Tutorial: Digital Image Noise Reduction

Digital Noise

Any digitization method introduces some amount of noise into the resulting digital image. The value of each scanned pixel is determined by one or more light sensitive photodiodes that convert the light which is shown through your film into an electrical signal which is processed into the color value of the pixel. Due to the nature of the sensor design, if the same pixel would be exposed multiple times by the same amount of light, the resulting color values would not be identical. Instead, there is a small statistical variation in the pixel values. We call this variation "noise". The result of this noise is an uneven appearance in the scanned images when examined very closely, particularly in areas such as sky that have very even coloring. Noise is more noticeable in uniformly dark or light areas. The grain structures of the film can also cause the scanned image to appear noisy.

Noise Reduction

We apply noise reduction techniques to scanned images to reduce the effect of particular grain structures in some films. This benefits many films when the "traditional" use of GEM does not. However, applying these techniques blindly or too aggressively can do more damage than good. Like ICE, ROC and GEM, we apply these techniques only when they will benefit your particular images, based on our experienced eyes and careful analysis.

Let's look at two example images showing the benefit of these techniques:

Before and after versions of two photos showing the benefit of noise reduction techniques. Noise reduction techniques can have drastic effects (and these examples are not unusual).

We do not charge extra for this service. Our goal is to give you images that you can use without doing a bunch of work.

[We describe ICE, ROC and GEM here.]