Digital images can be prone to noise due to the visual effects of electronic errors of the camera's sensor. Long exposures, higher ISO settings or higher temperatures can introduce noise, while some color channels â notably the blue channel - will be more affected by noise than others. This is because sensors are often less sensitive to blue light, and to compensate for this the blue channel is amplified more than others. JPEG compression can also amplify noise.
To reduce or remove noise, some companies produce (Photoshop) plug-in filters, while a few digicams feature built-in noise reduction methods.
Here is a technique to reduce noise caused by long exposures of 2 seconds or more,
which can manifest itself as fixed colored dots often denoted as stuck pixels.
These can be successfully removed by taking a âdark frameâ before or after the main shot, with the same shutter speed as the main exposure, and subtracting this from the original to replace the stuck pixels.
Open the two images in Photoshop and paste the dark image as a second layer. Apply a little Gaussian Blur and change Layer Options to âDifferenceâ. You will note that most stuck pixels are now invisible and the image has improved considerably.