Fundamental frequency modeling using wavelets for emotional voice conversion