Noise is an unwanted by-product of digital photography and usually shows itself as speckled pixels of an unexpected colour or brightness. It’s a disruption to the smooth gradations of colour that we have come to expect from our digital images. We generally encounter two types of noise: shot noise, which is created by fundamental variations in light; and electronic noise, which is the result of our camera’s internal processing. Regardless of the kit you use and the subjects you shoot, noise forms part of everyone’s workfl ow, so it’s essential to know how it’s caused, how to manage it and how to minimise its effects.
![]() |
Despite being shot at ISO 1600, this candid portrait shows no signs of noise as it was taken with a full-frame DSLR |
Shot noise
The raw material for all photography is, of course, light, and this is made up of tiny packets of particles called photons. Our brains process out the minute delay in these multiple packets of light reflecting off our subjects, so we are never aware of the quantum nature of light – we never think of it as a discrete amount.
An easier way of thinking about light hitting a sensor is to visualise it as rain falling into buckets (the photosite cells on your sensor). When a light shower of rain falls, it is collected in some but not all adjoining buckets, yet when there’s a downpour most of the buckets will capture a similar amount of water. So, during intense rainfall the variation in collected water between buckets is not very significant, but during intermittent rain some buckets will remain dry so the difference will therefore be greater.
Moving back to light, when the flow, or intensity, is great, there’s little variation or noise. However, when the intensity is low, there’s much more variation or noise. In photographic situations, different amounts of light fall on different parts of the subject, creating highlights, midtones and shadows. Noise is most noticeable in the shadow areas. To minimise shot noise, it’s important to make your exposures as accurate as possible. For optimum results, shoot raw and always expose to the right.
Electronic noise
In addition to this fundamental variation in light and how your sensor receives it, noise can also be introduced by your camera’s electronic processing of light, or amplification. As with other electrical systems, when the balance between useful information (strength of signal) and distracting background interference (noise) becomes visible, it can lead to problems.
Electronic noise occurs differently in different cameras: a small amount is introduced when light is converted into a signal by the photosite, but the most significant contributor is the amplifier, which comes into play at high ISO sensitivities – magnifying low levels of light.
Luminance and chrominance noise
![]() |
Chrominance noise is recognisable as random coloured pixels within solid areas of your image |
Although there are many different kinds of noise, generally all noise can be broken into two components: luminance (pixel brightness) and chrominance or chroma (pixel colour).
![]() |
Luminance noise is mostly monochrome and looks like film grain |
Luminance noise is the more difficult to correct as it’s linked to preserving fine details within an image, so when luminance is smoothed image sharpness falls. Chrominance noise is the easier of the two to remove, but its correction must be undertaken carefully to avoid desaturating areas of the image.
Sensor size and noise
![]() |
Small-sensor devices such as camera phones make a poor job of noise reduction when used at high ISOs. This was shot at ISO 2500 |
Not all digital sensors are the same size and neither are their light-receiving cells, or photosites, which are packed side by side on the sensor surface. Many smallsensor cameras that boast of high megapixel values have a high density of smaller photosites crammed together in a small space.
![]() |
100% |
At the other end of the market, fullframe sensors found in professional DSLRs are bigger (but not necessarily higher resolution) and have less densely packed photosites of a larger size. Yet the most important issue is that a full-frame DSLR has a sensor which is four times the area of a Four Thirds camera sensor, so the Four Thirds device produces an equivalent exposure using a quarter of the amount of light compared to a DSLR.
Therefore, smaller sensors densely packed with smaller photosites can be very noisy at mid and high ISO values, as they create a weaker signal that requires more amplification, which in turn creates more noise – hence they have a low signal-to-noise ratio.
Noise Control in Lightroom
For processing raw files, Lightroom provides intuitive noise-reduction tools in the Detail panel. On import, Lightroom automatically recognises raw files and applies a preset noise-reduction process, but you can customise this for your DSLR and, even better, for its different ISO speeds. The golden rule is to process out as much noise as possible before sharpening your image.
1. The starting point
This image was shot as a full-resolution raw file at ISO 6400 and viewed in Lightroom at 400%. On inspection, we can see luminance noise clearly in the smooth defocused areas in the background, as well as some further noise in the animal’s fur.
This image was shot as a full-resolution raw file at ISO 6400 and viewed in Lightroom at 400%. On inspection, we can see luminance noise clearly in the smooth defocused areas in the background, as well as some further noise in the animal’s fur.
![]() |
Before After |
2. Increase Luminance
Arrange your desktop to show a Before/After view. Increase the Luminance slider until the noise softens, but not so much as to lose sharpness. This example has had +25 increase. You don’t need to remove the entire noise texture otherwise the image will look odd.
Arrange your desktop to show a Before/After view. Increase the Luminance slider until the noise softens, but not so much as to lose sharpness. This example has had +25 increase. You don’t need to remove the entire noise texture otherwise the image will look odd.
3. Push Detail
By default, the Detail slider will be set at +50, so move this up until you see sharper detail. The downside is that as Detail is increased, unwanted noise can reappear. Low amounts of Detail will create a softer result, which may be acceptable for smooth-textured images.
By default, the Detail slider will be set at +50, so move this up until you see sharper detail. The downside is that as Detail is increased, unwanted noise can reappear. Low amounts of Detail will create a softer result, which may be acceptable for smooth-textured images.
4. Explore Contrast
The final control is to experiment with the Contrast slider. Contrast works by blending areas in your image together, so they appear less visible in the end result. Like Detail, use Contrast sparingly, and if it makes no difference leave it on its default value of 0.
The final control is to experiment with the Contrast slider. Contrast works by blending areas in your image together, so they appear less visible in the end result. Like Detail, use Contrast sparingly, and if it makes no difference leave it on its default value of 0.
5. Customising your DSLR
You can create preset NR for your DSLR and each ISO in Lightroom. Open Preferences and click on ‘Make defaults specific to camera ISO setting’. Open a file shot at a specific ISO and make the adjustments. Then choose Set Default Develop Settings from the Develop menu.
You can create preset NR for your DSLR and each ISO in Lightroom. Open Preferences and click on ‘Make defaults specific to camera ISO setting’. Open a file shot at a specific ISO and make the adjustments. Then choose Set Default Develop Settings from the Develop menu.
Color Noise Control in Photoshop
Most Photoshop users undertake their raw-file noise reduction in Camera Raw using an almost identical set of tools to those in Lightroom. However, Photoshop’s ability to work with channels and edit in the theoretical LAB colour space helps it to deliver an alternative way of dealing with chrominance noise.
This is an extreme example of chrominance noise taken with a small-sensor camera phone. Patches of coloured pixels seem to ‘infect’ the pure whites and blues of a daylight image. We are going to remove as much of the colour as possible without compromising sharpness.


Open your Channels palette and click on the single ‘a’ channel – making the desktop image turn monochrome. From the Filters menu choose Blur>Gaussian Blur and increase the Radius value until sharp shapes start to soften, as shown.

Repeat Step 3, but this time on the ‘b’ channel, as shown. Because the Lightness channel contains the sharpness information, but not colour, you can edit channels ‘a’ and ‘b’ without softening the image.
5. Final result
To reveal the results, click on the master Lab channel to switch it all back on, as shown. You will notice that random colour noise is largely absent, but the image will still remain sharp.
To reveal the results, click on the master Lab channel to switch it all back on, as shown. You will notice that random colour noise is largely absent, but the image will still remain sharp.
Film grain and digital noise
![]() |
Film grain can have a pleasing granular appearance, which becomes more apparent in bigger prints |
Although there are some visual similarities between film grain and digital noise, they are produced by very different circumstances. Grain in film becomes visible in a print as a result of the development of silver-halide crystals in the film emulsion. News and sports photographers typically ‘pushed’ the working ISO of film from 400 to 1600 by extending its chemical development longer than usual. Like copper-sulphate crystals grown in a test tube, the longer the film was developed, the bigger the crystals grew and the more visible the grain became.
While the grain characteristics of certain films, such as Kodak Tri-X, were used as a visual effect by photographers such as William Klein and Daido Moriyama, digital noise is seen much more as an unwanted electronic byproduct than a stylistic choice.
Practicalities of noise
During the rapid development of camera technology, noise has become a byword for all that’s bad about digital, yet in reality it’s much less of an issue than you think. Today’s professional DSLRs provide exceptional image quality with much lower levels of noise compared to earlier camera systems.
Shooting such quality at high ISOs results in many new possibilities. We can now shoot subjects and situations that would have been tricky to capture previously, such as moving subjects in low light, or using atmospheric ambient light rather than flash to retain the mood. For black & white photographers, noise has much less of a visual impact on the final product as it can be easily disguised using post-production techniques in Lightroom or Photoshop.
Controlling noise in-camera
There are two types of noise-inducing scenarios that better DSLR systems can mitigate against: noise created by long exposures and noise created by high ISO settings. As each of these scenarios produces slightly different kinds of noise, two separate in-camera processes address them.
![]() |
Images shot at ISO 800 show little noise when processed properly and printed at A4 or smaller |
Useful camera settings
File formats and NR reduction
Like all other on-board camera settings, noise-reduction settings affect only JPEGs and are not embedded into raw files.
Post-processing noise in raw files
If you use your DSLR’s propriety raw file processor, such as Nikon Capture NX-D, this will make full use of the data embedded in the raw file to trigger a more sympathetic noise reduction than a third-party postproduction application such as Photoshop or Lightroom. The third-party application won’t differentiate between different NR reduction settings the camera has used.
Noise and heat
Higher than usual sensor temperatures can also cause unexpected noise to appear in your images, even at low ISO values. Shooting with live view for extended periods of time can cause the camera’s internal temperature to rise, for example, so it’s always advisable to switch this off between shots, especially if you are also shooting long exposures and/ or using high ISOs. While this may not be an issue for mirrorless cameras, which are designed to be used with full-time live view, they are cited as a possible cause of noise in most recent DSLR camera manuals.
Auto ISO dilemma
Most DSLRs have an additional setting that keeps the ISO value fl uid throughout your shoot, so you can maintain a chosen aperture or shutter speed setting depending on the subject you’re shooting. Most cameras allow you to set a ceiling value for auto ISO so you don’t inadvertently shoot at too high a value, such as 6400. For high-speed shooting in changeable conditions, where the emphasis is on freezing motion rather than controlling noise, auto ISO can be very useful.
Noise-reduction limitations
Most in-camera noise-reduction techniques involve complex on-board processing that will slow down your workflow. Depending on the method used, NR processes can reduce your burst rate if you are shooting multiple frames per second, as the system struggles with the data before allowing you to shoot again.
So, if you are shooting high-resolution files of fast-moving subjects, and you have high ISO noise-reduction switched on, you may find that you fill your buffer up quickly, and are unable to shoot as many frames as you expect.
Noise and ISO values
ISO is a term used in photography to indicate the light sensitivity of your sensor or film stock. Compared to 10 years ago, digital-camera sensors are designed to produce acceptable images under a much wider range of low-light conditions. But generally, it’s still a good idea to use low ISO values, such as ISO 100 or 200.Like shutter speeds, ISO values follow a linear scale such as 100, 200, 400, 800, 1600, 3200, 6400, and as each of these values doubles the sensor requires only half the amount of light to work effectively. When the ISO value is halved from 400 to 200, the sensor requires twice the amount of light to obtain a correct exposure. Low values such as ISO 100 are selected when shooting in bright conditions; higher values such as ISO 800 are used when less light is available.
Native or base ISO
Unlike film, which has a fixed ISO, digital sensors operate across a wide range of ISO values, typically 100-6400, while some pro DSLRs keep doubling up to stratospheric values such as 102,400.
In order to facilitate this huge range of sensitivity, sensor manufacturers design their products to work in three different modes. At its native or base ISO, typically 100, the sensor operates at maximum quality and produces little or no noise at all. (Although camera manufacturers never advertise the native ISO of their DSLRs, this is almost always the lowest ISO value such as 100 or 200 – excluding
other settings marked ‘expanded’ or ‘low’).
In order to facilitate this huge range of sensitivity, sensor manufacturers design their products to work in three different modes. At its native or base ISO, typically 100, the sensor operates at maximum quality and produces little or no noise at all. (Although camera manufacturers never advertise the native ISO of their DSLRs, this is almost always the lowest ISO value such as 100 or 200 – excluding
other settings marked ‘expanded’ or ‘low’).
Once you increase ISO from its native value, the camera electronically amplifi es the signals from lower light levels. In this secondary mode, noise begins to appear as amplifi cation increases.
A third mode, sometimes referred to as boosted, expanded or extended ISO, makes shooting possible at extremely high ISO settings by employing a software process to simulate speed gain rather than electronic amplifi cation. In this setting, noise will be very visible and there will also be a drop in dynamic range.
Risky intermediate ISO settings
In the same way that analogue photographers once experimented with film speed and development times to determine optimum ISOs, digital photographers have tested noise performance across a range of ISO settings and unearthed some interesting findings.While we assume that the amount of noise increases as we increase the ISO, there’s plenty of evidence to suggest that intermediate settings can actually increase noise more than you might expect.
For recent Canon DSLRs, there’s evidence that intermediate ISO settings, such as ISO 125, create the same amount of noise as ISO 200, with the effects mostly visible in extreme shadows. This is due to a crude internal process that appears to amplify noise at these values.
Post-production noise control
If you shoot raw files, or prefer not to suffer buffer delays as a result of using in-camera NR, then controlling noise in post-processing is the best option.Best full-frame cameras for low noise
![]() |
Canon EOS-1D X Mark II |
Canon EOS-1D X Mark II
With a top native ISO speed of 51,200 (expandable to 409,600), this 20.2MP full-frame DSLR is perfectly suited to low-light shooting. Even in auto mode, it can achieve shutter speeds up to 1/8000sec.
With a top native ISO speed of 51,200 (expandable to 409,600), this 20.2MP full-frame DSLR is perfectly suited to low-light shooting. Even in auto mode, it can achieve shutter speeds up to 1/8000sec.
![]() |
Nikon D5 |
Nikon D5
Combining a 20.8MP FX-format sensor with Expeed 5 technology, the Nikon D5 has a native ISO of 100-102,400, expandable to ISO 3,280,000. The AF system is designed to perform in near darkness (-4 EV).
Combining a 20.8MP FX-format sensor with Expeed 5 technology, the Nikon D5 has a native ISO of 100-102,400, expandable to ISO 3,280,000. The AF system is designed to perform in near darkness (-4 EV).
![]() |
Sony Alpha 7R II |
Sony Alpha 7R II
The palm-sized Sony Alpha 7R II has a back-illuminated 42.4MP full-frame sensor, and an ISO range that can be set as high as 102,400. The Bionz X image-processing engine finely tunes data for clearer, more natural images.
The palm-sized Sony Alpha 7R II has a back-illuminated 42.4MP full-frame sensor, and an ISO range that can be set as high as 102,400. The Bionz X image-processing engine finely tunes data for clearer, more natural images.
In-camera noise-reduction (NR) options
ISO noise reduction is applied automatically by your camera’s on-board systems at all ISO values above its base or native setting.![]() |
These three examples were shot as JPEGs at ISO 3200, with varying levels of in-camera noise reduction applied |
Long-exposure noise reduction
This isn’t quick and will usually take at least twice as long as your original exposure time. Although there are different methods of long-exposure noise reduction, the process usually involves the creation of a second ‘dark’ frame, which is combined with the original to subtract the noise.
![]() |
On a two-second exposure, long-exposure noise reduction smooths out unwanted noise, but slightly at the expense of detail and texture |
High ISO noise reduction
Most DSLRs have a range of high ISO NR options such as Standard, Low, Strong or Disable. While Strong NR will be effective on high ISO speeds, it will also slow down your burst rate. Experiment to see which works best for your particular workflow.
Multi-shot noise reduction
Many camera systems offer a third method of NR, which combines multiple shots into a single frame. Suitable for non-moving subjects, it’s a bit like HDR for noise reduction.
Best APS-C cameras for low noise
![]() |
Nikon D500 |
The standard ISO range of the Nikon D500 (100-51,200) can be lowered to 50 or expanded to 1,640,000 (equivalent). The smaller sibling of the D5, this APS-C DSLR is not afraid of the dark, and
handles highlights well.
![]() |
FujiFilm X-T2 |
The Fujifilm X-T2 features a 24.3MP X-Trans CMOS III sensor with no low-pass fi lter, delivering superb image quality. The native ISO range stretches from 200-12,800 and can be expanded to 51,200.
![]() |
Sony Alpha 6500 |
The Bionz X image-processing engine in the Sony Alpha 6500 enhances picture clarity, while minimising noise. It boasts a 24.2MP sensor in a tough, palm-sized body, and an ISO range that can be expanded to 51,200.
0 comments:
Post a Comment