All Digital camera suppers from a common malady, called "High ISO noise". What does it mean is, if you select a higher ISO speed for your sensor, your will get some extra digital noise signal. This noise manifests like tiny colored specs scattered all over the picture area. It is kind of like getting grains when you use high-speed films. This happens due to the fact Image sensor cells produce electron charge over time and depending on the temperature. Cooler the temperature, lesser the noise, higher the temperature, higher the noise. Unfortunately, even in normal ambient temperature, the existing image sensors produces appreciable amount of charges to degrade the image at low light and / or longer exposure. In technical term this is also called Dark count and Dark noise of the sensor.

Dark count is usually expressed in electrons per unit of time at a given temperature. It can also be expressed as a current per sensor area at a given temperature. Dark noise is approximately the square root of the dark count. Both of them are a function of temperature. So by cooling the sensor, it is possible to reduce the Dark count. In Astro-photography, telescopes are fitted with special sensors, which are cooled externally to virtually eliminate Dark Count.

We are not so fortunate, as our Digital cameras are not equipped with micro Refrigerators! But there are few things we can do to minimize this annoying effect.

• The first and foremost effort should be to minimize the usage to ISO higher than 200. In Canon Digital SLR, this can be achieved by using a high speed lens and / or using lens equipped with Optical Image Stabilizer. The other option is to use a tripod
• If ISO 400 or higher is used, try to overexpose by 1/2 stop, which you may compensate in Photoshop later. Remember, darker the image, higher will be the noise.
• While post processing in Photoshop, try Despeckle and Noise filter.
• Noise distribution is not uniform across RGB channels. Some channel tends to be more noisy than others. So first inspect individual RGB channel and apply the above filters only in the noisy channel. This will help to preserve image sharpness to a greater extent.
• Those filters makes the image soft. So you may localize the filtration by first selecting the noisy area with a lasso tool.
• An excellent software tool named Neat Image works great on removing noise. Download the working evaluation copy from www.neatimage.com
• Use Unsharp mask carefully on a noisy picture, as noises will become more pronounced. So you should first reduce noise and then apply the Unsharp Mask.

The example below illustrates how a noisy picture is recovered by NeatImage.

Original Image with Low light, High ISO noise

Noise eleminated by NeatImage