Skip to main content

Topic: Creating a mood-scanner (Read 16536 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.
Creating a mood-scanner
Reply #25
Hi everyone,

I'm well aware that this topic hasn't received new answers for several years. However, some progress has been made with regard to automatic mood detection.

Consider, for example, the following papers:

Panda, R., & Paiva, R. P. (2011). Automatic Creation of Mood Playlists in the Thayer Plane: A Methodology and a Comparative Study.
Schuller, B., Hage, C., Schuller, D., & Rigoll, G. (2010). ‘Mister D.J., Cheer Me Up!’: Musical and Textual Features for Automatic Mood Classification. Journal of New Music Research, 39(1), 13-34.
Yang, Y. H., Lin, Y. C., Cheng, H. T., & Chen, H. H. (2008). Mr. emo: Music retrieval in the emotion plane.

In most of these articles, the basic idea is to model emotional responses in two dimensions through low-level audio descriptors, e.g. MPEG7 descriptors. Though this does not capture individual differences, it seems at least to result in a good approximation of emotional responses. Of course, it won't work flawlessly for every song (there will always be outliers) - but in my opinion, providing statistically testable values of emotionality is a huge step forward.