There were 40 rating levels, including video and audio; either multimodal, unimodal, or physically measurable events, or non-physical abstract events, the latter being the result of a pragmatic interpretation (for the development of the pragmatic characteristics of Corpus cf. Németh, 2011). Each grade level was performed independently of other remarks. Each file was checked by one annotator with comments, but by another. The agreement between the annotators was guaranteed by frequent consultations and debates. The remark of physical events was naturally quite unimodal, as they came from a direct visual or audio observation or measurement, including, depending on the video, the direction of the gaze, the blinking movements, the hand and the head, the posture and, depending on the audio, the pitch movements, the changes in intensity, the silence, the overhanging language, the beginning and the end of the speech. Emotions were referred to in three ways: multimodal (observation of video and audio) and unimodal, after audio and still unimodal, by video (facial expressions). Indeed, these three different observational modalities did show differences in the magnitude and intensity of the emotions observed on the modalities of the same recording. In addition, as expected, this approach also provides the opportunity to capture the specificity of the spokesperson`s expression of emotions. Table 10. The most common match patterns, which consist of 2 events. Top Photo Group Ilustrasi diskusiContoh dialog 3 orangSintia: This film was released last week.
Have you seen it? Ranti: Yes, I did. Bella: Oh, I watched it yesterday too. Sintia: Really? I`ve seen it too. How do you feel about that? Ranti: I don`t really like it. There are plot holes that annoy me. Sintia: I agree with you. At the end of the film, I was not satisfied. Bella: Yes.
But comedy A really helps, at least I could feel his emotions. Ranti: I`m not sure. I think because actress B is the co-star, actor A had his confidence. Bella: That`s true, but his comedy has been good since the previous film. Sintia: Exactly. Last year, he even received an award. Too bad a lot of people aren`t happy with his film. Locher, M. (2004). Power and courtesy in action: disagreements in oral communication. Berlin: Gruyter Sheep. Table 11 confirms what we have seen in the simplest two-event diagrams: head-shaking is both before and after the expression of disagreements.
In addition, conthesatatives and backchannel are among the most common events involved in more complex patterns, including integration. The following video classes were included in the determination of the corresponding models: the physical descriptor classes of v_gaze, v_hand, v_head and v_posture (the prefix “v_” in all cases, which represents video observation) and the interpretive class of v_embl (emblems, including various shades of concordance). Table 1 shows the occurrence of elements that represent the emblem class: when designing the HuComTech Corpus, we wanted to identify a large number of multimodal behaviors over a given observation period. Based on the resulting database data, this paper focuses on the discovery of time models related to agreement/disagreement. It describes the methodological basis of the structure of the corpus, the analysis and interpretation of the data. Particular emphasis is placed on the Theme research tool: we describe both its theoretical bases facilitating the analysis of multimodal behavioral data and some methodological questions of its application to the HuComTech Corpus.