This is the first post in a short series looking at Digital Noise in an image and how to reduce it.
You have been out shooting all day, you have been previewing your images and you know have some crackers. You head home plug in your camera, download the images and then look at your brilliant shots. Your heart drops as you see your images are covered in strange pixelated marks over your lovely smooth tones. Yes, you have noise.
I remember when I was going over to digital, noise was something I didn’t think about. I thought it would be great shooting digital, digital is clean, I would have lovely smooth tones, no more film grain and being forced to shoot at a low ISO.
Oh how my dreams were broken.
Film Grain was something of a by-product that you lived with and worked it into your images as an aesthetic. I remember shooting 3200 ISO black and white film for a fashion shoot to get a grungy look, in these dark damp abandoned buildings. Well noise is sometimes likened to grain and it is a good analogy. Noise is the digital version of grain, but we spend more time trying to suppress noise in our images than we did film grain.
Digital noise in photography can be more intense depending on several factors from sensitivity (ISO) the length of the exposure and even the temperature.
To start with noise is caused by the signal that the sensor receives. All electronic devices that either transmit or receive a signal will have noise. When watching TV this signal is sent down a cable or via an antenna and received by a decoder or another antenna and carried to the TV. In the case of a camera this signal is light coming through the lens and hitting the sensor. With a high noise to signal ratio, the noise is barely noticeable, whereas with a low ratio it can be hard to distinguish the noise from the signal.
A camera’s ISO setting controls the sensitivity of the cameras sensor. Usually ISO go up by a factor of 2 starting at 100, then 200, and then 400 ect (some cameras opt for an ISO setting of either 50 or 80 as well). This means is that an image taken at 100 ISO will have an exposure of half that of an image shot at 200 ISO. This is great but as you go up to 200 ISO the amount of noise is amplified. The amplification of noise is because as you raise the ISO you are amplifying the signal in the camera. Raising the ISO is like turning up the music it just gets louder.
Digital cameras create three different types of noise;
Random Noise, which looks like slight pixelation, it is a fluctuation in pixels above and below the image intensity. There will always be random noise, it will increase as the ISO is raised. The pattern of random noise changes and will never be the same even when two images are shot with the same settings. Random noise is the most common form of noise that will be found in digital photographic images.
Fixed Pattern Noise, consists of Hot Pixels, which are pixels that far surpass that of the random noise fluctuation. This phenomenon normally is greater with long exposures and is increased when shooting in hotter temperatures. Fixed Pattern Noise will show the same pattern even if two images are taken with the same settings, the opposite to that of Random Noise. Fixed pattern noise is less of problem with modern digital cameras.
Banding, is dependant on your camera. It is created by the camera when it reads the signal from the digital sensor. Banding is most visible in high ISO Images, in the shadows or when an image has been aggressively brightened. I have experienced banding more in photoshop than in camera but it is dependant on the make and model of the camera.
Fixed pattern noise is the easiest of the noise to remove. The camera has to learn to recognise the pattern and then replace the pixels to create a clean image. Fixed pattern noise is much less of a problem in the second generation of digital cameras than Random noise.
Random noise although is less objectionable than the other two types it is also the most difficult to remove from an image without degrading the overall image quality. Computers find it hard to distinguish what is digital noise and what is a pattern or a texture in an image. When attempting to remove random noise you also have a tendency to remove the textures in the image as well.
In the coming posts I will discuss different approaches to removing noise.