A comparative acoustic analysis of non-verbal vocalisations in film dubbing

Abstract

This study explores the translation of non-verbal vocalisations in dubbing and examines their role in maintaining the authenticity and emotional depth of audiovisual performance. It addresses a notable gap in translation and psycholinguistic research by focusing on vocal behaviours such as swallowing, sneezing, stuttering, wheezing, laboured breathing, and groaning, which have often been neglected within studies of paralanguage and kinesics. The corpus was compiled from four seasons of Friends (1994–2004) and their Ukrainian-dubbed versions, obtained from the Kyivstar TV platform. It included 1,050 instances of non-verbal vocalisations along with their corresponding dubbing solutions. A mixed-methods approach was adopted. Qualitative analysis focused on contextual use and acoustic characteristics, while quantitative analysis employed a corpus-based design. Acoustic parameters, including pitch, duration, stress, and amplitude, were measured using Praat, and Audacity was used to isolate background audio from dubbed tracks. The analysis also drew on sound analysis, functional equivalence, error analysis, and contrastive analysis frameworks. The findings demonstrate that vegetative vocalisations, such as coughing, sneezing, and swallowing, were frequently modified or omitted in the Ukrainian dub, resulting in the loss of emotional and comedic nuance. Aural vocalisations, including wheezing and laboured breathing, were similarly underrepresented, reducing synchronisation with visual cues and weakening emotional resonance. The study concludes that preserving the acoustic and emotional integrity of non-verbal vocalisations is essential for achieving greater authenticity and expressiveness in audiovisual translation.