The word "noise" is most commonly used to describe a sound, often a loud one, that is unpleasant or causes a disturbance. Noise, however, can describe an unwanted irregular pattern that disturbs any kind of signal.
In digital photography, noise is the granularity and false colors found in your image, leading to loss of detail and color fidelity. Noise becomes especially apparent as the signal from the sensor is amplified.
Noise is the same kind of electronic static you'd see on television, hear on the radio, or pick up on a phone call. In cameras, noise presents itself like granules of sand that increase in size, randomness, and color variation as this static increases. For the most part, better cameras are able to mitigate this static very well and full-frame, flagship bodies tend to do it best.
The noise in your images increases as you increase the ISO because you're increasing the signal-to-noise ratio: allowing more irrelevant information in. You're certainly able to capture more information as you boost the signal, but you begin to lose accuracy as you do so, creating noise. Noise also happens when you try to amplify or enhance a low amount of source's signal beyond its natural level. These two instances are exemplified when shooting the same scene at different ISOs or when brightening a dark image.
Your camera's ability to mitigate noise is a balancing act between the sensor size and technology, the camera's processor and software, and user input (file type, settings, etc.). It's important to learn this balancing act well in order to create images with less noise. You'll have a better time in post-production when optimizing your images because you've provided the best possible starting point.
Noise reduction is simply cleaning up the signals your device (your camera) has received. By reducing the noise, you're left with more signal. You get a cleaner picture. Applying a noise reduction algorithm removes the different types of interference that clutters up your photo. It interpolates (fills-in) the gaps and errors made by the noise.
Completely eliminating noise from an image isn't possible because no signal is pure as captured. Noise reduction does not mean noise elimination. Because noise reduction can reduce the quality of your image in other ways when over-used, the trick is finding the least amount of noise reduction that will produce an acceptible image for your purposes.
Also, because digital signals are binary -- a series of 1's and 0's -- you can't have signal and noise occupying the same exact space. Interpolation tries to smooth that out and make an educated guess as what should occupy that spot. If you apply too much noise reduction, you'll lose the detail and color fidelity you were aiming to preserve. The smoothing effect of interpolation turns into broad-brushed smudging.
Just like your camera, applying noise reduction in post-production is a balancing act between the camera's inherent capabilities, software capabilities, and user input. Understanding this and giving yourself the best possible starting point at each step will give you the ultimate flexibility in reducing noise in your photos.
Noise Reduction in Lightroom
In the Develop module of Adobe Photoshop Lightroom, you'll see a section called Detail where the Sharpening and Noise Reduction tools are located. The Noise Reduction tool separates noise into two main categories: Luminance and Color. Each of these categories has its own set of sliders to adjust for each aspect of that type of noise.
Luminance is the granularity itself and Color is the color of the granularity in the noise of the image. So, one will smooth out the grain and the other will even out the color speckles of the grain. Lightroom helps you control each of these separately so you can focus which kind of noise is more of a problem in your image.
Luminance Noise Controls
The first section of the Noise Reduction tool are the Luminance sliders. They are: Luminance, Detail, and Contrast. Each controls three aspects of Luminance noise.
- The Luminance slider focuses on the granularity of the noise. As you increase its effect, the grains blend together. If too high, fine details like hair, mesh, and fabric textures are lost.
- The Detail slider adjusts the threshold for the Luminance slider's effect. It helps regain some fine detail that the Luminance slider removed. It's somewhat of a counterbalance to the Luminance slider.
- The Contrast slider adjusts the contrast of each granule in the noise. Works much like conventional contrast but on a smaller scale. Its effects are quite subtle, even at 100.
Color Noise Controls
The second section of the Noise Reduction tool are the Color sliders. They are: Color, Detail, and Smoothness. Each controls three aspects of color noise.
- The Color slider adjusts the errant color specks in your image, applying more color uniformity as you increase it. But if you're too heavy-handed similar colors will blend and flatten.
- The Detail slider works much like the one in Luminance section. It works on per-pixel color separation in fine details.
- The Smoothness slider determines how gradual the transition between corrected colors is. It's much like the Contrast slider, but for color variations. Its effect is also very subtle.
Other Tools for Noise Reduction
Sometimes you want to selectively apply noise reduction to your image. Three tools in Lightroom enable you to apply noise reduction to portions of your image rather than globally. These tools are the Radial Filter, Gradient Filter, and Adjustment Brush. Each of these tools have a slider that will allow you to adjust the noise only the areas where these tools have effect.
Selective application of noise reduction with these tools is useful especially when global adjustments would degrade your image or you want more precise control over the noise reduction effects. For example, shadows that have been brightened tend to show more noise than highlights, you could use one of these tools to apply stronger noise reduction than to other parts of the image.
Sharpening & Noise Reduction
Older cameras or very high ISO images tend to have a lot of noise and require strong noise reduction to improve the image. However, many times strong noise reduction can cause a significant loss of detail due to the blurring effects of strong noise reduction. To counteract this effect, you can apply Sharpening.
Small amounts of sharpening an image go a long way, but it can reintroduce the noise you just removed without refining its application. A Masking slider in the Sharpening tool can help restore sharpness without undoing your noise reduction work. When set to zero, your settings are applied to the entire image. As you increase the masking, the effect's distribution is decreased until only the edges of high contrast areas have visible sharpening. So, you can restore the perception of sharpness by applying it only to the edges of people and things in your image.
What is great about combining Sharpening Masks with Noise Reduction is that there are instances where can use stronger Noise Reduction settings than usual and still maintain edge sharpness. This is great for skin, walls, skies or other smooth surfaces.
Noise Reduction Presets
The ultimate goal of learning these different aspects of noise reduction is to create reusable presets in Lightroom for each camera and ISO combination. Presets are a huge time-saver. For example, my Canon EOS 1D Mark IV's noise becomes bothersome to me from ISO 3200 and up, so I created presets for them in order to consistently apply batch-corrections for that ISO.
Additional tweaks to these presets such as sharpness, vibrance, or Hue-Saturation-Luminance (HSL) settings can also be integrated according to your camera's characteristics. So, if you notice a particular color goes flat at a certain ISO, you can boost it and save it into the preset.
If you want to be very precise, integrating accurate color profiles using an X-Rite Color Checker or X-Rite Color Checker Passport at different ISO settings is the way to go. To cut down on time, only start the process where the noise becomes bothersome and go up from there. Make your noise reduction, sharpness, etc. adjustments and save it as a preset for each ISO.
Although reducing noise in your image begins at the moment of capture -- using the lowest possible ISO and nailing your exposure -- it is great to have noise-reduction available in post-production. Your camera can only do so much and Lightroom's Noise Reduction is powerfully refined. Good Noise Reduction technique will give you cleaner images and enable you to display your photos larger both on screen and in print. Utilizing presets for Noise Reduction will save you a lot of time so you can work smarter.