Speech Rate Comparison When Talking to a System and Talking to a Human: A Study from a Speech-to-Speech, Machine Translation Mediated Map Task

This study focuses on the adaptation of subjects in Humanto-Human (H2H) communication in spontaneous dialogues in two different settings. The speech rate of sixteen dialogues from the HCRC Map Task corpus have been analyzed as direct H2H communication, while fifteen dialogues from the ILMT-s2s corpus have been analyzed as a Speech-to-Speech Machine Translation (S2S-MT) mediated H2H communication comparison. The analysis shows that while the mean speech rate of the subjects in the two task oriented corpora differ, in both corpora the role of the subject causes a significant difference in the speech rate with the Information Giver using a slower speech rate than the Information Follower. Also the different settings of the dialogue recordings (with or without eye contact in the HCRC corpus and with or without live video streaming in the ILMTs2s corpus) only show a negligible difference in the speech rate. However, the gender of the subjects have provided an interesting difference with the female subjects of the ILMT-s2s corpus using a slower speech rate than the male subjects, gender does not show any difference in the HCRC corpus. This indicates that the difference is not from performing the map task, but a result of their adaptation strategy to the S2S-MT system.

本研究聚焦于在两种不同情境下自然对话中人与人（H2H）交流时主体的适应情况。对来自HCRC地图任务语料库的16段对话的语速进行了分析，将其作为直接的人与人交流；同时对来自ILMT - s2s语料库的15段对话进行了分析，将其作为语音到语音机器翻译（S2S - MT）介导的人与人交流比较。分析表明，尽管两个以任务为导向的语料库中主体的平均语速不同，但在这两个语料库中，主体的角色都会导致语速出现显著差异，信息提供者的语速比信息跟随者慢。此外，对话录制的不同情境（HCRC语料库中有无眼神交流以及ILMT - s2s语料库中有无实时视频流）在语速上仅显示出可忽略不计的差异。然而，主体的性别呈现出一种有趣的差异，ILMT - s2s语料库中的女性主体语速比男性主体慢，而在HCRC语料库中性别没有显示出任何差异。这表明这种差异并非源于执行地图任务，而是他们对S2S - MT系统的适应策略所导致的结果。