This dataset was recorded in quiet office and home environments, with 345 speakers participating, including 143 males and 202 females. All speakers involved in the recording were professionally selected to ensure standardized pronunciation and clear articulation. The recorded texts cover keywords and other information.