eduzhai > Applied Sciences > Engineering >

A New Method Towards Speech Files Local Features Investigation

  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: There are a few reasons for the recent increased interest in the study oflocal features of speech files. It is stated that many essential features ofthe speaker language used can appear in the form of the speech signal. Thetraditional instruments - short Fourier transform, wavelet transform, Hadamardtransforms, autocorrelation, and the like can detect not all particularproperties of the language. In this paper, we suggest a new approach to theexploration of such properties. The source signal is approximated by a new onethat has its values taken from a finite set. Then we construct a new sequenceof vectors of a fixed size on the base of those approximations. Examination ofthe distribution of the produced vectors provides a new method for adescription of speech files local characteristics. Finally, the developedtechnique is applied to the problem of the automatic distinguishing of twoknown languages used in speech files. For this purpose, a simple neural net isconsumed.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...