Learn good practices for adjusting images and the advantage of working in 16 bits.
When I shoot images I shoot in RAW and I do everything I can to maximize their quality in terms of exposure and sharpness, and when I process them in RAW conversion I do the best job I can to maximize tones and colors. (I use Lightroom, but this applies to any converter.) But I still find reason to go to Adobe Photoshop on many of my images, for anything from final tweaking to extensive creative effects. To do the best job there, some understanding of Levels and Curves is needed, as they are probably the most commonly-used adjustment tools.
Some background: the histogram, bit depth and color gamut
But first, let’s look at the histogram. It’s what you are affecting with Levels and Curves (and, in fact, most adjustments). It’s an exposure meter after the fact. And if you refer to it on the back of your camera while shooting, it’s an exposure meter on steroids. It shows the tonalities in an image and that’s what you need to capture correctly by making the best choices in camera settings such as shutter speed, ISO and aperture.
The left end of the histogram represents the darkest tones and the right end represents the lightest, as indicated by the graduated scale under the graph. For an 8 bit image, the histogram consists of 256 vertical bars, the height of each one representing the number of pixels that have that tonal value. When you bring an image into Photoshop from a RAW converter such as Lightroom, Aperture or Adobe Camera Raw, you face the decision to bring it in as either 8 bits or 16 bits. Before that point, all the adjustments you do in the RAW converter are acting on all the tonal levels the camera captured, which gives you the highest-quality adjustments you can make to an image. Do as much as you can at this stage.
In Adobe Camera Raw the 8 bit vs. 16 bit option is set at the bottom of the window, shown in the top half of the figure below, and clicking on the underlined options gives you the dialog shown in the bottom half:
In Lightroom you set the option in the External Editing tab of Preferences.
Working with 16 bit images is highly recommended. Digital cameras of the last few years commonly capture 14 bits, and older ones 12 bits. (Some very high-end cameras can deliver 16 bits.) But to store any number larger than 8 bits in a computer it has to go into a 16 bit “container.” A 16 bit image has 65,536 levels, but your images will actually have the number of tonal levels the camera captured: 4096 for 12 bits and 16,384 for 14 bits. You can see that saving all the tonal levels your camera captured leaves a lot more information in an image than if you reduce it to the 256 levels of an 8 bit image.
These are the number of tonal levels per channel, and there are three color channels, red, green and blue, so for a color image you will have a total for all three channels of over 16 million levels (or colors) for an 8 bit image, over 68 billion for a 12 bit capture and over 4 trillion for a 14 bit capture. (The 16 bit value would be over 281 trillion.) So even if you have an older camera that only captures 12 bits, you’re sacrificing quite a few tonal levels (or colors) if you reduce an image to 8 bits for further work in Photoshop. The difference may seem insignificant but when you start to stretch the histogram by lightening dark tones or by darkening skies that may only have the few tonal values of a slight gradient, you are more likely to see posterization of values in an 8 bit image than in a 16 bit one.
You can’t help what your camera can capture, but why reduce it even more by going to an 8 bit image in Photoshop? Despite the huge increase in tonal levels, the size of a 16 bit image file is only twice that of an 8 bit image.
As a side note, you can convert an 8 bit image to a 16 bit one, but you won’t add any tonal levels. When they’re thrown away in bringing the RAW file into Photoshop they’re gone. Sometimes you need to convert a 16 bit image to 8 bits, for instance to make a JPEG, or to use some of the filters in Photoshop. If you need to do this, be sure to flatten the image first. If you don’t, you have lost all the advantage of being in 16 bits in the first place.
Color gamut, the range of colors available, is a different issue, not related to bit depth. The color gamut you start with is that which your camera is capable of capturing. It will fit into in the RAW converter’s larger color space. But when you take an image into Photoshop or any other image editor you will have to squeeze the gamut to fit that of your chosen working space. The default in Photoshop is sRGB, which is the smallest gamut of the working spaces. The middle gamut, which I recommend, is AdobeRGB. This gamut can be achieved by some of the newest monitors and is close to what most printers can produce. The largest gamut is ProPhoto, but working with it requires knowing some pitfalls, which is beyond the scope of this article.
Back to the histogram
The histogram represents the range of tones in an image. A dark (low key) image will have the histogram weighted to the left end and a light (high key) image will have it weighted to the right end, as shown here.
A low-contrast image will have it bunched toward the center and a high contrast image will have significant areas toward the dark and light ends. An image with “blocked up” blacks (no detail in the darkest tones) will have the histogram pushed up against the left wall, and an image with blown-out highlights (overexposed with no detail) will be pushed up against the right wall, as shown below:
Unless you are advanced enough to know you have a good reason to do otherwise, you should always do adjustments as adjustment layers. Go to Layers > New Adjustment Layer or click the black/white circle at the bottom of the Layers Palette/Panel. Don’t use Image > Adjustments; that “bakes” the changes into the pixels of the image itself. (A few adjustments can’t be done as layers, such as Shadows/Highlights. Do those on a duplicate of the pixel layer so at least you have an out.)
The advantage of an adjustment layer is that it can always be tweaked without causing any image degradation because the adjustments aren’t actually made until the image is flattened for output. If you need to flatten an image, do so by giving it a different name by using Save As, so your master file will be kept with the layers intact. (Saving a PSD file as a JPEG of the same name won’t overwrite your master file; they are two separate files because they have different file extensions.) You might want to change the adjustments next year, when you learn great new things you can do with Photoshop and its many wonderful plug-ins, or when you finally get a good monitor and actually calibrate it and figure out how to be sure your computer is using the calibration. (Hint: It may not be! If you are using Windows Vista it isn’t. If you are using Windows 7 you need to make sure a certain setting that is not the default. See my tutorial: Getting Consistent Color From Monitor to Print.)
Basic Levels adjustments
To darken or lighten mid-tones, pull the middle slider right or left.
To darken the darkest tones, pull the left slider to the right, and to lighten the lightest tones, pull the right slider to the left, as shown circled in red below.
If you pull the sliders just up to the “information” in the histogram you will increase the contrast of the image. But if you pull the sliders underneath the histogram you are blocking up blacks or blowing out highlights. All the tones to the right of the slider in the figure below will be pushed to full white. Sometimes this is what you want to do, for instance to brighten up a dull sky, but not always.
You can get a visual picture of the areas that will be affected as the slider begins to come under the information. Hold Alt as you move the slider and the areas being blocked up or blown out in the image will show as colors on your image, indicating the color channel(s) involved. (This also works with the sliders in Lightroom and Adobe Camera Raw.) In some cases the affected areas may not be significant, or may be areas where you want the tones to go full black or white.
With whites this will cause areas that were near-white, that may contain important detail and smooth tonal gradients, to be pushed to pure white, and soft gradients in light tones may be replaced by an ugly look. (If you have specular highlights, such as the reflection of the sun on a chrome bumper, there is nothing you can or should try to do to avoid blown-out highlights.) With blacks, moving the left hand slider under the information in the histogram will cause areas that were dark gray, and that may contain important detail and smooth tonal gradients, to be pushed to black, and they may take on a gritty look with harsh tonal transitions.
If the moves are slight you may need to inspect the image at some magnification to see the affected areas. You will also need a properly calibrated and profiled monitor to see them accurately on the screen. If your monitor doesn’t show what the histogram indicates, trust the histogram. It is a very accurate report of what is actually in the image. Your monitor is, at best, a decent representation IF it is a good monitor, properly calibrated and profiled and your computer is actually using the profile.
By the way, if the image on your monitor gets noticeably darker or lighter as you tilt it up or down, or move your head up or down, you aren’t seeing an accurate representation of your image. If you can keep your line of sight at 90 degrees, you may be close but I wouldn’t want to count on it. (And this would be the case even if it is calibrated and profiled.) The vast majority of monitors are made for home and office use, not for the digital darkroom. Only a few monitors are suitable for photographers. See the article cited above.
Basic Curves adjustments
The Curves adjustment gives you a much greater degree of control than does Levels because you can make several points on the curve to affect different tonal values differently.
Pulling the center of the curve up and to the left lightens the mid-tones in an image, and down and to the right darkens them. This is similar to moving the middle slider in Levels but doesn’t give exactly the same result. Moving the center point of the curve gives a higher-contrast, snappier appearance, while making a similar move with Levels gives a lower-contrast more subdued appearance.
When you lighten an image in this way, using a single center point to affect primarily the mid-tones, you will be lightening the darks also, often too much. The advantage of Curves over Levels is that you can make a second point to hold down (or even darken further) the darkest tones after you lighten mid-tones. This is done by pulling down a point near the left (dark) end, to place it back on, or near, the original straight line, which represents the original tonalities.
There is a tonal scale under the Curve the same as the one under Levels, dark on the left and light on the right. (Unless it is a grayscale image, in which case the black to white gradient is reversed.) In the tonal areas where the slope of the curve is higher than the original diagonal, the contrast is increased, and where the slope is lower, contrast is decreased. In the Curve above, contrast is held as it was in the darkest one-eighth, increased in the mid-tones and decreased in the lightest one-third. Curves in this general shape are called S-curves.
More advanced Curves adjustments
In the case of a low-contrast image, you can bring either end (or both) of the Curve in as shown below for a stronger contrast effect than doing an S-curve.
And you can decrease contrast by either making a reverse S-curve or by moving either end (or both) as shown below.
The power of Curves over Levels is that you can locate multiple points on a Curve for more control over limited tonal areas, but be careful to keep the curve smooth. Going overboard can give an artificial look. In recent years other techniques have appeared for affecting tonalities, and especially for reducing contrast, which is difficult to do well with Levels or Curves. Among these is the Shadows/Highlights adjustment, high dynamic range (HDR) techniques, and an assortment of tone mapping techniques in plug-ins such as those from Nik and Topaz.
Color channel adjustments
Both Levels and Curves have color channel dropdowns, near the top of the dialog window. You can see them in the figures above just above the graph window, in a box that says RGB. Click the arrows to the right and you will see the three color channels and a composite RGB channel:
You can do individual adjustments for each color channel just as for the RGB channel, which will affect the colors of the image.