MultiCo: MultiCo Multimodal Corpus

Multico - a uniqe multimodal corpus
The MultiCo Corpus consists of three Persuasion, Dialogue, and Sports subcorpora, and covers three categories of communicative events: monologues, dialogues, and multilogues. The audiovisual material comes, respectively, from the Polish TEDx and parliamentary speeches, from the recordings of task-oriented dialogues, and from a sports television broadcast. Each subcorpus offers approximately five hours of audio and video recordings along with time-aligned orthographic transcriptions and gesture annotation. The latter is done manually for the Persuasion and Sports subcorpora, and is based on motion capture technology for the Dialogue subcorpus. The dialogues are recorded with multiple camcorders and high-quality microphones. The MultiCo corpus is a freely available resource. Its potential applications range from research (e.g., studies on multimodal interaction), to design (e.g., the design of virtual agents), and education (e.g., training in interpersonal communication or rhetorics).