Noise Reduction Through Multiple Exposures

As camera technology has advanced, so has the high ISO performance, resulting in images with low noise even in poor light. Today, you can use ISO values as high as a million which were thought to be impossible just a few years ago. However, there is nothing like a free lunch as they say!

When you use high ISO values, the side effects are high noise (that appears like grain) and reduced dynamic range (that is, the ability to capture the tonal range). This problem becomes more acute as the sensor size decreases. You can use noise reduction (NR) software either built into the camera or in post-processing, but both these result in the image becoming soft, with fine detail being lost. Using a tripod to support lower ISO values is a good solution, but there are times you simply can’t use one, like in a museum or on a street.

Look at the test scene (Picture 1). This scene was captured both at ISO 6400 and ISO 640. You can clearly see the difference in the noise levels in this 2:1 crop (Picture 2).

Is there a way out? Of course, buying the latest and the greatest camera every year to get access to bleeding edge technology is a way, but that is an expensive and an impractical solution. And even the latest cameras have limits though the boundaries may be moving upwards every year.

There is an alternate way to capture high ISO images with less noise and at the same time not lose any detail. The penalty? You need to take a lot more images and then put up with some post-processing. That’s about it. The rest of the article will explain how to achieve this. In the end you will be surprised that it is so simple! Plus, this technique is applicable to all cameras starting from Point and Shoot to full frame digital cameras. Before we get into the nitty-gritty, just look at a few of the images you have captured in the past, while using your camera at various high ISO values. Observe the noise in them carefully and decide on the highest ISO at which the noise is still acceptable to you.

This technique involves two steps. The first is to capture multiple images and the second is to post-process them. Let us look at these in detail.

Image capture: