The paradoxical role of emotional intensity in the perception of vocal affect

Abstract

Vocalizations including laughter, cries, moans, or screams constitute a potent source of information about the affective states of others. It is typically conjectured that the higher the intensity of the expressed emotion, the better the classification of affective information. However, attempts to map the relation between affective intensity and inferred meaning are controversial. Based on a newly developed stimulus database of carefully validated non-speech expressions ranging across the entire intensity spectrum from low to peak, we show that the intuition is false. Based on three experiments (N = 90), we demonstrate that intensity in fact has a paradoxical role. Participants were asked to rate and classify the authenticity, intensity and emotion, as well as valence and arousal of the wide range of vocalizations. Listeners are clearly able to infer expressed intensity and arousal; in contrast, and surprisingly, emotion category and valence have a perceptual sweet spot: moderate and strong emotions are clearly categorized, but peak emotions are maximally ambiguous. This finding, which converges with related observations from visual experiments, raises interesting theoretical challenges for the emotion communication literature.

The variably intense vocalizations of affect and emotion (VIVAE) corpus prompts new perspective on nonspeech perception.

“…First, classification accuracy may vary as a function of emotion intensity, and higher ambiguity has previously been described for instances of both weak and extreme emotions (Atias et al, 2019; Juslin & Laukka, 2001), represented in this corpus. (Note that a detailed analysis of the effects of intensity on perceptual evaluation are addressed in a separate study, Holz et al, 2021). Second, in contrast to corpora presupposing diagnostic emotion expression (Belin et al, 2008; Cordaro et al, 2016; Hawk et al, 2009; Lima et al, 2013; Maurage et al, 2007; Sauter, Eisner, Calder, & Scott, 2010; Schröder, 2003), individual expressions were not selected to best fit a characteristic pattern.…”

The ACM Multimedia 2022 Computational Paralinguistics Challenge

“…Vocalisations -The Vocalisation Corpus VOC-C: It is provided by the MPI for Empirical Aesthetics, Frankfurt am Main, featuring vocalisations -such as laughter, cries, moans, or screams -with different affective intensities, expressing different emotional states. The data from the female speakers have been made available to the public, see [12,13]; the male speakers are so far unseen. We partition the female vocalisations into Train (6 speakers, 625 samples) and Dev(elopment) (5 speakers, 460 samples), and the male vocalisations (2 speakers, 276 samples) into Test, modelling a 6-class problem with the emotional classes achievement, anger, fear, pain, pleasure, and surprise.…”

“…In the Vocalisations Sub-Challenge, non-verbal vocal expressions from the Variably Intense Vocalizations of Affect and Emotion Corpus [12,13] are used (VOC-C) for classifying the expression of six different emotions. Such human non-verbals are still understudied but are ubiquitous in human communication [24].…”

Audio-Tokens: A toolbox for rating, sorting and comparing audio samples in the browser

“…3) by allowing both horizontal and vertical dragging operations. An example would be rating emotional content of speech or music along the dimensions valence and arousal (Holz et al, 2021;Paquette et al, 2013). Lines appear during dragging Fig.…”