Digital Black and White Mistakes
A good black and white image can create a strong impression but can be a challenge to produce. Here are a few potential digital black and white mistakes you should be aware of.
Long ago, in a world before color images were even possible, there was black and white film, and thus there were black and white images. This was the medium photographers worked in since that's all there was. Over time, photographers developed ways of altering the rendering of their black and white images by using colored filters over their lenses. A red filter would strongly pass red light and reject green. A green filter naturally did the reverse. In this way, various hues could be rendered as desired from bright to dark merely by the selection of an optimal colored filter. Landscape photographers used red filters to darken skies and foliage and increase contrast. Portrait photographers loved the way orange filters smoothed out the appearance of freckles and other skin imperfections. You get the idea.
But times have changed, and few of us still shoot with film, let alone black and white film. Digital rules the day. Many photography text books and schools though still teach the use of colored filters for black and white, intent on covering such foundational techniques, and some digital photographers turn to the use of such filters in their quest to "do things right." Don't fall for it.
Every photosite in a digital camera already has a colored filter over it — red, green or blue — in a patterned array known as a Bayer mosaic. The individual sensors underneath each of these Bayer filters are all exactly the same. Each one is capable only of measuring the amount of light that falls on it during the interval over which the shutter is open. They can't directly sense color any more than the black and white film could years ago. By covering each one with a single Bayer filter in red, green or blue, that photosite sensor selectively records light of that predominant color, rejecting other wavelengths, altering its sensitivity in much the same way that colored filters do with black and white film.
But were you to place your own colored filter over the camera lens in an attempt to emulate traditional black and white techniques, you would be stacking that filter on top of the Bayer colored filter over every pixel photosite on your camera sensor. Photosites whose Bayer filter color matches that of the filter color you placed over your lens would reinforce the filter strength. Photosites whose Bayer color differs from the lens colored filter would effectively block almost all light transmission. A filter than only passes red light stacked on top of one that only passes green leaves little of the spectrum to get all the way through to that photosite. Again, you get the idea.
Your camera processor or raw conversion software will still do its level best to convert the data it has available to it into a useful image, but by filtering out so much of the possible input data you'll be making its job much more difficult. Rather than getting better quality through the use of traditional black and white techniques, you'll be severely hurting quality through the use of inappropriate techniques that no longer apply.
Another common method used for creating black and white images is to make use of the "black and white" mode built into many current cameras. In my experience, this method rarely produces optimal results since the conversion typically just desaturates all colors equally, removing color while retaining only luminance information. In a sense, you could equate this to the use of black and white film in the old days without a color filter — in both cases you're at the mercy of the inherent sensitivity across the spectrum of the recording medium.
Curiously, if you leave the black and white image in raw format, the color information is still in there. It's just that your camera has been told not to show it to you. Most raw conversion software programs will either not honor the instruction to show only black and white or won't even be aware of it in the first place. Open a Nikon raw NEF image shot with black and white mode in Nikon CaptureNX2 and it will show you the requested black and white version. Open that same image in Adobe Lightroom or Camera Raw and it will show full color. In any of these programs though you can freely do whatever you really want with the raw data to make the best black and white rendition you can.
In-camera conversion to black and white may be convenient, but you're almost certainly better off converting to black and white post capture, on your computer. There are numerous techniques and software programs and plugins available to help with this, and I don't have the space in this article to cover them all. But there is one more potential mistake I want to warn you about here that is relevant.
Once converted from raw format, digital images are usually worked on as RGB files with three independent color channels. But if that image has been converted to black and white, the value in all three channels will be the same by definition. If they weren't the same, the image would have a color cast to at least some degree. Given this fact, you really don't have anywhere near as many different possible values as you might think.
A standard 8-bit per channel RGB image has 256 different possible red values, 256 different possible green values, and 256 different possible blue values for each pixel. Since these three values are completely independent of each other, there are a total of 256 x 256 x 256 = 16,777,216 hues and shades available to represent each pixel. But if all three values have to be the same for a black and white image, there are only 256 total values possible. Only neutral combinations are possible in a black and white image. You can have (0,0,0) and (67,67,67) and so on up to (255,255,255) and that's it. All those combinations of RGB values with unequal values aren't available since they wouldn't be neutral hues.
If you don't like seeing banding and other artifacts in what should be smooth gradient areas of an image, never store your converted black and white images in 8-bits per pixel mode. Using math similar to the above, it should be clear that 16-bits per pixel will be far safer when working with black and white. This is generally good advice for any image of course, but becomes even more important for black and white work.
Black and white images create strong impressions by simplifying images through the removal of color. But this color removal also robs your image files of much of the information usually present. It's up to you to be aware of this and employ methods that allow you to keep quality high from capture in-camera through final processing on your computer.