Image Noise

The electronic sensors used in digital cameras do not always record light perfectly. This can be because the sensor has been damaged in some way, or it can be simply because the light levels during the exposure are too low. A damaged sensor may introduce dark pixels that appear as a small dark dot in every photo. Poor light conditions can instead cause a grainy texture in the photograph, which is referred to as image noise.

You can think of a camera sensor as being made up of millions of tiny light collectors, each of which is keeping a count of how much light hits a particular part of the sensor. This count is not always 100% accurate (due to various quantum and electrical effects), and will usually be a little bit higher or lower than the true amount of light seen.  If light levels are good then each collector will receive a  lot of light, and this small inaccuracy will have an almost imperceptible effect.  But if light levels are low then this small change may be enough to slightly change the colour of the pixel.

Take a look at the image taken below, which shows some fish swimming in a dark tank of water.   This was a challenging situation to shoot well – a short shutter speed was necessary to capture the fish without motion blurring, but because of the darkness of the tank and room it was not possible to increase the aperture enough to create a good exposure.  This made it necessary to try shooting in high ISO – a topic we will cover in more detail in the next lesson.  Unfortunately, as we will see, a high ISO can mean a noisy image.

Notice how the magnified image shows a kind of grainy structure with many pixels showing slightly different shades. Although in reality the water has a smooth dark blue colour, the camera has detected small amounts of light (either real or as a result of various electrical processes in the sensor) and has tried to represent this in the image. This causes the grainy pattern in the water, and has the effect of lowering the overall picture quality.  This kind of random pattern is referred to as “grain” or “noise”.

Noise can be caused or made worse in several ways, and we will now take a look at some of these causes.

Poor Light Conditions:  Noise is most commonly introduced into photographs when shooting in low light levels. The less light the camera sensor receives, the more random fluctuations will dominate, and the more noise will end up in the final image. To reduce the impact of this try to increase the amount of light entering the camera. One way this can be done is by using a longer exposure or a larger aperture. Another way is to increase the lighting in the scene, usually through adding additional artificial lighting sources such as flashbulbs.

ISO: You might remember that there are three common ways to increase exposure – shutter speed, aperture and ISO. To reduce noise both shutter speed and aperture are useful, but increasing the ISO can actually make noise even worse. We will go into more depth about ISO in the next lesson, but essentially ISO makes the sensor more sensitive to light. Doing this can increase sensor performance in low light levels, but it also means random fluctuations have a bigger impact. This is especially an issue at higher ISOs, and depending on your camera you may find that noise starts to become a significant problem at ISOs above 1600.

Megapixels: Although camera manufacturers like to market their cameras by the number of megapixels, increasing megapixels is not always a good thing. The more pixels a sensor has the less light each pixel will receive (unless the sensor size increases in proportion). This means that each pixel will be more affected by noise – and so increasing the number of megapixels can result in increasing the amount of noise in your photographs! This can be a real problem with small compact cameras or with smartphone cameras, which often claim a high amount of megapixels but have small sensors, and because of this often struggle in poor light conditions.

Noise can to some extent be dealt with during post-processing. This involves smoothing the image, and lowers noise at the expense of image resolution. However, the best way to reduce noise is to select the right combination of settings when shooting the photo – so keep ISO as low as possible and try increasing exposure through aperture or shutter speed. Alternatively, try to increase the light available, either by adding extra light sources or by making use of the camera flash.

Finally, higher quality professional cameras often contain features designed to reduce the impact of noise. This allows such cameras to use a larger range of ISO before noise starts to become such a problem.

No comments yet