This paper presents the creation and perceptual validation of the EVE corpus, a resource, in English and French, for speech emotion recognition of audio and audiovisual content and for the generation of verbal and non-verbal behaviour.For each language, ten actors performed 10 linguistically and semantically neutral sentences with different emotions (fear, anger, happiness, sadness, disgust, surprise, confidence, confusion, contempt, empathy) and a neutral condition.Each was expressed at two arousal levels, with two trials per level.The emotional content of the corpus was perceptually validated by 600 participants per language.The corpus is accessible through this link: https://doi.org/

Public speaking (PS) self-assessment remains challenging due to multiple behavioral dimensions and a lack of objective evaluation methods.Virtual reality (VR) offers immersive training environments with automatic performance analysis capabilities.However, current evaluation systems use ad hoc metrics lacking transparency and reproducibility.No comprehensive set of multimodal, contextindependent behavioral cues exists for interpretable user feedback.We propose verbal and nonverbal features meeting three criteria: automatic measurement capability, context independence, and user interpretability.Using a multimodal corpus of VR presentations by 60 participants, we extracted 47 behavioral features via the Meta Quest Pro headset.Expert assessment used 7-point Likert scales.Correlation analysis and machine learning models demonstrate this feature set provides a relevant basis for automated PS assessment in VR.

Virtual Reality (VR) has demonstrated its potential as a training and education tool through its immersive capabilities.Combined with Intelligent Virtual Agents (IVAs), applications for acquiring public speaking skills have notably been developed to replicate distressing social situations, yet crucial in one's life.The training experience can be tailored to the users' needs by manipulating the IVAs' behavior.However, building challenging and immersive experiences requires in-depth knowledge of how virtual agents are perceived.In this context, the present article investigates the perception of medium-sized virtual audiences' valence and arousal in VR.A total of 35 audience conditions were considered, varying in IVAs' gender and nonverbal behavior, and were evaluated by 70 participants to assess the influence of these characteristics on user perception.Specifically, 7 attitude types were designed, assigning distinctive behaviors to each IVA to convey the desired levels of valence and arousal.Additionally, 5 variations in audience gender composition were included.Overall, this study provides valuable guidelines for designing virtual audiences by identifying validated attitudes associated with specific valence and arousal levels.Notably, 4 out of the 7 designed attitudes successfully elicited the intended perception ratings.The misperception of some attitudes reflects the complexity of designing virtual audiences using a priori characteristics drawn from the literature.Importantly, the attitude types remained valid regardless of the audience gender composition, as no significant influence of gender on perception was observed.

The multimodal behaviour of IVAs may convey different socio-affective dimensions, such as emotions, personality, or social capabilities. Several research works show that factors may impact the perception of the IVA’s behaviour. This paper proposes a systematic review, based on the PRISMA method, to investigate how the multimodal behaviour of IVAs is perceived with respect to socio-affective dimensions. To compare the results of different research works, a socio-emotional framework is proposed, considering the dimensions commonly employed in the studies. The conducted analysis of a wide array of studies ensures a comprehensive and transparent review, providing guidelines on the design of socio-affective IVAs.