T_Visionarium II
T_Visionarium II offers the means to capture and re-present televisual information so that viewers can explore and edit a multitude of narrative components within the 3-D, 360-degree AVIE projection environment.
Twenty-four hours of digital, free-to-air Australian television was captured over a period of one week, and this footage was segmented and converted into a large database containing over thirty thousand video clips. Each clip was then manually tagged with metadata to define its semantic and combinatory properties. The encoded metadata includes the gender of the actors, the dominant emotions they express, the pace of the scene and specific actions such as standing up, lying down and telephoning. The segmentation of the video data breaks down the original linear narrative into components, allowing them to become the building blocks for a new kind of interactive television.
Two hundred and fifty video clips are simultaneously displayed and distributed in 3-D around AVIE's circular screen. Using a special interface, the viewer can select, rearrange and link these video clips at will, creatively composing them into combinations based on his or her personal inclination. When the viewer selects a particular clip, it enlarges on the screen and the T_Visionarium II engine then searches the database to look for other video clips that share similar metadata. The most-similar clips (the morphological synonyms) display in proximity to the selected one, the most-dissimilar clips (the morphological antonyms) appear on the opposite side of the circular screen and a grading of similarity between the two distributes on both sides. At this point the viewer can begin to combine these clips into narrative sequences by simply dragging and dropping any of the displayed video clips into the area of the main selected clip.
When initiating a further search for similar/dissimilar clips, the viewer can also use an additional menu that fine-tunes the prerogatives of this function. In effect, it emphasizes certain metadata properties above others during the search, for example by looking mainly for occurrences of a certain colour when the viewer chooses ‘colour’ from that menu, or occurrences of a particular emotion when he or she picks ‘emotion’. This menu also allows users to request combinations of these properties. Additionally, T-Visionarium II enables the viewer to look at the complete original recording from which the segmented clip was extracted.
To use the T_Visionarium II apparatus is to experience a space within which screen imagery is dynamically reformulated and reimagined. When television is thus stripped of its mundane narrative context, its aesthetic, behavioural and media qualities become strikingly apparent. T_Visionarium II is not merely the control of a televisual database; it is a mode of inhabiting the televisual imaginary—a sphere of pure and endless mediality authored by its viewers.
Australian Research Council Project Title: Interactive Narrative as a Form of Recombinatory Search in the Cinematic Transcription of Televisual Information.
Australian Research Council Discovery Project: DP0345547