SMYLE

SMYLE is a multimodal French-language corpus (16 hours) that includes audio-video and neuro-physiological data collected from 60 participants engaged in face-to-face storytelling (8.2 hours) and free conversation tasks (7.8 hours). The corpus integrates all modalities, precisely synchronized, offering a unique opportunity to investigate the cognitive characteristics of spontaneous conversation, including brain-level processes. The storytelling task comprises two conditions: one where the storyteller interacts with a normal listener and another where the listener is distracted.A total of 60 participants took part in the experiment (mean age = 22.77 years, SD = 3.29, range = 18–36). The sample consisted of 43 females and 17 males. Fifty-four participants were students from various levels and fields of study, five were employed, and one was unemployed. All participants were native French speakers, right-handed, and reported no neurological or language disorders. None of the participant dyads were acquainted prior to the experiment.Several levels of annotation were performed on the corpus, including both automatic and manual annotations:Automatic Annotations:Smile detection (neutral face, low-intensity smile, high-intensity smile)Momel-IntIntOpenSmileOpenFaceManual Annotations:GazeEnriched orthographic transcription in IPUs (including laughter, broken words, and elisions)Head movements (nod, shake, tilt, other)Feedback (generic and specific feedback)Feedback components (eyebrows, facial expressions, hands, head, type of lexicalization)Derived Annotations (from manual transcriptions):Aligned tokensSyllablesPhonetic transcriptionsPart-of-speech tagging with phonetic alignmentsSpeech activity (speech, laughter, silence)Please contact us for access to the corpus. You will be asked to sign a user agreement; we reserve the right to refuse/restrict access to certain types of data.