论文部分内容阅读
本文介绍基于时长为N的样品重排的语音置乱器。业已发现,“均匀”(U)变换(由地址映射i→ki(模N)确定,K从1到N)在破坏语音的可懂度上与伪随机(PR)变换一样有效。分析结果表明了输入信号与已扰乱信号频谱之间的关系,而计算机模拟对PAM样及ADM,APCM,ADPCM码的打乱效果进行了比较。打乱效果是按上述胜序递增的,且ADPCM码的时延可小到1～2mS。最后,将这种置乱与倒频作了比较倒频相当于对波形的奈奎斯特样品每隔一个进行符号变换。
This article describes a speech scrambler based on sample rearrangement of length N. It has been found that a “uniform” (U) transform (determined by address mapping i → ki (modulo N), K from 1 to N) is as effective as a pseudo-random (PR) transform in destroying speech intelligibility. The analysis results show the relationship between the input signal and the disturbed signal spectrum, while the computer simulation compares the disruptive effects of PAM samples and ADM, APCM and ADPCM codes. Disruptive effect is increasing according to the above sequence, and ADPCM code delay can be as small as 1 ~ 2mS. Finally, scrambling this scramble and scrambling is equivalent to symbolically transposing every other Nyquist sample of the waveform.