Voice regulation due to background noise variations has been studied extensively—the Lombard effect. However, most such studies do not involve naturalistic communication between participants. Moreover, studies of the effect of voice support—room reflected sound of one’s own voice—are relatively scarce. While verbal communication can involve multifactorial interaction of room acoustics, background noise, and other contextual factors, this study considers a subset of these to characterize voice regulation in two-person conversations. The experimental setup included pairs of participants (n = 20) performing a communication task to elicit naturalistic conversations, while being seated 2 m away from each other in a hemi-anechoic room. The independent variables included two levels of voice support and five levels of background noise (i.e., a 5 × 2 experimental design). Each talker heard their own and the other talker’s voice, floor and simulated room reflections of their own and the other talker’s voice (in real-time), and simulated background noise representing over open-back headphones. The participants’ voices were analyzed for various acoustic and phonetic measures along with their subjective responses of speaking and listening efforts.
Yadav et al. (Wed,) studied this question.