Font Size: a A A

Voiceprint Speaker Multidimensional Feature Extraction Platform Design

Posted on:2016-08-24Degree:MasterType:Thesis
Country:ChinaCandidate:D M ZhangFull Text:PDF
GTID:2308330470961480Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
With the development of technology and social, people would like more and more comfortable lives. As an essential feature of human civilization, voice’s role in the life is very important. At present, people’s need based on the application of the voice grow greater. Some applications based on voice appear in market, such as acoustic control switches, voice recognition security systems, autonomous voice chat system, independent voice services. The ideal level of these functions depends on the accuracy of speech recognition. Voiceprint feature reflected the characteristic of the sound signal, so the accuracy of extraction voiceprint characteristics is very important.Voiceprint identification is a hot issue in today’s research; it has great room for development. But the premise of voiceprint identification is that the speech signal is analyzed, extracted voiceprint. Only accurate analysis, feature extraction voiceprint, to provide a good foundation for the voiceprint recognition experiments.In the 21 st century, computer network technology and communication technology and other high technology continues to develop rapidly, most of the software and electronics products are becoming more and more in line with human use, the volume is more portable, more and more advanced technology-based, positive because of this trend, most of the world’s countries have a voice processing as a key research projects, the first when it is filled in speech recognition, important work is to extract features voiceprint.This design put forward multidimensional sound signal extraction. Voiceprint feature is extracted from the different directions and angles. This design sets up nine speech signal acquisition system. Voiceprint feature extraction platform based on LabVIEW is set combined with the data acquisition card PCI6221. The platform has the advantages of LabVIEW and MATLAB. Experiments show that multidimensional sound acquisition plat can provide a more comprehensive and accurate voiceprint characteristic. This design has the certain innovation and the reference value.
Keywords/Search Tags:Multi-channel, Voiceprint feature, MFCC, LabVIEW
PDF Full Text Request
Related items