Font Size: a A A

Research & Implementation Of DSP On Speaker Recognition

Posted on:2007-01-16Degree:MasterType:Thesis
Country:ChinaCandidate:C Y HeFull Text:PDF
GTID:2178360182980495Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Speaker recognition technology belongs to biometrics recognition technologies. Speaker recognition is the process of recognizing who is speaking on the basis of individual information included in speech signals. Compared with other kinds of biometrics technologies, speaker recognition has many advantages: convenient, natural, accurate, economical and well expansibility. Speaker recognition plays an important role in many areas, and has extensive applied foreground.This thesis introduced the analytical and processed foundation knowledge of the speech signal first. And then, the fundamental theory of speaker recognition is discussed. The thesis develops a DSP-based speaker recognition system, and introduces the design of hardware circuit and software development flow in detail.The main process of speaker recognition goes along as following: pre-emphasis, drawing the characteristic of speech signal, pattern matching. Linear prediction coding analysis (LPC) is used to draw the characteristic of speech signal. The system in this thesis is a small- vocabulary speech sound control system. Allow for the feature of textual system, dynamic time wrapping(DTW)can meet the requirements of the system, so I select the DTW for arithmetic. The kernel algorithm of the whole system is LPC and DTW.In the hardware circuit, TI Company's TMS320VC5402 DSP is used as its core circuit, which can ensure the real time realization of the speaker recognition task. This thesis introduces the characteristics of DSP and the configuration of TMS320VC5402, and develops a DSP-based speaker recognition system. This thesis stated the structure of the system and analyzed the working progress of the system. In the whole system, in which TMS320VC5402 is used as its core circuit, the TLC320ADSOC accomplishes the transformation of A/D and D/A of the audio signal, and TMS320VC5402 accomplishes the training and recognition of the audio signal. The DSP platform has strong processing ability, compact structure, and high efficiency. The design idea and structure form has common and practical values.
Keywords/Search Tags:Speaker recognition, Linear prediction analysis, Dynamic time wrapping, Digital Signal Processing (DSP)
PDF Full Text Request
Related items