Generating method for shielding signals used for protecting Chinese speech privacy

The invention provides a generating method for shielding signals used for protecting the Chinese speech privacy. The method comprises the following steps that (a) a sentence probability table is obtained through statistics; (b) a sentence paragraph probability table is obtained through statistics; (c) a phase probability table is obtained through statistics; (d) a Chinese character probability table is obtained through statistics; (e) a syllable probability table is obtained through statistics; (f) text information is generated according to the sentence number in a definite paragraph, the sentence paragraph number in the sentence, the phase number in the sentence paragraph, the Chinese character number in the phase and the syllable in the Chinese characters; (g) sound synthesis is carried out. The generating method for shielding signals used for protecting the Chinese speech privacy has the advantages that the meeting room sound shielding requirements and the Chinese voice features are sufficiently considered, the traditional mode of adopting stationary noise and the like as the shielding signals is omitted, the shielding signals which do not have the practical measuring but are very similar to the normal speaking voice are generated on the basis of each statistics features on the words, the phases and the sentences in the Chinese language are generated. Compared with traditional shielding noise, the shielding signals have the advantages that various kinds of adverse influence on auditory sense can be greatly reduced, and the sound shielding effect is improved.