Technique for Obtaining the Spectral Distribution of a Communication Transmitter Modulated by a Stochastic Process
Stanley A. Cohen, William T. Shelton · 1967
Transmitter spectrum models are being developed for systems modulated by typical random signals such as speech. Typical random signals are described by their statistical probability functions rather than the waveform description of deterministic signals. The resulting spectral distribution of the transmitted signal is vastly different from that of a single tone signal or even a signal of multiple tones. Since the waveform of a random signal is unknown, its Fourier transform does not exist and its spectrum is unattainable by way of the conventional route used for deterministic signals. Rather, a meaningful power density spectrum is available from the Wiener-Khintchine theorem via the intermediate step of autocorrelation. Such a procedure can be used to describe voice signals, TV signals and communication signals of other types. This method has been applied to conventional amplitude, phase and frequency modulated systems, and their single sideband counterparts. It has also been used to analyze a PAM/FM telemetry system.