WHAT IS S/N RATIO & WHY IS IT IMPORTANT?
WHAT IS S/N RATIO & WHY IS IT IMPORTANT?
By Frank McClatchie
The Signal-to-Noise Ratio specification sometimes noted as “S/N Ratio”, “SNR”, or just “S/N” for cameras, amplifiers, and wireless equipment is sometimes overlooked and miss-understood. The S/N Ratio notation will show up in many specification sheets for the equipment in a CCTV system, so it must be important, but what does it really mean and how does it effect your installation?
The specification of S/N Ratio indicates exactly how much noise you can expect to see superimposed upon your picture. The S/N specification is written in logarithmic notation expressed as some number of “dB” deci-Bells, so very large ratios can be written with a small number of digits. For instance, a S/N ratio of 60 dB would indicate that the signal power amplitude is 1,000,000 times greater than the power level of the noise, and also the signal voltage is 1,000 times greater than the voltage of the noise that is present.
But first where does noise come from? Noise is caused by heat and the absents of all heat is know as Absolute Zero the null point of all thermodynamic temperature scales. It is defined as zero K on the Kelvin scale and as −459.67°F on the Fahrenheit scale. It is the theoretical point at which atoms stop moving. At absolute zero (zero Kelvin) there is no noise. The amplitude of noise goes higher as the temperature goes higher. The atoms that make up the semiconductor amplifiers and all the conductors of electrical energy within the product, even the atoms that make up the cables that connect products together move in a random manner increasingly as the temperature goes up. At room temperature, the power level of the noise is X. Most Cameras, Amplifiers, and active (powered) devices have a higher noise output than the temperature of the device would stipulate. This noise power must be overcome by a large margin in the picture signal to reduce the visible “snow” in the picture.
The visual effect of poor S/N ratio is to cause the picture to seem like it was taken during a snow storm. The lower the S/N number, the more “snow” will be seen in the picture. The picture may stand still, but the “snow” is constantly moving about in a random manner. If the interference has some structure or shape to it, it is not noise. Noise is total random and has no repeating sequence. All pictures will contain some “snow”, but pictures with greater than 62 dB S/N, will have so little “snow” that it will not interfere with the picture. Once the “snow” appears, the only way to remove it is to reduce the bandwidth of the video signal, but that also reduces the picture definition, so a choice must then be made. Is the reduced “snow” worth the reduction in picture definition? The magnitude of the “snow” in a picture will relate to the actual bandwidth of the device through which the picture is being transmitted. In AM transmission systems, the increase in the noise power of the “snow” is directly proportional to the bandwidth of the transmission system that is reducing the signal level. The same could be said of the bandwidth reduction provided by a low pass filter (the frequency of the 3dB loss point).
There are other ways to reduce the “snow” in the picture, but they entail reconfiguration of the connection between the Camera and the Recorder. Most of the “snow” that creeps into a picture after the Camera generates the video signal is cause by the fact that all systems such as Coaxial Cable, Unshielded Twisted Pair wire, Fiber-Optic, Microwave, and Wireless Radio connections will cause the original video signal to decrease in level, and thereby close the Ratio between the signal (now at a lower level) and the noise which is present everywhere with (no reduction of level).
If you have two components in tandem that have no gain or signal amplification, then the output of the second device will display a noise level 3 dB higher than the first device, and thus suffer a 3 dB reduction in S/N ratio. If the first device has a gain of 20 dB, then the second device will have a noise level 20 dB higher than the first device, but any signal going through these devices will also be 20 dB higher, therefore will not suffer from the 3 dB decrease in S/N ratio that the first example suffered. If the gain is a lot less than 20 d/B, the second device will add to the total noise level according to the rules of power addition.
There are other ways to reduce the noise in a video image and that is by reducing the temperature of the environment (a whole lot), or by reducing the bandwidth of the video signal, or by amplifying the signal before it is reduced to too low a level by transmission losses. The first solution reducing the temperature, is impractical in most environments, the second solution reducing the bandwidth, is possible but only if you are willing to accept the reduced picture definition that accompanies bandwidth reduction. That leaves the third possibility, amplifying the signal to keep it well above the noise level which is forever present at any temperature above absolute zero. Restoring the level of a video signal at the receiving point in direct proportion to the loss caused by the cabling system is the best method to go the distance. Selective amplification is used to completely equalize the natural cable slope loss of the cable span. To equalize a length of cable, the gain of the amplifier must be greater at high frequencies than it is at lower frequencies. If the gain of the amplifier matches the loss of the cable at all frequencies then there will be no loss of picture definition as the length of the coaxial cable is increased as long as you can compensate for the loss.
Longer transmission distances can be attained, by adding equalizing amplifiers at optimum distances along the cable to amplify and correct for the cable slope loss before the S/N Ratio becomes un-usable. A margin of 3dB below the minimum allowable S/N ratio is the point at which the amplifier is placed for optimum performance. You can also gain an advantage by Pre-emphasizing the video signal before it starts its trip down the cable.
The bandwidth of the video signal will affect the reading on any meter but doubling the bandwidth will only increase the noise measurement by 3 dB, so small differences of bandwidth do not affect the accuracy of the S/N measurement very much.
Remember that video circuits are always terminated with 75 Ohms, so any S/N Ratio measurement must always include a 75 Ohm termination at the end of the line for an accurate measurement.
