Naturalness and intelligibility of rule-synthesized speech, supplied with specific spectro-temporal features derived from natural continuous speech.