Security Analysis of Scalable Speech Coders

Scalable speech coders have been developed in order to achieve efficient speech communication over mobile ad hoc networks (MANETS). In practice, selective encryption is used with scalable speech coders to provide security of speech data. This paper analyzes the scalable ADPCM coders with selective encryption scheme where only core bits of coded speech are encrypted. The parameters of the coder have been identified by analyzing the selectively encrypted speech. The methodology developed is based on signal processing techniques. Autocorrelation method and power spectral density have been used to extract the bit rate and align the core and enhancement bits. Further zero crossing rate and Fourier Transform operations are used to detect speech-pause regions.