Abstract

Voice conversion finds applications in a wide variety of areas such as customisation of text to speech systems, voice editing and dubbing, voice restoration systems, etc. in addition to its initial applications of speaker conversion and conversion of speaking styles. The basic steps required for the implementation of voice conversion however remain the same. The alignment of features using various techniques such as dynamic time warping, time sequence matching etc., the development of the mapping function, and finally the conversion of features, are computation intensive and require the researcher to have in-depth knowledge of the various techniques used. In this work, in an effort to reduce the tasks of a researcher interested in using voice conversion for his application, a Matlab GUI-based tool has been designed, implemented and tested for carrying out spectral feature conversion for three applications, conversion of female to male speech, whispered to normal speech and synthetic to natural speech. The tool not only provides an easy to use interface by carrying out feature conversions using various mapping functions but also enables the user to view, save and compare the results graphically.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call