The cause of sensor noise

I'm back after my long hiatus due to heavy school commitments. Yes, JC life is real tough, especially with useless subjects like project work. Anyway, the purpose of this post is to discuss the causes of CCD (Charged Coupled Device) noise. For those that do not know, sensor noise is what causes grainy images when pictures are taken at high sensor sensitivities, for example, ISO1600.

Sensor noise is more visible in images with uniform areas of colour and brightness

To understand how sensor noise comes about, we must first understand how a CCD captures an image. Light is made up of discrete particles of energy, photons. When these photons come into contact with the photosites, aka pixels of the CCD, a voltage is produced. More photons reaching the photosite result in more voltage produced. As such, the camera is able to distinguish bright areas from dark areas, to create an image.

How is the noise produced? Basically photons are quantum particles, they adhere to the laws of quantum mechanics. It is not possible to count the number of photons hitting the photosite with 100 percent accuracy , there will always be a certain amount of uncertainty. This is one of the fundamental laws of quantum mechanics, the uncertainty principal. This uncertainty creates what we call shot noise. One way to counter such noise is to increase the amount of photons used to make the image. This can be done by simply increasing the exposure, letting more light into the camera. With more photons, the noise will increase but at a slower rate than which the signal increases. Thus, the signal to noise ratio increases, resulting in a cleaner image which is less noisy.

Another major factor of sensor noise is sensor read noise. After photons are captured by the photosite on a CCD to produce a voltage, the voltage is then amplified by an amount proportional to the ISO gain set in the camera, and digitized in an analog-to-digital converter (ADC). Each step in the signal processing chain from counting the number of photons captured by the photosite, to amplification of the voltage, to digitization results in noise added to the image. None of the electronic components are perfect. All of them contribute noise to the final image.

One other facet of the sensor noise problem is thermal noise. At the crux of the issue is that at higher temperatures, electrons have a higher kinetic energy and are more likely to be accidentally released by the photosite . This causes the voltage produced by the photosite to be higher than it actually is supposed to be. To alleviate this problem, one should not take extremely long exposures as the thermal noise increases proportionately with the length of the exposure.

And that's all for my discussion on sensor noise, hope you enjoyed it. For those of you who would like to find out more on this technical subject, feel free to visit http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/index.html#shotnoise to find out more!
Look out for my next post on my new in ear canal phones, the ATH-CK52.

To cleaner images,
Bryan