This dataset was recorded in quiet office/home environments, with the participation of 150 speakers, including 76 males and 74 females. All speakers involved in the recording were professionally selected to ensure standardized pronunciation and clear enunciation. The recorded texts cover commonly used keywords in voice applications.