Brandon G. Morton

Acoustic Features for Recognizing Musical Artist Influence

Musicologists have been interested in the topic of influence between composers for years and have developed methods and heuristics for recognizing influence in classical music. While these methods work well for music where the score is the primary source of information, this type of analysis is not well suited for modern popular music where the audio recording itself is arguably the primary representation. This paper presents two audio content-based systems for influence recognition: a system using a spectral representation (Constant-q transform) and support vector machines and another system that obtains features by using a deep belief network and then logistic regression for classification. The system using the spectral representation provides a baseline for future comparisons and evidence to support the idea that influence recognition can be performed using information extracted from the audio signal. The other system attempts to improve performance by using a deep belief network to learn features useful for influence recognition by mapping data extracted from the audio signal to labeled influence data. A dataset of about 77,000 30-second audio clips, consisting of retail previews of popular music tracks was gathered for this work. These songs were chosen from expertly-labeled influence relationship information gathered by the editors of the AllMusic guide.

L.E.A.P.: Localized energy awareness program through collaborative K-University STEM projects

As the worlds energy usage continues to rise while the natural energy sources continue to dwindle, it is becoming more important that the general publics awareness of energy is increased. The National Academy of Engineering, NAE, has deemed this issue so important that five out of the fourteen Engineering Grand Challenges deal with energy. Raising awareness with children about energy consumption and generation issues will help create a sense of familiarity and potentially, a desire to follow career paths that deal with solving these future energy challenges. Introductory energy modules are being taught in K-12 classrooms in California, New Jersey, and Pennsylvania with support from the National Science Foundation, NSF, GK-12 Ph.D. Candidate Fellows of Drexel University and the Dragons Teach Program, which stems from the UTeach Program in the country. A Localized Energy Awareness Program, LEAP, has been formed through a K-University STEM collaborative project. Curriculum is being taken from the university level and broken down to each K-12 level to give insight into the topic of energy. In the new Next Generation Science Standards, NGSS, for science and engineering practices there are eight practices that should be taught at each level of K-12. Through these practices and the NAE Grand Challenges, energy awareness modules have been developed to bring into K-12 and Drexel University freshman engineering classrooms. The modules focus on the current energy usage situation, current active methods of energy harvesting including renewable and non-renewable energy sources, and modeling (drawing, building and computing depending on age group) new solutions and awareness for the energy crisis. Surveys are being used to gauge the effectiveness of our activities with the students, teachers, and participating undergraduate and graduate students. This is a work in progress and results of the surveys and work are still being processed.

Analyzing the Perceptual Salience of Audio Features for Musical Emotion Recognition

While the organization of music in terms of emotional affect is a natural process for humans, quantifying it empirically proves to be a very difficult task. Consequently, no acoustic feature or combination thereof has emerged as the optimal representation for musical emotion recognition. Due to the subjective nature of emotion, determining whether an acoustic feature domain is informative requires evaluation by human subjects. In this work, we seek to perceptually evaluate two of the most commonly used features in music information retrieval: mel-frequency cepstral coefficients and chroma. Furthermore, to identify emotion-informative feature domains, we explore which musical features are most relevant in determining emotion perceptually, and which acoustic feature domains are most variant or invariant to those changes. Finally, given our collected perceptual data, we conduct an extensive computational experiment for emotion prediction accuracy on a large number of acoustic feature domains, investigating pairwise prediction both in the context of a general corpus as well as in the context of a corpus that is constrained to contain only specific musical feature transformations.