Noise has long been a bête noire for digital photographers, as it causes randomly colored pixels or patterns that interfere with the purity of the image. In early digital cameras, noise very seriously impaired image quality. Today, with the availability of larger files and improved technology, high ISO noise is much less of a problem than it used to be.
Noise arises from the electrical nature of the camera’s circuitry, from the properties of the sensor, and from image processing. It unequivocally reduces image quality in three distinct ways: because it’s distributed randomly, it obscures fine detail; its grainy appearance reduces the smoothness of tonal changes and the purity of colors; additionally, an indirect effect of noise is that it increases the file size of compressed images.
Types of Noise
All images exhibit noise: the important factor is whether it’s visible and detracts from image quality, or whether it’s invisible or is aesthetically acceptable. Noise may contribute to the texture of an image, and may help reduce banding artifacts during printing.
📷But usually it’s not welcome, particularly if it occurs in fixed patterns such as stripes or bands (banding noise), or as large and irregular specks of color. It tends to increase with higher ISO settings, longer exposure times, and with hot cameras— those with a high operating temperature. It’s often worst in shadows, and not so bad in bright parts of an image.
Noise also increases with image manipulation such as large exposure corrections, large color changes, Shadow/Highlight corrections, and after in-camera adjustments such as Dynamic Range Optimization. Sharpening may also increase the visibility of noise, as will enlarging the image.
High iso Noise Reduction
Noise becomes visible when the affected pixels are about the same size as fine detail in the image. Any procedures you use to reduce noise may also reduce fine detail—software includes Topaz DeNoise, Noise Ninja, and PR Noise Corrector. Some work as plug-ins, some as stand-alone applications. They aim to distinguish details from noise in order to reduce noise selectively — making tones smoother without blurring details.
Some software use “noise profiles” descriptions of the noise patterns specific to a camera and ISO setting, derived from analyzing sample images to improve the selectivity of noise reduction to reduce loss of detail. At the same time, these applications may apply sharpening and other adjustments, all of which you can control. Ultimately you need to strike a balance between tolerable loss of detail, and reduction in noise.
Despite being digital photography’s main enemy, noise can also be useful. For example, it’s essential when you want to recreate the texture and tone of film-based prints; for the most realistic results, software from DxO Optics and Lightroom can simulate film grain based on noise profiles obtained from actual film.
Image averaging reduces noise in low-light exposures of static subjects. To do this, make two or more correctly exposed shots using a tripod. Open the images and copy them all into a new image, each as a layer. Change the opacity of each layer according to the number of layers below it. Set the bottom layer to 100 percent; if there’s one layer below, set this to 50 percent; for two layers, 33 percent; for three, 25 percent; and for four, 20 percent. Then, flatten the image. Some cameras carry out this process automatically, although making the multiple exposures will greatly increase the capture time.