Investigation on applied principles of story architecture in Asian genre films using audio analysis

Content-based video analysis which involves understanding of the semantic meanings in video has been an active research topic. In light of the success in content-based audiovisual analysis for movies and curiosity in understanding narrative structuring, this paper proposes to conduct conten...

Full description

Saved in:
Bibliographic Details
Main Author: Lee, Hui Qing.
Other Authors: Andy Khong Wai Hong
Format: Final Year Project
Language:English
Published: 2013
Subjects:
Online Access:http://hdl.handle.net/10356/54417
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Content-based video analysis which involves understanding of the semantic meanings in video has been an active research topic. In light of the success in content-based audiovisual analysis for movies and curiosity in understanding narrative structuring, this paper proposes to conduct content-based audio analysis of Asian Vengeance films to determine the validity and applicability of dominant structure principles as established in film literature. In this project, a set of computable audio features including standard time-domain features is proposed and extracted from widely accepted contemporary examples of the Asian Vengeance genre. Sequence structuring system is used to analyze the narrative events of the films. An interesting relationship between major events of the story and the objective measures is observed. Key findings include identification of dialog-driven scenes, music-driven scenes, change of narrative sequences and possible narrative events such as Rising Action and Intensified Conflict where tension build up. The change in energy levels also indicates the start or the end of important narrative events such as the Inciting Incident and the Climax. This paper elaborates on the observations made and also discusses on subjective measures that can be derived from audio features. With the promising results demonstrated, this paper also proposes further research opportunities.