This dataset was recorded by 12 speakers with authentic pronunciation and diverse vocal qualities (6 males and 6 females) in a professional recording studio. The recorded texts span the full range of phonemes, and the annotators possess a professional linguistic background, ensuring the data meets the research and development needs for voice synthesis.