Read nearly any smartphone or camera review and you'll come across a discussion of "image noise." Typically, the writer will say the camera shoots decent images in good light, then fault it for producing noisy images in low light. However, the review may not explain what noise is, why it occurs, or how to prevent it.
What is image noise?
In digital photography, image noise consists of specks of false color, called artifacts, that degrade the quality of the picture. For instance, if you look closely at a photograph showing a blue sky at sunset, you may see a random pattern of tiny red dots, giving the image a blotchy and mottled appearance. If you're photographing people's faces, they may end up looking like they suffer from rosacea.
This effect is most common in pictures shot in low-light situations, when the camera's internal metering may set the camera on a high ISO setting to capture more light. The camera's circuitry tries to squeeze more information out of whatever illumination is available—essentially, it guesses at how to interpret the light that enters the lens. That works, up to a point, but it also ends up producing irrelevent specks of color.
Find the best model for your needs and budget: Check our camera buying guide and Ratings.
Something similar occurs with high-ISO film—but the results aren't nearly so ugly. A digital camera's ISO setting is basically an analogy to old-fashioned film's ISO ratings. (The letters stand for the "International Organization of Standardization," the association that developed specifications for film sensitivity decades ago.) But, like all analogies, it's imprecise. And that causes some confusion.
With film, tiny, light-sensitive crystals are embedded in the emulsion—the thin layer of goop that coats the plastic substrate. These grains undergo chemical changes when exposed to light and then they change further when they're being developed using a series of chemical baths. The larger the crystals, the more sensitive the roll of film: High ISO film, such as ISO 3200, has larger crystals than low ISO film, such as ISO 100, and it allows the photographer to shoot in lower light. On the downside, the crystals in high-ISO film are so large that in the final photo you'll actually notice the grain, especially when you enlarge and print those images. Those large grains can compete with photographic detail, and this is particularly noticeable if you're shooting subjects with very fine, subtle patterns. However, the grains don't distort the color—and a somewhat grainy film photograph can impart a particular kind of artistry.
In digital photography, the tiny elements that together capture the image aren't light-sensitive grains of silver. Instead, they are pixel sites located on the camera sensor. They detect how much or little light is hitting them, and send corresponding electrical signals to the camera's processor. The processor converts each electrical signal into a pixel, one of millions that will make up the image. It's during this process that errors can occur, resulting in image noise.
There's a big difference between film and digital cameras: When a digital camera boosts its ISO, it doesn't switch in a new image sensor, the way a photographer would swap in faster film in an analog camera. Instead, the circuitry performs some electronic magic to increase each pixel's reaction to light. In fact, the ISO on digital cameras should really be called "ISO equivalency." We'll discuss the details some more, but the result is that the distortion produced by low-light, high-ISO shooting in digital photography tends to be much more distracting than what's seen with film photography.
Why does low light produce noise?
When it comes to digital photography, there isn't just one kind of image noise. In fact, if you shoot enough digital photos, you'll start to notice that different shooting situations produce different patterns or types of visual artifacts. That's because some problems originate on the sensor, while others may come from the processor misreading the signal or from another source.
One big cause of image noise is the fact that individual pixel sites on your camera's image sensor are never completely uniform in size and how they function (due to slight manufacturing defects and variations). In bright light, the pixel sites produce strong signals, and you don't notice these differences. But in marginal conditions, when each pixel isn't registering a strong signal, slight variations become more pronounced, and the resulting image includes pixels with the incorrect tone.
It's up to you to decide when an image is too noisy to keep. That judgement will depend, to some extent, on the subject. A photograph with a lot of small textures and details may not look as noisy as an image with few details and large, continuously toned sections. The size of the print is also a factor—the bigger the print, the more conspicuous the noise.
Of course, ideally you'll just produce photos that have little visual noise to begin with. Here are five tips to help you do just that.
Six tips for minimizing image noise