Auditory morphing based on an elastic perceptual distance metric in an interference-free time-frequency representation
Hideki Kawahara, Hisami Matsui · 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03). · 2003
An elastic spectral distance measure based on a F0 adaptive pitch synchronous spectral estimation and selective elimination of periodicity interferences, that was developed for a high-quality speech modification procedure STRAIGHT [1], is introduced to provide a basis for auditory morphing. The proposed measure is implemented on a low dimensional piecewise bilinear time-frequency mapping between the target and the original speech representations. A preliminary test results of morphing emotional speech samples indicated that proposed procedure provides perceptually monotonic and high-quality interpolation and extrapolation of CD quality speech samples.