Abstract:

Films have attracted interests from different disciplines. One crucial question discussed by psychologists is whether a film interacts with the human emotional system. Experiments on film viewers’ self-projection, identification with film characters, or immediate neural responses have been evidence of film’s affective power. However, the study on the contents of viewers’ viewing experience is less explored due to the shortage of their immediate comments on ongoing stories.

In the past decade, the advent of bullet-screen, an interactive video comment featured by many Asian video-sharing platforms, has made it possible to explore time-synchronic video comments. With bullet-screen, which is laid directly over the screen, viewers can leave their instant comments towards a specific video element that evokes their emotions or feelings.

This talk will present an ongoing study that aims to uncover Chinese film viewers’ cognitive processes, which unfold dynamically in time, by digging into a considerable number of bullet-screens using Natural Language Processing methods. The specific phenomenon studied here via textual clues left by movie viewers is the emotional contagion between movie content and the audience. The study aims to see whether, without direct interaction, emotions can also be shared between the film character and the observer.

The researcher is curious about whether certain events, objects, or situations in films are responsible for evoking viewers’ particular emotions and the intensity of their emotional responses. By utilizing the corpus built with bullet-screens, the study will show how to interpret the changing emotional intensity and valence of viewers’ instant responses through movie presenting time with sentiment analysis approach, and what they perceive, feel, and judge about a particular concrete aspect of the film at a specific moment with Topic Modelling method.

Zoom meeting link:

https://hku.zoom.us/j/99064741928?pwd=eU9BRTRZcVo4b3poKzVnNTcvM1N1dz09

Meeting ID: 990 6474 1928

Passcode: 659637