Study And Implementation Of Speech Interference Method Based On DSP

The implementation methods of telephone security in the secure communication were studied and a homology sound scrambling method for secure communication was proposed in this paper. Compared with white noise interference, the proposed method has higher security, because it can’t obtain the original speech semantics by increasing the SNR of the signal (such as filtering). The principal study work is listed as follows:(1) To solve the problem of privacy analog speech communication, a homology sound scrambling method was proposed. By real-time sampling and splitting the original speech, lots of speech fragments were derived and stored. Those speech fragments were then extracted randomly and mixed to the original line as interference. The feasibility of the proposed method was validated by simulation.(2) The speech fragments extracting method based on entropy function was put forward to resolve the problem of speech splitting.By using the amplitude entropy and spectral entropy algorithm, each frame of speech was classified into acoustic/silent class.Saving the successive frame of acoustic class respectively, lots of speech fragments were derived and the speech splitting problem was solved. The simulation experiment results showed that the proposed method, compared with the method using amplitude entropy, was more effective in speech splitting.(3) The hardware platform based on TMS320VC5509A and CCS Integrated development environment was established and the speech real-time scrambling secure communication system programming was developed. The measured result indicated that both the software and the hardware platform implemented the homology sound scrambling method (including speech fragment splitting, extracting and mixing).