Reflection XII Adagio

high noise / low noise


column and row noise / severe under-exposure


hot pixel noise / very long exposure or very hot conditions


Bayer pattern noise / substantial under-exposure


random noise / high ISO

While it’s best to eliminate noise in images at the point of capture (by choosing optimum tools and making exposures), taking the steps to do this may be impractical and/or lead to unacceptable trade-offs, so you may need to make a compromise settling for reducing it (first during exposure and second during post-processing). But which compromises should you make? Knowing the types of noise that are produced in digital images and how they are produced will help guide you to solutions that will eliminate, reduce, or remove it.

There are three types of noise; random noise, fixed pattern noise, and banding noise.

Random Noise

Random noise appears as both luminance (light and dark) and chrominance (hue/saturation) variations not native to an image but produced by the electrical operation of a capture device. The electrical signal produced in response to photons is comingled with electrical variations in the operation of the capture device. Random noise patterns always change, even if exposure conditions are identical. Random noise is most sensitive to ISO setting. Again, digital cameras have one native ISO setting; higher ISO settings artificially boost the signal produced by the sensor and the noise accompanying it. The results? You get a brighter picture from less light and exaggerated noise. Since the pattern is random it is challenging to separate the noise from the image, especially texture, and even the best software used to reduce it through blurring may compromise image sharpness; how much depends on the level of reduction.

Fixed Pattern Noise

Fixed pattern noise (“hot pixels”) is a consistent pattern specific to an individual sensor. Fixed pattern noise becomes more pronounced with longer exposures. Higher temperatures also intensify it. Since the pattern is consistent, it can be easily mapped and reduced or eliminated.

Column & Row Noise

Banded noise is introduced with the camera reads the data produced by the sensor. It’s camera-dependent. Banding noise is most visible at high ISOs, in shadows, and when an image has been dramatically brightened. This type of noise quickly becomes obvious and objectionable; the regular row and column patterns from the sensor quickly call attention to the capture device; it is challenging to reduce without severely compromising image sharpness.

Noise can be broken down into two kinds; chromatic (hue/saturation variances) and luminance (brightness variances).


chrominance / color noise

Chromatic Noise

Chromatic noise produces a more ‘unnatural’ appearance, it is easier to reduce without compromising image sharpness than luminance noise. Chromatic blurring is less noticeable than luminance blurring, as human perception tends to see color contained within contours, even when it is not precisely true. It’s a convenient optical illusion. Larger chromatic variances may result from bayer pattern demosaicing. (Digital sensors typically capture photons with an array of two green, one red, and one blue photosites that register separate luminance values for each site. This data is then processed, ‘averaged’ if you will, to generate a final color, such as brown or lavender, or even a specific green, red, or blue. If done under suboptimal conditions, such as underexposure, larger areas of color variances may occur and will require additional post-processing. Extreme amounts of chromatic noise reduction may results in reduced saturation, especially along contours separating strongly contrasting colors.


luminance / light & dark noise

Luminance Noise

The presence of luminance noise is more readily accepted than chrominance noise. Luminance noise is harder to reduce than chrominance noise. Luminance information encodes contour, volume, and texture, key elements in representational images. Aggressive amounts of luminance noise reduction subdue image texture, creating a synthetic or overly smooth appearance, and blurs contours, lessening the appearance of focus. Camera noise reduction tends to be crude. Raw conversion software produces significantly improved results. For more extreme noise reduction, third-party software (such as Define, NoiseNinja, and NoiseWare) offer superior functionality and results.

Noise also varies in both magnitude and spatial frequency. Noise occurring over short distances has a high frequency (it’s ‘fine-grained’), while noise occurring over long distances has a low frequency. Noise magnitude, often described by the statistical measure of ‘standard deviation’, quantifies the variance a pixel will have from its ‘true’ value. Higher magnitude noise overpowers fine texture and becomes exceptionally difficult to remove.

The noise floor dominates other forms of noise. It is created by the type of read circuits in the sensor, the transistor characteristics, and support circuits such as the analog to digital converter. As light levels increase the noise associated with light (‘photon shot noise’) exceeds the noise floor. If the signal is increased by a factor of two (one f-stop), then the noise increases by a factor of one and the signal to noise ration increases by one. A higher signal to noise ratio makes noise less visible. When the signal exceeds the maximum value the sensor is capable of capturing (dynamic range is a measure of the largest ratio of the capture signal to the noise floor), the noise drops because the signal is pinned at the saturation value.

With a thorough understanding of what produces noise, how it is produced, what kinds and types of noise to be on the lookout for, you can take steps to reduce it at the point of capture. You want to start with as little noise as possible. If you want noise, you can always add it later, which gives you the possibility of customizing it with almost infinite precision. If you begin with noise in your originals, you’re locked in, and it can be challenging to reduce it without compromising image quality – sharpness, texture, saturation, and hue variety. Given that noise isn’t the only concern you balance, for some uses this may be an acceptable trade-off.


Reflection XXXVII

high noise / low noise

Noise happens. There is always some degree of noise present in any electronic device that receives or transmits a signal. Though noise is unavoidable, it can become so small relative to the signal that it is no longer visible. It’s all about a good signal to noise ration (SNR). The image is the signal. The capture and carrier mediums’ biproducts are the noise. (Film grain is noise.)The higher the SNR the more the image overpowers the noise; the lower the SNR the more the image becomes confused with noise. While some noise increases the apparent sharpness of images, the vast majority of noise found in images degrades quality. It’s best to avoid it.

There are a number of things you can do to reduce noise in your digital images at the point of capture.

1       Expose to the right

2       Use lower ISO settings

3       Use bigger sensors

4       Use faster exposure times

5       Keep equipment cool

Start here. Use software to reduce the appearance of noise only when needed. (Of course, reducing noise may not be the only deciding factor when selecting conditions, tools, and techniques. As with all things, you will have to balance many concerns simultaneously. Adhere to the above principles when practical.)

Knowing how noise is produced will help you avoid it.


avoid the red zone, caution in the yellow zone, go for the green zone


Underexposure results in more visible noise. Darker regions contain more noise than lighter regions in digital capture; the opposite is true of film. This is because the darker regions are recorded with fewer photons and less bits of data. It stands to reason that darker regions of images are produced with fewer photons. But why are they recorded with less bits of data? Digital cameras record data in a linear progression. If a digital sensor is capable of recording 14 bits of data (or 16,384 shades of gray) with a dynamic range of 8 stops, the lightest stop contains half the data in the file; the next lightest stop contains half as much data; and so on; it generates this progression from dark to light – 32 / 64 / 128 / 512 / 1024 / 2048 / 4096 / 8,192. This means you want to expose to the right. Avoid the lower 2 stops whenever possible; they contain less than 1% of the total data in the file. With so little data the signal becomes confused with the noise; the signal to noise ratio is very low. What’s more, when these tonal regions are adjusted, brighter or contrastier, the noise contained there quickly becomes more pronounced and with so few bits of data it also has a tendency to posterize. In addition, significant under exposure greatly increases the chances of producing Bayer pattern noise, a type of noise that is seen in larger areas and is more challenging to remove in post-processing. You’re much better off making a light exposure and darkening the Raw file during post-processing; this way your shadows will be represented with much more data and contain less noise. When light becomes challenging, which should you choose – under expose (without clipping) or raise the ISO? The precise cut off point is different for every camera, but in general avoid the lower 20% of the histogram. You’ll get less noise if you boost ISO and move the histogram up.


Higher ISOs amplify noise. ISO (International Standards Organization) is a descriptor that signifies absolute sensitivity to light. A digital sensor has one native ISO. Higher ISO settings simply boost the resulting signal. This is useful, but not ideal. When the brightness of the image is boosted, the noise is too.

Use “Shorter” Exposure Times

Longer exposures generate more noise. Hot pixels become hotter. All sensors have a few pixels that heat up faster than others, producing brighter than expected values. Some even have a few dead pixels that never fire, producing only black pixels. During longer exposures hot pixels are given more opportunity to heat up, growing brighter still; slightly hot pixels not visible at shorter exposure times become visible. As digital sensors age, hot pixels may become hotter and more pixels may become hot. Hot pixels produce a consistent fixed pattern of noise that can be recorded for given exposure times, making it easy to reduce. There is a duration at which a sensor suddenly produces a lot more noise; in older models this could be seen in exposures lasting only one minute, but with newer sensors this is greatly extended and most users will not encounter this; if you’re making very long exposures it’s useful test this and find the duration at which this happens for a your specific camera.

Use Bigger Sensor Sites

When it comes to noise, bigger is better – theoretically. Bigger sensors have more light gathering capacity, producing a higher signal to noise ratio, or cleaner images. More isn’t necessarily better. Cameras with more photosites (yielding more megapixels) packed into smaller areas tend to produce a lower signal to noise ratio, or noisier images. That said, a stronger signal does not necessarily guarantee lower noise. It’s the relative amounts of signal to noise that determines how noisy an image appears. The way a camera processes the file it makes has a significant impact on the final quality. Consequently, many medium format cameras that produce beautiful files in daylight typically produce noisier files than DSLR’s with smaller sensors in low light.

Stay Cool

High temperatures exacerbate noise. Thermal energy (leakage current) in semiconductors can generate an electrical signal that is difficult to distinguish from the optical signal. Ambient temperature increases leakage current by a factor of 2 for every 8 degrees Centigrade. Whether due to ambient temperature (You might start to see some effect over 90 F / 32 C degrees.) or storage (Don’t leave your camera in the sun, especially on a car seat, for long periods of time.), your camera can get hot and this will increase noise, which you can reduce by cooling off and keeping your camera cooler.

The best way to avoid noise is to produce as little of it as possible during exposure. These five tips will help you do just that.



Before & After Noiseware

Who doesn’t have noise? If you don’t run into noise in your digital images, at least once in a while, you may not be pushing the envelope enough. You can photograph long after dark; if you haven’t tried it, you owe it to yourself to experience this—it’s magical. And if you find you don’t have a DSLR on hand, this should be no reason not to make pictures with a point-and-shoot or cell phone.

Whether you’re using a cell phone, a point-and-shoot digital camera or a DSLR at high ISOs or with very long exposures, you’re bound to run into some noise. Noise happens. When you have it, there’s a lot you can do about it. There are many ways you can reduce noise during postprocessing; you could even say there’s an art to it. Learning these techniques can improve good exposures and save others.
If Lightroom and Photoshop fail to adequately reduce noise in your images, it’s time to move to third-party plug-ins. For years, they’ve done a superior job of reducing noise, and they still do. While there are many fine third-party plug-ins for Photoshop (Noise Ninja, Neat Image, Dfine, etc.), one stands out from all the rest: Imagenomic Noiseware Professional.

For me, Noiseware is the most robust noise-reduction software available. Ironically, while it offers the most sophisticated feature set, very often the default settings when you first open an image are all you’re likely to need. In many cases, very little, if any, additional tweaking is necessary.

In part, this is because Noiseware analyzes the images you process and creates “profiles” or saved settings that it uses every time you open a new image. It intelligently learns your needs by tracking your past images and analyzing your new images. You can also use Noiseware’s tools to create your own profiles, which can be saved and reused. You can save your own Preferences for how you’d like Noiseware to behave and learn. Noiseware also offers 13 default settings (like Landscape, Night Scene, Portrait, Stronger Noise, etc.) and allows you to save your own custom settings, which can be created from scratch or by modifying the provided presets.

PresetsGeneric Standard Presets


Custom Preset

Noiseware’s ability to target noise reduction to specific aspects of an image is what makes it unparalleled. You can adjust Noise Reduction based on Luminance or Chrominance; higher settings produce stronger noise reduction. You can target Noise Level based on Luminance or Chrominance; higher settings tell the software there’s more noise. You can target Color Range; Noise Reduction and Noise Level can be customized by hue—reds, yellows, greens, cyans, blues, magentas, neutrals. You can target Tonal Range; Noise Reduction and Noise Level can be customized for shadows, midtones and highlights. You can target image areas based on Frequency (or amount of detail); Noise Reduction and Noise Level can be customized to High, Mid, Low and Very Low frequencies. Finally, you can enhance detail, first, by using Detail Protection to reduce the effect based on Luminance or Color, and second, by using Detail Enhancement, which provides Sharpening, Contrast and Edge Smoothening.




Noise Level helps prepare the filter by analyzing the image

Noise Reduction is the blurring effect


Detail Enhancement – turn it off and use Photoshop instead


Frequency (of detail) targeting


Tonal & Color Range targeting

Noiseware’s ability to provide this level of selectivity is extraordinary. It allows you to easily customize noise reduction for separate areas of an image without making complex masks. You’ll want to do this. Here’s just one example, among many, of why you want to do this. Smooth image areas reveal noise much more readily and they support more noise reduction, while highly textured image areas hide noise, but don’t support as much noise reduction without compromising apparent image sharpness.

Use Noiseware’s sharpening sparingly (if at all) and only for the most modest boosts to image sharpness, as you can create much more sophisticated and selective results in Photoshop—and almost every image can use a little sharpening after noise reduction. Always reduce noise before sharpening.

Combine today’s digital cameras with the latest software, and you’ll find that you’ll rethink many things about when and where you make exposures. You’ll shoot at higher ISOs that you once thought were unusable. You’ll shoot in low levels of light where you once thought it was impossible to get an exposure, much less a usable one. You’ll look at your digital files, and where once you thought noise was a deal-breaker, you’ll find it no longer is. Noise-reduction skills and noise-reduction tools are essential to any photographers skill set and toolkit. Master them, and liberate yourself.




Lightroom’s Detail panel

Reducing noise in Adobe Camera Raw and Lightroom (the controls and results are identical) is easy.

The Detail panel provides tools to reduce two kinds of noise – Luminance (light and dark) and Color. Results can be targeted with the Detail slider into smoother (low setting) or more textured (high setting) areas. The effects for luminance noise reduction can be further modified by adjusting the Contrast slider; a higher setting affects only high contrast noise, while a lower setting affects even closely matched values. And finally, the effects for color noise an be further modified with the Smoothness slider, a higher setting creates a more aggressive effect.

Zoom into an image at 100% magnification and move the sliders until noise is reduced, but image quality isn’t compromised. Use restraint. In a majority of situations, it’s better to preserve a little noise than to blur the image substantially.

All noise reduction blurs images. Sharpening after noise reduction during RAW conversion is recommended. Knowing that you’ll sharpen an image after noise reduction, you may reduce noise slightly more aggressively initially.


no noise reduction


appropriate noise reduction


excessive noise reduction

There are limits to how far you’ll want to go. Noise can be so aggressively reduced that surfaces within images become textureless and begin to seem synthetically rendered with software rather than optically captured photographically. This effect may become more pronounced if contours are strongly exaggerated during sharpening. While sharpening, take care not to accentuate noise further. Develop a sensitivity for texture and contour, and use your best judgment. You know what things look like. Make your images look convincing to you, and you’ll quickly convince others.

RAW converter tools have limits. RAW converter tools do a good job with moderate amounts of color noise. Even high settings don’t tend to compromise image quality; sharpness, saturation and hue variety are all preserved. But sometimes they don’t go far enough. For aggressive noise reduction, especially for larger noise produced by Bayer pattern demosaicing, turn to Photoshop and possibly third-party noise-reduction software.
RAW converter tools do a reasonable job with luminance noise, but aggressive applications may compromise sharpness (some, but not all of this can be compensated for with RAW converter sharpening tools), and at times they don’t go far enough. When you encounter situations like this, turn to Photoshop and third-party noise-reduction software.
Most images can benefit from a little noise reduction and sharpening during RAW conversion. For many situations, this is all the noise reduction you’ll need. Many exposures don’t require substantial post-processing. However, some exposures require more power and finesse than these tools can deliver. When you encounter these, move to more sophisticated tools found in Photoshop and third-party plug-ins. But always start here.


Get the RSS Feed  

Subscribe by Email