This dataset was recorded in both quiet and noisy environments, with the participation of 34 speakers, including 19 males and 15 females. All speakers involved in the recording were professionally selected to ensure standardized pronunciation and clear enunciation. The recorded texts cover wake-up words and interrogative sentences.