Improvement in the quality of speech received at suppressed carrier SSB

Abstract

A receiver should tune to the carrier frequency accurately at suppressed carrier SSB (SCSSB). Manual tuning, however, is very difficult due to the lack of the carrier signal. If fine tuning fails at reception, the quality of received speech is degraded owing to the shift of speech spectrum. This article proposes a method to improve the quality of speech degraded at SCSSB. Received speech is transformed into cepstrum. The remarkable peak on the quefrency indicates the fundamental period of speech. If tuning is carried out correctly, the shape of the peak is symmetrical. Mistuning gives an asymmetrical peak. Furthermore, the shape is different to the direction of mistuning. Employing this feature, the mistuning frequency is estimated. Finally the spectrum of the received speech is shifted and the naturalness is improved by employing the overlap and add method.