Emotion Perception in Adverse Listening Conditions

Thumbnail Image
Issue Date
Journal Title
Journal ISSN
Volume Title
Much (cross-linguistic) research has been done on nonverbal emotion perception (e.g., the perception of laughter and facial expressions). Another important aspect of emotion perception is verbal emotion perception. Nonverbal emotion perception has been found to contain specific universal characteristics which make it possible to recognise emotions across different cultures. Research shows that verbal emotion perception may also contain universal characteristics. If so, people should be able to determine which emotion is being expressed in an utterance regardless of whether they speak the language or not. Emotion perception is typically investigated in optimal laboratory conditions. However, in everyday listening conditions, background noise is prevalent. Moreover, in today’s multilingual society, many people regularly communicate in a language other than their mother tongue. This study therefore investigates verbal emotion perception in less than optimal conditions. Specifically, this study investigates the effect of not knowing the language and the effect of the presence of background noise on emotion perception. To investigate these questions, Dutch listeners who have no experience with Italian have participated in an emotion recognition task. In the future, Italian listeners will be tested as well. The experimental stimulus set consisted of 120 emotionally-coloured utterances selected from the Italian acted EMOVO Corpus. This selection consisted of 10 different utterances (five nonsense and five semantically neutral utterances) produced in one of five emotions, i.e., anger, sadness, joy, fear, or a neutral state, by both female and male speakers. Utterances were presented in the clean, and crucially, in two babble noise conditions at 5dB and -3dB. Each utterance-emotion combination was presented only once to each listener, and thus each utterance-emotion combination only appeared in one listening condition. Utterances, emotions, and listening conditions were randomized and counterbalanced over participants. After listening to each utterance, listeners had to determine which of the five emotions they thought the utterance conveyed. The emotion decisions were compared between the two listener groups and between the different listening conditions. In line with previous cross-linguistic emotion perception research, I found that Dutch participants were able to recognise emotions in Italian. Furthermore, a negative effect of noise on the perception of emotions was found. The perception of emotions in Italian became more difficult for Dutch listeners when listening in the -3dB SNR condition. Crucially, this only applies for anger and fear. Furthermore, effects of sentence type, speaker and trial were found for specific emotions, indicating that these factors may influence the process of emotion perception as a whole. The findings indicate that noise does influence the perception of emotions in a foreign language. However, to what extent noise influences the perception and if this really is only applicable to specific emotions should be further investigated. A follow-up study will focus on emotion perception by native Italians, and compare these results with the results of the current study.
Faculteit der Letteren