Power requirements for speech communication systems

This paper presents a simple method for predicting the power required for audio, AM, DSBSC, and SSB speech communication systems which are peak-power limited and are operating in the presence of "white" Gaussian noise. The effects of simple speech processing such as pre-emphasis, filtering, volume compression, symmetrical peak clipping, and frequency translation are considered both singly and in combination. The method is based on empirical data and is directly applicable to many practical speech communication problems encountered by the radio engineer.