Eliminate Image Noise

One of the main problems photographers face today is digital noise in images. For many amateurs, noise is a very confusing word, which they associate only with sound. So what is image noise? What can you do to reduce noise in images? Let us try to find out.

Noise has nothing to do with sound, though the term is used mainly in audio. The association of the term noise with sound can also be explained with the basic understanding of the term. Noise can be explained as any unwanted signal that appears in a circuit. When these unwanted signals overwhelm the main signal, thus making it difficult to decipher, it becomes a real problem. In the case of sound, all background or ambient sound that makes it difficult for any meaningful audio to be clearly understood, constitute noise.

Similarly, image noise is any unwanted signal (or charge) that gets generated in photo diodes or associated circuitry. In normal circumstances, these noise signals are too weak compared to the signal or charge generated by the light from the subject to be of concern. But when the main signal is weak, these noise signals become significant and show up in the final image as grainy spots. We know that a digital image is formed by light striking the sensor and producing corresponding charges on the photo diodes. But when the light is low (as in shadows), the charges generated are weak. So random background signals show up in the shadow areas as grains or digital noise.

Properly Exposed Image, ISO 100 Underexposed image, ISO 100
   

Noise become prominent when signals are amplified. A point to note here is that in digital cameras, all sensitivities except the base ISO are the result of signal amplification. In case of low-light images, since background noise is significantly higher than useful signal, these also get amplified when choose a high ISO.

How do we reduce noise?

The primary method is to use the base ISO in most of the situations. Most cameras have the base (or native) ISO as 100. This eliminates unwanted signals at the capture stage.

Another way of reducing noise is to capture in Raw format with slight over-exposure. This is also known as exposing to the right (right side of the histogram represents over-exposure). But why Raw? Because Raw is an uncompressed format and hence has more details embedded in it compared to JPEG.

Overexposed Image from Raw, ISO 100

So it allows us to recover details more efficiently in shadows and highlights. A slight over-exposure (1 or 1.5 stops) will capture slightly more light so that the shadow areas also have enough image signal to suppress noise.

Under-exposed Correctly Exposed Over-exposed Raw
     

When you process Raw images, you can recover the shadow and highlight details in Adobe Camera Raw or another software.