It is becoming increasingly important for authorities concerned with the generation, distribution, and radiation of television signals to have available an accurate method of measuring random noise in the presence of a signal. This monograph describes two realizations of a simple method which is based upon sampling the random noise in the known minimum-energy regions of the video spectrum. The first has been used quite extensively and has proved to give very consistent and reliable, while the second and more sophisticated version us at the moment still in the course of development.

The advantages of the method are: a) The apparatus is simple and state and the measurements can be repeated to within 1dB. Also, no undue degree of skill required on the part of the operator. b) Both the noise spectrum and the total noise power in the band frequencies are measured at the same time. c) Measurements can be made accurately in the presence of pilot control frequencies and interference of a relatively continuous nature such as pick-up from broadcast stations and mains hum. Furthermore, the accuracy is not impaired by relatively large amplitude non-linearity such as is encountered, for example, in gamma correction amplifiers. d) One is not restricted to a particular test signal of an artificial type but, within limits, measurements can be made during the transmission of a wide range of picture signals. Also, it is simple to measure the noise level at various amplitudes of the picture signal.

The disadvantage is that there is a minimum noise level, depending upon the type of signal present and the frequency of measurement, which can be measured to a given degree of accuracy. This limitation, however, has not been found at all restrictive in practice and, in general, an overall accuracy of approximately ±1 dB can be expected.