Research Achievement Report 2010

Novel Computing Project

Wednesday, February 5, 2011
Reported by Nhung Nguyen, M2, Student ID 80925784
Kiyoki Laboratory, Graduate School of Media and Governance, SFC, Keio University
Email: nhungntt@sfc.keio.ac.jp or sofpast@gmail.com

A Cross-Cultural Music Analysis System with Culture-Dependent Semantic Metadata Extraction Mechanisms

This year, I have continued to work in the topic of the creation of a cross-cultural music environment [Research Achivement Report 2009 ]. After getting succeeded in semantic metadata extraction for non-Western classical music, this year I have developed an environment from these extracted metadata which helps real users able to understand the cultural differences. The summary of my work is described as follows:

  • Improve Semantic Metadata extraction method: perform open tests to confirm the E-I matrices.
  • Data Structure: design a unified database structures for all E-I matrices.
  • Use cases: provide users two use cases for them to understand: (1) How different cultures interpret the same music and (2) How different cultures express their emotions through music

    Working with my professor, my assistant professor, we wrote a paper about this topic named “A Cross-cultural Music Museum System with Impression-based Analyzing Functions”. The paper was accepted and will be presented in IADIS International Conference, e-society conference, Avila, Spain, March 10-13, 2011.

    Research's Summary

    This research proposes a cross-cultural music analysis system (CCMA), which is a new computational environment to promote cross-cultural understanding and communication through the use of multimedia. This system provides a new platform for evoking the imagination of users regarding cultural diversity by presenting the variety of emotional responses to music that different cultures express. The system consists of music analysis, search and visualization functions, and is characterized by three main features as follows: (1) a culture-dependent semantic metadata-extraction method, which extracts musical elements (e.g., key, pitch and tempo) and generates impression metadata (e.g., sad, happy and dreamy) corresponding to properties of each musical culture, (2) a cross-cultural computing mechanism to represent the differences and similarities of various musical cultures, and (3) easy-to-use interfaces which help users join the system’s database creation process. In our implemented prototype system, traditional music is used as the common cultural determinant among different cultures. The significant contribution of our system is its multimedia database technology, which applies Kansei, a Japanese term that means psychological-feeling to compute and increase awareness of cultural differences.

    By using the system’s functions, a user can acquire insights about other cultures by comparing their own impression of a musical piece to how a different culture interprets the same musical piece. Furthermore, users can understand the differences and similarities in musical expression of a particular impression (e.g., sad, happy and dreamy) as well as in interpretation of a music piece among various musical cultures.

    To extract impression metadata from non-Western classical music, this study builds a model, called the “Musical Elements–Impressions E-I Operator” to represent the complicated relationship between musical elements and impressions from composers’ viewpoint. The significant features of our extraction method are (1) to normalize musical elements according to standard parameters of a musical-culture, and (2) to use typical music samples and create "cultural music filters" which apply culture-based weighting to musical elements before converting those elements to impressions.

    This paper also presents the experimental results and insights derived from those results, which have been observed through the operation of our music analysis prototype system. The experimental results make it possible to realize how groups of musical-elements show up as having the same impression-meaning across different cultures.

    Publication

    Nguyen, N., T., Sasaki, S., and Kiyoki, Y., 2011, “A Cross-cultural Music Museum System with Impression-based Analyzing Functions,” . Accepted and to be presented in IADIS International Conference, e-society conference, Avila, Spain, March 10-13, 2011.
    http://www.esociety-conf.org/

    Reference
    1. Research Achivement Report 2009
    2. Nguyen Thi Trang Nhung, Master's Thesis, 2011. "A cross-cultral music analysis system with semantic metadata extraction machenisims".
      Nhung Nguyen
      Kiyoki Laboratory, SFC