Digital Imaging Artifacts: Image Noise

Hello and welcome back to the second post in our Digital Imaging Artifacts series. Our first post in the series talked about Image Aliasing & Color Moiré. If you haven’t checked it out yet, you can do so right here. As promised, today we’re gonna be talking about Image Noise.

Noise in digital images has always been one of the most troubling dilemmas photographers have to deal with. Noise can be seen as low-frequency irregular arrangements of color patches (as illustrated in the image below), and can also be more bothering if combined with other image color artifacts.

Noise in digital photography can be thought of as grain in film photography. And sometimes, on print-out or slide, it actually is (as can be seen in the image below). Causes of image noise are numerous, and some of them cannot be easily identified. Those causes, however, fall mainly into two categories: Fixed Pattern Noise (FPN), and Temporal noise.

Fixed Pattern Noise

Fixed pattern noise is seen as a spatially fixed pattern across the whole image, and can be caused by 3 main factors:

1. Dark conditions: fixed pattern noise can be introduced in dark conditions, when there is no sufficient lighting and the photographed subject or scene is not well illuminated.
2. Long exposure: fixed pattern noise can also be introduced to your images when the exposure time is long, such as in night or low light photography.
3. High sensor temperatures.

The dark photographic conditions can also significantly reduce the available image sensor’s dynamic range. For those of you not familiar with Dynamic Range, you can check it out and learn more about it in my earlier post Photography and Dynamic Range right here.

Temporal Noise

Temporal noise is a variation in image noise from one image to the next. Pixel values vary randomly over time. Main factors that contribute to image temporal noise are: shot noise, reset noise, read noise, and dark current shot noise. These are measured by the ratio between signal to noise, known as signal to noise ratio or alternatively called SNR.

If the SNR value is high, the image quality is good and noise is at a relatively low level. On the hand, if the SNR value is low, image noise is at high levels that are more visible, degrading the image quality.

1. Shot Noise: Shot noise is most problematic under low-light conditions, where the SNR ratio is rather high. At high-light intensities, SNR ratio is high and shot noise is subsequently negligible.
2. Reset Noise: reset noise occurs at the sensor level, before the signal is converted from analogue to digital. Most camera and image sensor manufacturers are actually able to completely control reset noise by using a special signal processing circuitry.
3. Read Noise: read noise is noise added to the signal when it’s being read out of the sensor and converted from analogue to digital. Read noise is independent of exposure time. Slower analogue to digital converting and data processing cameras introduce higher levels of read noise.
4. Dark Current Shot Noise: dark current shot noise, which is noise associated with dark conditions previously discussed in FPN or fixed pattern noise can be largely controlled. The shot noise component of this kind of noise however, cannot be reliably controlled or removed.

Image Sensor Size & Image Noise

Smaller image sensors are likely to introduce more noise into your images, Especially in low light conditions when the ISO is pumped up to higher levels (400 and above). In high-light conditions, when the ISO is set at its lowest levels (80-100), image noise can hardly be noticeable. Larger image sensor produce less image noise than do smaller image sensors.

Noise Reduction

Most digital cameras perform some sort of internal noise reduction processing to the image. However, this processing does not actually get rid of all image noise, but rather control it a little.

You can also control image noise using special photo-editing software, or even using adobe photoshop’s noise reduction filters. It is worth mentioning that high levels of noise reduction can lead to losing fine details or edge details of your image. So you need to be careful not to go over the limit when trying to edit your image noise, for its better to have some level of noise rather than lose image details.

For those who are not familiar with noise reduction techniques, I’ve previously written a step-by-step tutorial that will guide you through the process of reducing image noise in adobe photoshop. You can check it out right here.

This concludes our second post (Image Noise) in the Digital Imaging Artifacts series. Next up, we’re gonna be talking about the third kind of digital imaging artifacts called Blooming and Clipping. Stay tuned…