Is there a performance disparity between men and women in a patriarchal institution? We investigate this question concerning the current debate on gender stereotypes and egalitarianism. Informed by a recent study of the vocal data of the US Supreme Court justices, we apply a machine-learning algorithm to an original audio dataset of 1,400 criminal trials in Chinese courtrooms, identifying cues indicating the Standard Mandarin (\textit{Putonghua}) proficiency of judges. We found that men and women judges, despite working in a patriarchal institution, are equally competent in oral Standard Mandarin. This finding suggests that women’s position and performance in the Chinese judicial system are not necessarily determined by explicit gender bias. Our analysis sheds light on the research approach of audio data and supplements the existing scholarship on the relationship between gender and language use in politics.