A proposed analog speech scrambler based on parallel structure of wavelet transforms

The main aim of this study is to model and simulate a proposed analog speech scrambling system for analog speech signals, by using parallel structure of wavelet transformations.The simulated scrambling system consists of three main parts, i.e. transmitter, receiver and noisy channel parts. The processes that achieved in transmitter and receiver has an inverted nature. The performance evaluation process of the proposed system was taken into consideration via the using of noise power level (15 dB) in the simulated communication channel. The words "MATLAB" and " مساء " which pronounce be female were selected to be examined. The system performance is tested by using the following methods:
1. Subjective test, by playing back the original, scrambled, and descrambled signal to a number of listeners in order to estimate the intelligibility of the played signals.
2. Objective test, by using the (linear prediction coding, cepstral distance and segmental signal to noise ratio).
The proposed system investigates four types of wavelets (db2, haar,sym2 and sym4) each one with three different decomposition levels (such as Haar wavelet to correspond to Haar wavelet with the same level). Also, the system depends on the parallel structure by using two different types of wavelets on condition that they have the same level. In our system, Db1 wavelet along with Haar wavelet, Db2 wavelet along with Sym2 wavelet and Db4 wavelet along with Sym4 wavelet were used for the same level.In the proposed speech scrambling system, the algorithms are designed and implemented specific for the proposed scrambling system successfully by using MATLAB programming language version 7. The results of this work are promising , and they can be considered as a backbone for future programs, since for the first time the concept of parallel structure for transformation was introduced in the analog peech scrambling techniques that based mainly on using the transform domain philosophy. This approach provide more security for the resulted scrambled speech signal in comparison with the conventional scrambling approach that using only a single orthogonal transformation like FFT, DCT, or Walsh Hadmard transform (WHT), and the DWT in classical structure.