Spoken interaction conveys more than words: vocal qualities such as pitch, timbre, and rhythm strongly influence how speakers are perceived in terms of warmth, competence, and trust. Although this has been shown in many laboratory studies, most have relied on short, isolated utterances, leaving open how impressions unfold in extended conversations. In addition, emotional expression has typically depended on the speaker's performance, which makes systematic control difficult.


To address these issues, we employed DAVID (Da Amazing Voice Inflection Device; Aucouturier et al., 2016), a real-time transformation system that allows parametric modulation of prosody. Seventy-eight Japanese university students engaged in scripted dialogues in which their voices, and those of their partners, were gradually shifted toward either happy or sad tones. Each dyad completed one session with convergent prosody (both happy or both sad) and another with divergent prosody (mismatched tones).


After each conversation, participants rated their partners and the interaction, while physiological data were collected, including heart rate, inter-beat interval, and electrodermal responses. Linear mixed-effects analyses showed that divergence, particularly when partners spoke with a positive tone and participants with a negative tone, led to higher trust ratings in opposite-gender pairs and was accompanied by elevated autonomic arousal. Convergence, by contrast, was associated with greater enjoyment of the dialogue but weaker physiological effects.


These findings indicate that conversational impressions depend not only on whether prosodies converge or diverge, but also on the emotional valence and relational context of the pair. Importantly, the results suggest practical implications: vocal alignment may play a role in counseling and therapeutic dialogue, support rapport-building in workplace communication, and inform the design of voice-based conferencing systems that aim to promote trust and engagement. Studying prosodic dynamics therefore contributes both to psychological theory and to applied settings where effective communication is essential.