Radial Basis Function Networks for Conversion of Sound Spectra

In many high-level signal processing tasks, such as pitch shifting, voice conversion or sound synthesis, accurate spectral processing is required. Here, the use of Radial Basis Function Networks (RBFN) is proposed for modeling the relationships among sets of spectral envelopes. The identification of such conversion functions is based on a procedure which learns the shape of the conversion from few couples of original target spectra (training set). The generalization properties of RBFNs provides for interpolation with respect to the pitch range. In the construction of the training set, mel-cepstral encoding of the spectrum is used to catch the perceptually most relevant spectral changes. Moreover, singular value decomposition (SVD) is used to reduce the dimension of conversion functions. The RBFN conversion functions introduced are characterized by a perceptually-based fast training procedure, desirable interpolation properties and computational efficiency.