• No products in the cart.
  • No products in the cart.
Scroll to Discover
back to top
Image Alt

Kit Konnect

Background noise (or floor noise) refers to an unwanted disturbance in an electric signal. There are several noise types and they can be generated by different effects. It is always present in an electronic signal, whether it is a video, audio, or any other electric signal.

The Signal-to-Noise ratio (or S/R ratio, or SNR) is a measurement comparing the amplitude of the desired signal to the amplitude of background noise on top of it. It’s usually expressed in Decibels (dB); In video, it can vary from 40 dB (VHS quality) to 60 dB (Broadcast quality) and above.


Why does this matter and what does it tell us?

The analogue signal is inconveniently fragile and stages such as A/V processing, recording and broadcasting deteriorate its quality rapidly. So, the noise element is directly linked to the quality of a signal/image. Now let’s break things down…


What is a Decibel?

It is a unit of measurement that expresses the relationship between two physical values1 on a logarithmic scale (different than a linear scale). 

Here’s where it started: the acoustic power which a human ear can perceive, from the lowest hearing threshold2 (1 microwatt/m²) to the one of pain (1 megawatt/m²), can take ratio values of up to 10123. Therefore, acousticians have developed a notation in the “power of ten” in order to simplify things and avoid manipulating huge strings of numbers, expressed in the following notation: 10.log(P2÷P1). The Decibel notation was then extended to ratios of voltage, which can express amplifier gain, transmission line losses or signal-to-noise ratio (defining the quality of a signal). In this instance, its notation is 20.log(V2÷V1), as a voltage ratio is twice the equivalent of a power ratio.


Analogue vs. Digital signal

It is important to understand that an analogue signal represents the most naturally and accurately a physical phenomenon but is particularly vulnerable when going through the processing and transmission phases (as well as variations in temperatures, vibrations, ageing…), which in turn acts in degrading the signal itself.

On the other hand, a digital signal is binary: information is combined into sequences of 1s and 0s. The disturbances caused by the various treatments do not harm the information held within a digital signal, as long as the levels of 1s and 0s are recognised appropriately.

Circuits working at high-frequency pace have been developed to process the analogue-to-digital signal conversion as early as possible after the birth of the analogue voltage outputted from the sensor’s amplifier(s). This means that, after its digital conversion, the signal shouldn’t face many problems (such as the introduction of noise) from the processing and transmission phases. That’s why cameras nowadays have great signal-to-noise ratios at their base ISO.


Signal types at the output of sensors

In general terms, a camera device generates an electric signal deriving from light collected from the lens. At the sensor’s level, a physical quantity (light=photons) is translated through the sensor and outputted into an analogue signal whose amplitude (expressed in Volts) holds the information relative to the captured scene. But there’s a difference;

  • for CCD chips, the signal in its output is analogue.
  • for CMOS chips, the circuitry is attached to it meaning that the analogue-to-digital conversion occurs at the sensor’s level, allowing the sensor to output a digital signal.

Subsequently, the signal is processed before being recorded, broadcasted and/or archived.


What’s the relation between a sensor and noise generation?

The output amplifier of a sensor/photosite generates noise (see our previous post: What You Need To Know About Camera Sensors [Part 1]).

A CMOS sensor contains as many amplifiers and converters as it has pixels4. Therefore, the circuitry introduces a low fixed pattern of noise (higher than a CCD chip) to the signal which when augmenting the gain/ISO to increase the exposure/brightness levels of a shot scene, it also increases the noise floor level.

The illustration below expresses the relation between the “pixel value” and “light intensity”. Yet, it is unsure for which type of sensor this representation accounts for, but it should help in understanding the relationship between these elements, noise and dynamic range. Here, the SNR is represented linearly rather than in decibels (logarithmic scale).

(Source: Understanding Dynamic Range And Signal-To-Noise Ratio When Comparing Cameras )


To help, here is an interpretation of the terms used in this illustration:

  • Pixel Value: Capacity of the sensor’s photosite (or pixel) to store photons (sometimes called “well depth”).
  • Light intensity: Light brightness (given for a given pixel).
  • Pixel Saturation: The point at which the photosite (or pixel) has reached its maximum capacity to absorb photons.
  • SNR: Signal-to-Noise Ratio.
  • Dynamic Range: The ratio between the lowest and highest quantity of light that the photosite (or pixel) can absorb.


Why does it matter?

In fact, the level of noise present in an image is usually a factor – amongst various others – defining its quality5. In that instance, noise is identified as ‘grain’. Take a look at this snapshot6:


Any cinematographer’s advice would be to use the cameras base ISO as a starting point (for optimal S/N ratio and exploiting the camera’s dynamic range at its best potential). That said, it’s quite likely that not all situations will allow you to do this. You may have to work your way up7 (e.g. low light environment) or down8 (e.g. extreme brightness) on this setting depending on the brightness of the given scene and your capacity to access and use accessories (e.g. filters, lights).

If you’re struggling to illustrate what this means, here is another analogy: turn on your speakers without playing any tunes through them and ramp up the volume to its maximum. You’ll hear a never-ending “ksshhhhhhhhhhhhh[…]” – that’s noise! Same as the grainy image above but from an audio signal. When the audio signal inputted to your speakers is weak, you’ll naturally increase the volume and the noise level at once, therefore, affecting the overall quality of sound outputted from your device. If you want a strong signal, increase the signal’s amplitude from the source (e.g. get more light to hit your camera’s sensor, or increase the volume playing from your device to your speakers).


Let’s wrap it up

So the bottom line is, as long as your signal holding the ‘true’ information (as referred to here as ‘captured signal’) is strong enough and well above the noise floor, then you’ll keep a clean and accurate image, as opposed to a grainy image.

Still got some questions? Let us know what’s on your mind in the comments.



  1. As explained above: ‘power of the captured signal vs. power of noise floor’.
  2. Which is used as a reference point for all other measurements.
  3. Note that we’re talking about power (Watts) ratios here
  4. As opposed to a CCD sensor which amplifies the signal at a single output (low noise)
  5. … depending on your artistic intentions, I suppose. However, if it was truly intended, it would probably be best to generate artificial noise in post-production as it offers more control over your image’s quality and intended result
  6. You’ll realise that I’ve added noise to this image in post-production (rather than being generated by the camera’s circuitry), and the exposure is constant. It’s just another example of what noise is for those who can’t picture it.
  7. Therefore amplifying the noise floor. Though some high-end professional camera circuits are so efficient that visible noise may be fairly negligible even at a high gain level.
  8. Therefore getting closer to the noise floor

Post a Comment

Subscribe to Kit Konnect and receive newsletters, offers and invitations. You can unsubscribe anytime. For more details, review our Privacy & Cookie Policy.