Degradation from Noise Reduction


Today’s Question: What is the impact of noise reduction in terms of the degradation of image quality?

Tim’s Quick Answer: Noise reduction can reduce the apparent sharpness and level of detail in a photo and can also cause colors to become less saturated and potentially bleed into surrounding areas within the image.

More Detail: Noise reduction essentially involves averaging neighboring pixel values in an image to reduce the appearance of noise. The specific process is a bit more complicated than this, but if you think about it as a process of averaging neighboring pixel values you can get a better sense of the potential degradation in image quality that can result.

Luminance noise reduction can be particularly problematic, because it involves averaging tonal values in an image. This can quite literally involve reducing contrast between neighboring pixels, which reduces perceived sharpness and can also result in a loss of fine detail. It is therefore critically important to be very careful when applying luminance noise reduction. You will need to compromise between the level of noise reduction being applied and the degradation in image quality.

With color noise reduction the risk relates to altering color values. With modest color noise reduction, the primary risk is that colors may become less saturated. However, with strong color noise reduction you can also create a situation where colors bleed into each other. For example, if you have a photo with vibrant green leaves on a tree set against a bright blue sky, along those edges you may see the greens blending into the sky and the blues blending into the leaves.

The key is to carefully evaluate the image, and to only apply enough noise reduction to mitigate the effect of the noise without using such strong settings that the image quality is significantly degraded.