Categorizing normal and pathological voices: automated and perceptual categorization.