American White PelicansPurple PetuniasSandhill CraneHorse's EyeAmerican BisonSparring American ElksAurora Borealis

Photography Talk
What's All the Noise About?
by Bob Dean

Back in the days of film photography, we worried about grain in our images. It was the wisdom of the time that the faster the film, the more pronounced the grain and the less sharp the image. Exposure times, film processing parameters and such had some impact but that was really a second order effect. The guys in the green and in the yellow boxes were constantly working to improve grain structures to reduce the impact of faster speeds on image quality. That was then, this is now.

The digital revolution has introduced the photographer to the wondrous technologies of electronics and digital signal processing. But, one of the things we now must deal with in our photography is noise, or as some authors call it, digital grain. Let’ discuss its sources and how to deal with it.

What is noise? It’s all of the additional extraneous signals that come from the sensor, the analog electronics and the digital circuits in the camera that show up in our images (and usually have a negative impact on the image). This is particularity visible in images of large single color such as the sky.

Noise has a number of sources. These range from dust on the sensor to the intrinsic errors in converting an analog signal to a digital data stream.

Dust on the sensor is the easiest to understand but can be quite difficult to control. Point and shoot cameras have a pretty good sealed environment around the sensor so dust is normally not a problem, but if you do get dust problems, correcting the root cause can be tough. Digital SLR’s are much more prone to sensor dust but most have a “clean setting” that allows the photographer to blow away dust (carefully!). The other sources of noise are a bit more technical, but let’s start with light striking the sensor. Light comes in little packets call photons. When these little guys strike the sensor they are converted to electrons which are then used to generate the electrical signals containing image data. Since the sensor is warm (that is at a temperature above absolute zero) other “thermal” electrons are generated just by the heat. These electrons create a baseline signal (noise) that contaminates the image information. The signal coming off of the sensor is very small and cannot be effectively used by the circuits that convert the analog data to digital information. To remedy this, analog amplifiers are used to boost the signal. The background noise is amplified as well as the image data plus the amplifier adds its own noise contribution. The amount of amplification is controlled to a large degree by the ISO setting on the camera. The higher the ISO, the more the noise contribution. Higher ISO means the camera will use less light to generate the image. Less light means less signal from the sensor (fewer photons converted to electrons), more amplification is needed and the ratio of signal (image data) to noise is less. This shows up as higher noise or grain in the image.

Another noise source, one that’s easier to relate to film grain is the inherent nature of digital imaging. The process of producing a digital image means we are taking an infinitely variable light spectrum and converting it to discrete pieces of information. The first place is the sensor itself, made up of a fixed number of sensors, not unlike the grains in film. The analog data generated has, as we’ve seen, a noise component. This signal is then once again digitized into discrete pieces of data for processing. Each time this analog to digital conversion takes place, information is lost and edge detail is compromised. When the signal is then compressed into a jpeg file, more information is lost.

So, we have a lot of noise sources. How can we deal with them? The component and camera manufacturers have done very good things to reduce the impact of noise on our images. The first is the development of larger and larger sensors. Larger image sensors have more light gathering capability with a constant noise level. A 6 mega pixel sensor that is APS size generates more effective noise that a larger full frame sensor. The second, and probably most important improvement, is the noise reduction software in the camera itself. This “post processing” activity in the camera (with the RAW data) actually removes a lot of the noise from the signal, producing a cleaner and more grain free image.

What can the photographer do to reduce noise? Well, first use a camera with the biggest sensor you can afford. Unfortunately for the wallet, this usually means a digital SLR. If that’s not in the budget, then shop for the camera with the biggest sensor you can get and do research on the quality of the camera manufacturer’s noise reduction software.

With any camera, the following tips will also help.

1) Shoot at the lowest ISO setting you can for the conditions. Remember, this means less amplification and more light converted to image data.

2) Try to avoid very long exposures. The longer the shutter is open, the more total sensor noise is generated.

3) Use the lowest compression setting you can (high resolution jpeg has less inherent noise than low resolution – but it takes more storage space).

When you do get your images on to the computer, there are several really good noise reduction software packages available. Do some research and see what you can find. More are coming out every day.

Dealing with digital grain is really not that much different from film grain. Learn the causes and the remedies and keep on shooting.