Johann Sebastian Bach is considered one of the great composers of Western classical music. Now, researchers are trying to analyze the music of this German composer using information theory to find out why he is regarded as a great composer.
Suman Kulkarni and his colleagues from the University of Pennsylvania in the United States wanted to understand the relationship between the ability to remember or predict music and the structure of music. They chose Bach's works as the subject of analysis because he created a large number of works with different structures, including religious hymns and toccatas. The relevant study was published on February 2nd in Physical Review Research.
First, the researchers treated each note as a node, represented each transition between notes as an edge, and connected them to transform the works into an information network. Using this network, the researchers compared the amount of information in each composition and found that free and improvisational toccatas, which are intended to entertain and surprise, contain more information than religious hymns created for more contemplative environments such as churches.
Kulkarni and his colleagues also compared Bach's music with the listeners' perception of music using the information network. The researchers used an existing computer model based on participants' reactions to a series of images on the screen to construct it.
In the experiment, the researchers adjusted the information network based on the model to make it applicable for music analysis, where the connections between each note represent the likelihood perceived by the listeners that the two connected notes are played continuously or the level of surprise if it were to happen.
Because humans do not learn information perfectly, the information network that shows people's speculation about note variations rarely matches the information network directly based on the composition. This allowed the researchers to quantify this mismatch. The results showed a low level of mismatch, indicating that Bach's works effectively conveyed information.
However, Kulkarni hopes to fine-tune the computer model of human perception to better match the real brain scans when people listen to music.
According to Randy McIntosh from Simon Fraser University in Canada, this study has the potential to fill the knowledge gap in the field of neuroscience between complex structures like music and brain responses. However, there are still many factors that can influence a person's perception of music, such as the length of a piece of music and whether they have received musical training. He said that these factors still need to be explained clearly.
Related Paper Information:
![[Financial Morning News] Friday, January 26, 2024](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAIAAAACCAYAAABytg0kAAAAFElEQVQYV2N8+vTpfwYGBgZGGAMAUNMHXwvOkQUAAAAASUVORK5CYII=)
