Understanding Noise

Random Colours

Noise is the digital equivalent of film grain but where grain can enhance the appearance and mood of a picture, image noise is rarely as welcome. 

Noise isn’t unique to digital cameras it’s generated by all electronic devices. The key factor is the signal-to-noise ratio (SNR). When the signal  the intensity of light captured by the sensor in this case is greater than the background noise generated by the electronic circuitry, then the image will show little interference. When the SNR is low, it’s harder to differentiate the image information and the unwanted noise.

There are a number of factors that cause image noise to bloom. Increasing the ISO has a significant effect, as it amplifies the noise as well as the signal. The point at which noise becomes intrusive varies between cameras some are much better performers at high ISOs than others but you’ll typically get clean, low-noise results in images shot below ISO 800, with an acceptable amount of noise from ISO 800 to 3200. North of this your mileage may vary, as the size and resolution of the camera’s imaging sensor plays a part: the smaller and higher-resolution the sensor is, the sooner the effects of noise become apparent.

Long exposures lead to more noise too, as the sensor heats up during the exposure especially when the camera is being used in warm conditions. Generally, though, it’s far easier to remove the type of noise that results from a long exposure than it is the noise associated with high ISO sensitivities.

Noise is more obvious in the darker parts of a picture than it is in the bright areas. This is one of the reasons why it’s worth spending the time to get the exposure right in-camera even if you’re shooting raw files, which offer greater scope for rescuing in photo-editing software. If you brighten up an underexposed shot later, you’ll also make the noise that’s been recorded in the dark areas become more prominent.

Rough Texture

Noise is rather a catch-all term. When it comes to digital photography, it manifests itself in two distinct ways: luminance noise and colour noise. Colour or ‘chroma’ noise is the most objectionable. It can cover a picture in a curtain of coloured specks, blotches and bands that obscures details and makes an image look digital and artificial. Thankfully, colour noise is easy to remove in-software without affecting the underlying image too much.

That’s not the case with luminance noise. While this type of noise is more palatable than colour noise it’s closer to grain in film it’s more challenging to reduce it without making a picture look softer in the process. Noise reduction technology finds it tough to discern between fine detail and luminance noise, and ends up smoothing both. You can, at least, adjust the strength of the effect to a greater degree and apply it selectively to parts of an image when you use noise reduction in-software.

The noise reduction setting on a camera is more of a blunt tool, although there are typically several options to choose from. For instance, you may be able to select different strengths of high-ISO noise reduction or deactivate it completely. As it’s an image processing setting, similar to white balance, it’s only applied permanently to JPEGs; shoot raw files and you have the freedom to adjust it when you process the raw files later.

Your camera may offer a different take on high-ISO noise reduction, in the shape of a multi-shot noise reduction. With this activated, the camera takes a sequence of images using a shorter exposure than is required for a normal photo. These files are then combined to create a single image that has low noise but retains detail. The downside is that you need to keep the camera in the same position throughout. 

Finally, there’s long-exposure noise reduction. After you take a slow exposure, the camera records an exposure without opening the shutter. This ‘black frame’ lets the camera map where the hot pixels are, and subtract these from the image.

Sources of Noise

NOISE can ruin the quality of your pictures. While there’s a lot you can do about it post-shoot, as always it pays to get things as close to perfect in-camera to cut down on your editing time. High ISOs
and long exposures are the main causes of noise, as you’ll see below. The noise they generate may only be obvious when you’re scrutinising a picture close-up, and may not be that obvious when the image is seen at normal size. Besides, it’s better to have a sharp noisy picture than no picture.

High-ISO Noise

While aperture and shutter speed control how much light is let into the camera, the ISO determines how much light is required. Low ISOs give you optimum quality, but in situations where there’s little light or you can’t use a wider aperture to get a faster shutter speed for sharp shots, you may have to turn up the sensitivity. Cameras perform differently when it comes to controlling noise at high ISOs, so carry out some tests to work out which setting produces unacceptable results on your own model.

Long-Exposure Noise

Taking pictures in a warm location, using Live View extensively or shooting video before you take a long exposure can cause the camera’s internal temperature to rise to the point where noise develops. You can activate long-exposure noise reduction on a digital camera, but this essentially doubles the time it takes to create a picture. If you’re shooting pictures you want to combine later, such as a sequence of star trails, then this pause can cause a break in the sequence, so you should deactivate the function.

Exposure WOES

The pale sky caused the camera to underexpose (left). Shooting in raw gave us
the flexibility to ‘push’ the exposure in-software (centre), but a cleaner result
was achieved by exposing correctly in-camera (right).

Brightening underexposed shots in-software makes more noise. IF you fluff an exposure when you take it, there are steps you can take to correct the problem in photo editing software. It’s easier to brighten up images that are underexposed (too dark) than it is to claw back detail in shots that are overexposed; but while bringing up the value of shadow areas to reveal hidden details is straightforward, you run the risk of pumping up the volume of noise…

The pale sky is represented by the right-hand peak on the histogram, but it’s too far from the right (bright) side of the graph to be rendered as a bright tone in the photograph.

In Lightroom, we dragged the Exposure slider to the right to correct the brightness of the near-white sky and mid-tone stone. The result is a curtain of noise obscuring shadow detail.

Here, we took another shot of the scene but used exposure compensation on the camera to override the meter and let more light in. As you can see, more light means less noise.

Noise Reduction

Hide luminance and colour noise using the noise reduction software rather than your camera. DIGITAL cameras have settings for reducing noise caused by high ISOs and long exposures, but these don’t offer the control offered by the noise reduction options in photo-editing software. Here we walk you through the sliders available in Lightroom, and what happens when you take each setting to the extreme on a raw file.

Luminance noise reduction needs to be used with care as, taken too far, it
smears fine detail as well as the salt-and-pepper ‘grain’. You’ll need to view the
image at 100% magnification to monitor the strength.

The Detail slider sets the threshold for what Lightroom identifies as luminance
noise. Higher settings reveal more detail – such as in the eyes on the right – at
the expense of some lingering noise.

A lower Contrast setting gives softer results, whereas higher settings add
more ‘bite’ – although tonal changes can look choppy. To check the effect
without the distraction of colour, hold down Alt as you drag the slider.

Lightroom automatically applies a colour noise reduction setting of 25, which
is usually enough to remove the sheen of coloured pixels. High settings suck
away at some underlying colours, as in the eyes here.

Similar to the Luminance Detail slider, this acts as the threshold for
colour noise. The trade-off for increasing the colour detail, is an
increase in colour artifacts, as seen in this chap’s eyelids.

The Smoothness slider lets you reduce low-frequency colour. If you look
closely, you can see splotches of colour in the left shot; raising the
Smoothness removes these, but also knocks back some image colour.

Tips & Trick 3:48 PM