Erik M. Schmidt: Modeling and Predicting Emotion in Music

Date: 
Wed, 03/20/2013 - 5:10pm - 7:00pm
Event Type: 
Colloquium
This work seeks to relate core concepts from psychology to that of signal processing to understand how to extract information relevant to musical emotion from an acoustic signal. The methods discussed here survey existing features using psychology studies and develop new features using basis functions learned directly from magnitude spectra with deep belief networks. Furthermore, this work presents a wide breadth of approaches in developing functional mappings between acoustic data and emotion space parameters (e.g., conditional random fields). Using these models, a framework is constructed for content-based modeling and prediction of musical emotion.

Full abstract:

The medium of music has evolved specifically for the expression of emotions, and it is natural for us to organize music in terms of its emotional associations. But while such organization is a natural process for humans, quantifying it empirically proves to be a very difficult task. Myriad features, such as harmony, timbre, interpretation, and lyrics affect emotion, and the mood of a piece may also change over its duration. Furthermore, in developing automated systems to organize music in terms of emotional content, we are faced with a problem that oftentimes lacks a well-defined answer; there may be considerable disagreement regarding the perception and interpretation of the emotions of a song or even ambiguity within the piece itself.

Automatic identification of musical mood is a topic still in its early stages, though it has received increasing attention in recent years. Such work offers potential not just to revolutionize how we buy and listen to our music, but to provide deeper insight into the understanding of human emotions in general. This work seeks to relate core concepts from psychology to that of signal processing to understand how to extract information relevant to musical emotion from an acoustic signal. The methods discussed here survey existing features using psychology studies and develop new features using basis functions learned directly from magnitude spectra with deep belief networks. Furthermore, this work presents a wide breadth of approaches in developing functional mappings between acoustic data and emotion space parameters (e.g., conditional random fields). Using these models, a framework is constructed for content-based modeling and prediction of musical emotion.
 
Bio:
Erik M. Schmidt is a Post-Doctoral Researcher in the Music and Entertainment Technology Laboratory (MET-lab) at Drexel University in Philadelphia, PA.  He received his Ph.D. from Drexel University in 2012, and also holds a Master's in Electrical Engineering from Drexel University and a Bachelor's from Temple University.  During his undergraduate career, Erik also worked for Aviom, Inc., a company in the market of digital audio networking technologies.  His current research focuses on the automatic prediction of emotional content in acoustic musical signals.  Erik has general research interests in the areas of signal processing and machine learning for machine understanding of music audio.
Open to the Public
Syndicate content