Archive
IJMSTA - Vol. 5 - Issue 2 - July 2023
ISSN 2612-2146
Pages: 7
Music Generation by Direct Sonification and Musification Using
EEG Data
Authors: Hideaki Inoue
Categories: Journal
Abstract - This study attempts to generate music with real-time EEG data. I used the EEG data collected from people who listened to music that was previously measured and I created a music generation program using a Support Vector Machine (SVM). A headset-type device called Emotiv Insight (Emotiv, San Francisco, CA, USA) was used for the EEG measurement. For musical stimuli, I used the third movement of Beethoven's No. 6 and Barber's Adagio for Strings. Arousal-valence was adopted for the impression assessment. Based on the results, a composition algorithm was created, and real-time generation of songs was attempted. The algorithm combined two methods: "direct sonification", which directly interprets EEG data to sound, and "musification", which identifies the plus-minus rating of arousal-valence using an SVM and appoints a major or minor key accompaniment.
Keywords: Brain-computer music interface (BCMI), Electroencephalogram (EEG), Algorithm composition, Emotiv Insight
Download paper
