数据堂—20人中文多情感合成库

20人中文多情感合成库，由中文母语发音人录制，覆盖不同年龄段音色和性别。七种情感文本，文本均来自于小说，语料音素覆盖均衡，专业语音学家参与标注，精准匹配语音合成的研发需求。共计57.3小时

This is a 20-speaker Mandarin multi-emotion speech synthesis corpus recorded by native Mandarin speakers, covering timbres and genders across diverse age groups. It includes seven types of emotional texts, all sourced from Chinese novels, with balanced phonetic coverage. Annotated by professional phoneticians, it is precisely tailored to the R&D requirements of speech synthesis, with a total duration of 57.3 hours.

搜集汇总

背景与挑战

背景概述

该数据集是一个由20位不同年龄和性别的中文母语者录制的多情感语音合成库，包含快乐、愤怒、悲伤等7种情感类型的句子，总时长57.3小时，专为语音合成研发与测试设计。数据采用专业录音设备采集，并包含拼音和韵律标注，适用于语音合成应用场景。