AISHELL-2 Chinese Speech Database

The speech duration of AISHELL-2, the Chinese Mandarin speech database of Hill Shell, is 1,000 hours, of which 718 hours are from AISHELL-ASR0009-[ZH-CN] and 282 hours are from AISHELL-ASR0010-[ZH-CN]. The recorded texts involve 12 fields such as wake-up words, voice control words, smart home, unmanned driving, and industrial production. The recording process was carried out in a quiet indoor environment, using 3 different devices at the same time: high-fidelity microphone (44.1kHz, 16 bit); Android system mobile phone (16kHz, 16bit); iOS system mobile phone (16kHz, 16bit). AISHELL-2 uses voice data recorded by iOS system mobile phones. 1,991 speakers from different accent areas in China participated in the recording. After transcription and annotation by professional voice proofreaders and passing strict quality inspection, the text accuracy of this database is above 96%. (Support academic research, commercial use is prohibited without permission)