditions and you have moose silhouette. Spot meter on the moose and you've got a reasonably exposed moose in what will look like cotton batting. So spot meter on the moose for one shot, on the snow for the other (do it fast so the moose doesn't move) or bracket the exposure by at least 2/3rds of a stop each way, and start to work.
The first step is to identify the image in which the background exposure is roughly correct. Ignoring the moose for the moment, do your color balancing and other adjustments to get the best image of the background that you can get. The moose will be a blob. Now, find the exposure in which the moose is best and ignore the rest of the scene to optimize your image of the moose. Generally, for both these images, you'll want to be a tad underexposed to get good color saturation. Now, copy the background-scene image and paste it over the moose image as a layer. Set the opacity of this layer very low so you can just see the moose outline, and line up the moose with the layer below. You'll have to zoom in to get this exact, and take your time doing it. When you're done, restore the top layer to full opacity; you won't see your lower moose exposure at all.
When you have the layers aligned, zoom in on the moose itself. Select the eraser tool of Photoshop, with a brush that has hard edges and that is relatively small compared to the image. Set the opacity of the eraser tool low, perhaps to 20%, and carefully start to remove the upper (dark) layer to reveal the properly exposed moose underneath. When you have the proper amount removed, you'll have a good moose on a good background! Now just flatten the image, tweak the adjustments for the final exposure and color settings, sharpen, and you're done.
This approach will work pretty well where you have a single area that's much darker than the rest of the shot, like our moose. If you have a picture with a scattering of light and dark areas throughout, you'll be erasing forever to get a clean image. For that situation, you have two options.
First, if you use Photoshop's select by color range command (in the SELECT menu), and use the pull-down to select shadow areas. Feather the result and do your erase. This sometimes works really well, but you'll have to select the feathering really carefully to avoid developing a halo around the dark areas. It's sometimes helpful to fiddle with modify-expand, modify-contract, and feather to get the selection just right. Again, it can be a lot of work.
For those who (like me) are work-adverse, the new Shadow/Highlight adjustment in Photoshop, and the equivalent sliders in the Camera Raw conversion, take a lot of the work out of most dynamic range problems. This tool flattens the image and then darkens the highlights and shadows individually, with control over mid-tone contrast as well. But this works with only one image, and so if you've totally lost highlight or shadow detail on that single image there's nothing you can do.
A similar approach is to simply create two raw conversions of the same image, setting one to expose for highlights and the other for shadows. These can then be combined as described above. This works a little better than Highlight/Shadow because you typically carry more exposure data in the RAW file, but it is no improvement over what you can achieve with the sliders during the raw conversion.
For those who shoot JPG or who don't want to fiddle a lot during conversion, the "Shadow/Highlights" control is the best option available. The key is to use it correctly. To start with, you'll need to insure that the image is converted with a tad darker tone than you'd normally want; underexposed by a half stop or so. Then open the Shadow/Highlights control. You'll see the image is "flattened" completely in dynamic range, so everything's the same degree of lightness. It will look awful, make no mistake, but it won't stay that way. Now you bring the Shadows slider back to the left and the shadows will darken. Pushing the Highlights shadow right will darken them. The Tonal Width controls will set how bright or dark the area of the image must be for the other sliders to impact them. You can also set mid-tone contrast at the bottom of the window. It will take some fiddling, but with this control you can fix a lot of blown highlights or opaque shadows.
Where you need more local correction, you can work wonders with Dodge and Burn tools. The dodge tool will open up shadows, and the burn will darken bright areas. The burn also has the apparent effect of increasing color intensity (saturation) so it actually has the effect of burning off haze and fog when used carefully. These tools do a lot of changing, so you want to proceed carefully.
But there's a catch to this stuff. What you are doing is compressing the dynamic range of an image. Since the output media can't record the full range, capturing it doesn't do anything other than let you map the range of the scene to the smaller range of your output media in a controlled way. The result can be good, or it can be "muddy" because the snap of higher contrast is lost.
Fortunately, you can diddle the exposure, contrast, brightness, saturation, and other stuff in Photoshop to make the compressed range look as good as possible. You can also (somewhat) control the mapping process by selecting what you use as your "highlight" and "shadow" exposures.
Where the scene has only slight problems, you might also want to consider the excellent highlight/shadows adjustment tool that's provided in Photoshop CS. This tool lets you adjust the dynamic range in a kind of indirect way, by controlling highlight-shadow balance in the scene. It won't give you all the flexibility of the dual-image-mix approach, but it's a lot faster and easier!
|
|
|
