Drawing on the notion of musical intervals, recent studies have demonstrated the use of precise frequency ratios within human vocalisation. Methodologically, these studies have addressed human vocalisation at an individual level. In the present study, we asked whether patterns such as musical intervals can also be found among the voices of people engaging in a conversation as an emerging interpersonal phenomenon. Fifty-six participants were randomly paired and assigned to either a control or a low-trust condition. Frequency ratios were generated by juxtaposing nonlocal fundamental frequency (f0) productions from two people engaged in each given dyadic conversation. Differences were found among conditions, both in terms of interval distribution and precision. These results support the idea that psychological dispositions modulate the musical intervals generated between participants through mutual real-time vocal accommodation. They also underscore the inter-domain use of musical intervals.