This paper presents a new approach for designing aconcatenative text-to-speech (TTS) system based onmulti-domain unit selection. The method achievesgood synthetic quality with reasonable ...