This dataset was recorded in a car noise environment, with 20 speakers participating, including 10 males and 10 females. All speakers involved in the recording were professionally selected to ensure standard pronunciation and clear enunciation. The recorded text covers information such as strings, personal names, and stock names.