In the days of film, we worried about grain, and higher ISO speeds resulted in larger, more pronounced grain. Now we have noise. Digital noise is also affected by ISO, but unlike grain there are many ways to eliminate it before and after exposure.
Advances in DSLR sensors allow photographers to now shoot at ISO 4000 much as they once used ISO 400. The first step is to ensure you’re working with a state-of-the-art digital camera capable of making high-ISO exposures with amazingly low-noise characteristics. Still, it’s always best to shoot with the lowest ISO possible to achieve the least amount of noise.
Noise builds up with long exposures, so whenever possible choose a shorter exposure. Also, adjusting in-camera noise reduction to a more powerful setting can help. The trick is to strike the perfect balance between eliminating noise and maintaining detail. When in doubt, remember that it’s always better to get the shot than to miss it because of inadequate, albeit low-noise, settings. If you’re shooting RAW, remember that exposing correctly sometimes means subtle overexposure to maximize detail and minimize noise in shadow areas. It’s like artificially lowering the working ISO.
Working with RAW offers you additional noise-reduction capabilities in the computer. Programs such as Aperture (www.apple.com) and Photoshop Lightroom and Camera Raw (www.adobe.com) actually encourage photographers to minimize noise by making it so easy. These tools all incorporate the same basic sliders to fine-tune noise reduction and maintain image detail.
Even photographers who don’t shoot RAW can utilize extensive noise-reduction capabilities built into Photoshop. The Reduce Noise filter subtly blurs noise at the pixel level. Within that dialog, it’s the Strength slider that adjusts the amount of noise reduction, while the Preserve Details slider maintains edges to mitigate the blur. The filter also has sharpening options built in, although you can always resharpen the image with other approaches in a subsequent step. Click the Advanced toggle to remove noise in individual channels—immensely helpful since most noise is found in the green channel.
Another method is to remove noise with manual pixel blurring. Selectively reducing noise where it’s most evident within a frame—usually the darkest areas—can be an ideal way to eliminate noise without sacrificing detail throughout.
Simply duplicate the image to a new layer, blur it with the Gaussian blur tool (or your favorite technique) and then paint away a layer mask to allow noiseless areas to show through. The manual approach is a little more involved, but the results are completely customizable.