CAJ | 학술논문

为了实现语音的端到端安全加密通信，提出了一种新的可通过RPE-LTP压缩编码的语音加解密算法。该算法基于DCT变换域，在发送端对语音信号先后进行频域置乱和时域置乱，接收端进行解置乱和重建信号。运算过程限制在实数域，不会涉及相位的估计问题，避免了语音数据的损失。仿真结果表明，加密算法实现了加密的语音信号非噪声化，保持了语音信号的特性，满足通过声码器进行编码的信号输入要求。进一步降低了加密后的语音可懂度，提高了解密后的语音质量和信号重建率。
위료실현어음적단도단안전가밀통신，제출료일충신적가통과RPE-LTP압축편마적어음가해밀산법。해산법기우DCT변환역，재발송단대어음신호선후진행빈역치란화시역치란，접수단진행해치란화중건신호。운산과정한제재실수역，불회섭급상위적고계문제，피면료어음수거적손실。방진결과표명，가밀산법실현료가밀적어음신호비조성화，보지료어음신호적특성，만족통과성마기진행편마적신호수입요구。진일보강저료가밀후적어음가동도，제고료해밀후적어음질량화신호중건솔。
For the realization of the end to end secure encrypted communication of voice,this paper proposes a new voice encryption algorithm,which can penetrate RPE-LTP compression encoder.In the sender,the algorithm scrambles with the voice signal in frequency and time domain based on DCT transform domain .In the receiving end,it decrypts and reconstructs the signal.Opreation process is limited in real number domain and does not involve the phase estimation problem,it avoids the loss of voice data.The results show that the encryption algorithm can realize the non-noise signal of the encrypted voice,maintains the characteristics of the speech signal,meets the requirements of input signal to penetrate RPE-LTP compression encoder.Reduce the encrypted speech intelligibility further,improve the quality of the decrypted voice.