This dataset was recorded in a quiet office environment, with 20 speakers participating, including 10 males and 10 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.