You’ve probably read or heard about unacceptable noise in digital photos. What does that mean and how does it occur in photography?
What is Photo Noise?
In digital photographs, the term noise refers to a certain type of visual distortion. It looks similar to grain found in film photography, but can also look like splotches of discoloration when it’s really bad, and can ruin a photograph. Electronic noise, or digital noise, tends to get worse when you’re shooting in low light.
From a technical point of view, image noise is the visual manifestation of a lower signal-to-noise ratio, which is measured in decibels. While the amount of noise that you may consider acceptable may be different from what the next guy might accept, most professional photographers want to see photos with at least a 30dB signal-to-noise ratio. Generally, the less noise in an image, the better.
What Are the 3 Common Types of Image Noise?
If you are curious about what noise is in photography, you have to know how to spot it. In digital photography, there are three types of noise: random, with a fixed pattern, and banding.
Random noise is an unpredictable spread of imperfect pixels with either color or intensity different from the surrounding pixels. It appears when you photograph in low light conditions and use high ISO values. The distribution of noise in two photographs taken in the same conditions differs.
Fixed pattern noise is more obvious than random noise because the imperfect pixels have a higher intensity. It appears when you expose the camera to a lot of light and high temperatures. In this case, the distribution of noise in two photographs taken in the same conditions is the same.
Banding noise looks like stripes and is camera-dependent. You can’t do much about it, but it accentuates when having high ISO values, large areas in shadow, and certain white balances.
What causes image noise in photography?
High ISO
Higher ISO, which you may need when shooting in low light, is the main culprit in causing more image noise. Think of ISO as the gain knob on an amplifier for an electric guitar. The more gain, the louder the guitar’s sound, but it also becomes distorted compared to the clean sound of the guitar without gain.
In a photograph, the higher the ISO, the more distortion (noise) shows in the image, regardless of what shutter speed you use. You may feel tempted to increase the ISO to get more light to the sensor. But along with that higher ISO, you’re risking increased noise.
Small Camera Sensor Size
When it comes to noise, sensor size matters. Smaller camera sensors like those found in cell phones and compact cameras are thumbnail-sized. On these cameras, noise can reach unacceptable levels even at ISO 400.
By the time you reach ISO 800 or higher, the picture may end up looking like an impressionist painting and lose sharpness, detail and color fidelity. Even professional-level crop sensor cameras present this issue. Cameras with larger sensors, such as DSLR or mirrorless cameras, produce lower grain at higher ISOs. The larger the sensor, the better the grain at comparable speeds.
Pixel Density
A sensor with 14 million pixels (megapixels) will produce more digital noise than an equal-sized sensor with 10 megapixels. That’s because, in order to squeeze those extra 4 million pixels, the actual pixel size has to shrink, which means each pixel will let in less light (think of smaller apertures in lenses letting in less light than larger apertures). To compensate, the “gain” is turned up, and this causes distortion. Conversely, a larger sensor with 14MP will produce less noise than a smaller 14MP sensor.
Shutter Speed
Long exposures can introduce static, which can also be a cause of digital noise, specifically luminance noise. When using a longer shutter speed, try to mitigate noise by reducing ISO or taking advantage of in-camera noise reduction.
Shadows
If you are shooting in broad daylight at a higher ISO, the grain might not be so obvious… unless you look at the shadow areas. Grain shows up more against darker subjects or backgrounds. It gets even worse if, using image-editing software such as Adobe Photoshop, you lighten an image. Then the grain in the shadow areas will become even more obvious.
The photo above was shot at ISO 4000 with a Canon 7D DSLR at the Kennedy Space Center, Cape Canaveral, Florida. While the image technically has more noise than what is considered acceptable, the shot noise level doesn’t really interfere with the image, in our opinion.
What do you think?
How much noise in photography is too much?
While there are absolute numbers that are used to define levels of noise in photography, the amount of shot noise that would be considered unacceptable to one photographer may be fine for someone else. A snap-shooter, for instance, may not care about noisy images as much as an enthusiast or professional photographer. It also depends on usage. Higher noise may be tolerated in, for example, a nighttime sports action shot, but that same amount of noise would render an architecture photo, where detail and color fidelity are critical, unusable.
The size of the prints you make also factor into the “how much noise is too much” equation. Bigger prints will show noise artifacts more than small ones. Noise that might be huge and distracting in an 11×14-inch print might not look so bad on a 4×6-inch print, and might be almost invisible on a computer monitor.
So, “acceptable” noise is a very subjective concept!
How can you avoid too much noise or reduce noise in photos?
Shoot at a low ISO
Higher ISOs produce more noise. Therefore, choosing the lowest ISO you can while still getting a good exposure helps reduce noise. For a compact camera, that might mean only shooting at ISO 100. For a DSLR, you might get away with ISO 400-800, or even higher on some recent models. If you shoot only at low ISOs, however, you will likely need a tripod or a flash for low-light photography.
Get a camera with a larger sensor
Give your compact camera to your kids and move up to a DSLR, and you’ll see a major improvement in image quality. DSLR too big or heavy? Consider one of the new breeds of smaller interchangeable-lens (MILC) cameras.
Use in-camera noise reduction
Many modern digital cameras, both compacts and DSLRs, have in-camera noise reduction. Many compact cameras will apply image noise reduction to JPEG images as a default. DSLRs tend to have the option of noise reduction on, off, or on at high or low settings. If you definitely want noise reduction off, shoot in RAW mode.
Use noise-reduction software on your computer
If your camera doesn’t have image noise reduction built in, or you prefer shooting in RAW vs JPEG and adjusting the noise levels after the fact, use noise reduction software such as Nik Dfine. The advantage of using noise reduction software over in-camera noise reduction is that you have more powerful tools such as reducing grain selectively to parts of the picture where it’s needed while leaving other areas alone.
Many noise reduction software options include adjustable noise reduction sliders. These let you control luminance noise reduction and color noise reduction. With sliders, you can get the exact level of adjustment you prefer.
Shadow Play
Digital noise can be seen more readily in shadowy areas of shots. In this photo shot indoors at ISO 2000 with a Canon 7D at the Kennedy Space Center (full shot above) notice in the 100% how the grain is more obvious in the darker areas of the image than the lighter ones.
Luminance noise vs color noise
Noise in photography isn’t ideal regardless of what type it is. However, some types of noise are easier to deal with than others. For starters, let’s review what the two types of noise in images look like.
Luminance noise looks similar to the film grain that many photographers are used to. It’s a variation in brightness. Color noise looks like brightly-colored pixels that contrast from the colors in the rest of the image. With color noise, you’ll see a blue pixel next to a red one next to a green one. Color noise also tends to look splotchy and blurry.
Luminance noise is usually easy to fix using the luminance noise reduction slider in Lightroom. Color noise is a bit more finicky and can result in some excessive blurring and smoothing in your images. Be gentle, though, because applying too much noise reduction can make your images look “plastic.”
Is photo noise bad?
Noise in photography isn’t ideal. Noise seriously complicates your ability to print images or share them digitally in high resolutions. When it comes to noise, the best advice is to avoid it in the first place using the techniques mentioned above.
As you’re shopping for cameras or planning photography excursions, consider how noise may come into play for you and the shots you intend to take. Avoiding noise on the front end will make your editing sessions infinitely easier.
The best way to de-noise an image in post processing is to practice! Clean noise reduction takes some work and finesse. Don’t be afraid to reprocess old, noisy images until you’re confident in your techniques. Make sure as you work that you aren’t losing detail in your images. Keeping some noise is always preferable to losing significant image detail.
Plug ins and special software make the process faster and more precise, so do a little legwork before you start.
How do I find out how well my camera handles digital noise?
DxO Labs’ DxOMark service tests digital cameras for overall image quality, but they also do specific tests to determine the camera’s noise levels at all ISOs. You can search the Sensor Rankings database for your camera model. The DxOMark Sensor Score Sports (Low-Light ISO) will tell you the highest ISO at which DxO measured acceptable noise levels (which they define as 30dB or higher). Click on the “SNR 18%” tab to view a graph that shows the camera’s signal-noise ratio, measured in dB.
DxO only measures cameras with RAW capabilities, so they won’t have info for most compact cameras. You can assume, however, that noise will start to increase on a typical compact camera by ISO 200.
The good news is that the latest generation of digital cameras does a better job producing images with lower noise levels. As technology evolves, future generations may reduce noise levels even more or eliminate it altogether.
What is Photo Noise Reduction?
Photo noise reduction is a way to reduce image noise in post-processing. Photo editing software smoothes the image by evaluating the differences between values of adjacent pixels. They may use the intensity or the color of the pixel as a reference. In terms of smoothing methods, most editors use filters, such as a Gaussian blur filter or a Median filter.
Smoothing filters don’t make the difference between noise and details. The more aggressive the filter is, the larger the detail loss. Therefore, advanced photo noise reduction tools include artificial intelligence (AI) algorithms that help them accurately detect and eliminate noise. Most of them focus on random noise.
