A Line Spectrum Pair (LSP) Based Algorithm For Helium Speech Enhancement

Helium speech enhancement is one of the essential conditions that ensure diving carries working on the deep sea. Although more work has done by the foreign scientists, few works has done in our country. Furthermore with the development of digital speech processing and system on chip technology, more and more working has done in the Helium speech enhancement. Also the deep sea diving has developed in our country, it needs to get the core technology in the Helium speech enhancement.Helium speech is researched based on the classical acoustic tube theory of speech production, some important conclusions used for designing helium speech enhancement algorithm are drawn in this thesis. The transformation of the formant central frequency and bandwidth is the main causation cause the Helium speech. So an algorithm based on LPC is discussed in detail at first in this paper, in which the formant central frequency and bandwidth can be controlled independently by direct operation on the poles of all poles model of vocal tract filter.But the intelligibility of Helium speech with the algorithm based on LPC is enhanced not very well. After researched with LPC algorithm, A Line Spectrum Pair (LSP) Based Algorithm for Helium Speech Enhancement is proposed in this paper. Line Spectrum Pair is another form with LPC, belong to the frequency parameter. Line-Spectrum Pair associated with speech spectrum leads to synthetic filter based on LSP will never degrade system steady-state performance. Moreover LSP are important parameters of vocal track model in speech coding, their quantization are closely related to synthetic speech quality.New algorithm for Helium Speech Enhancement controlled the formant central frequency and bandwidth independently. Moreover it has more practical value for its better quantization character, interpolative character and degree of sensitivity. After researched the result data, the author ascertained the universal LSP coefficient in the new algorithm. By compared with the LPC method, it can be seen that our algorithm can correct helium speech and the intelligibility is enhanced greatly by independently controlled the formant.