情感语音合成数据库

希尔贝壳中文情感语音合成数据库 AISHELL-TTS0013 共 27 小时，28883 条语音。录音语言，中文；录音地区，中国。以中国北方口音区域为主邀请 100 名发音人参与录制。每个录音人分别录制悲伤、愤怒、恐惧、快乐、中性共 5 种情绪。安静录音室录制，设备覆盖高保真，音频采样率为 44.1kHz & 16kHz，比特率 16bit。此数据库经过专业语音校对人员转写标注，标注包含字、拼音以及韵律，并通过严格质量检验，合格率达到 99%。

The Xier Beike Chinese Emotional Text-to-Speech Database, AISHELL-TTS0013, contains 27 hours of audio data with a total of 28,883 speech utterances. The recorded speech is in Mandarin Chinese, and the recording was conducted in China. A total of 100 speakers were recruited, primarily from northern China. Each speaker recorded utterances in five distinct emotional states: sadness, anger, fear, happiness, and neutrality. All recordings were made in a soundproof studio with high-fidelity equipment, with audio sampling rates of 44.1 kHz and 16 kHz, and a bit depth of 16 bits. This database was transcribed and annotated by professional speech proofreaders, with annotations including characters, pinyin, and prosody. It passed strict quality inspections, with a pass rate of 99%.