Computerized encoding of audio information is highly complicated and not entirely understood. We utilized a hybridization of a machine learning algorithm and human auditory grouping and segmentation to further advance machine-based audio perception and grouping models. Human subjects listened to audio clips of musical selections and performed auditory grouping and segmentation of the clips. Data collected from the subjects’ grouping/segmentation were utilized by our machine learning algorithm to enhance the algorithm’s ability to emulate human auditory grouping and segmentation. A survey was also administered to collect information on demographics and musical experience. Overall, it was difficult to establish a direct correlation between the demographic data and the human-performed auditory grouping of the audio clips, with one exception concerning the number of groupings placed by subjects and the number of musical genres they enjoy. Suggestions for future research are discussed.

• This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator

• Wagner, Zachary

Publisher

• Worcester Polytechnic Institute

Identifier

• 65601
• E-project-042822-144737

Keyword

• Machine Learning
• Audio Grouping
• Music
• Psychology
• Computer Science
• Audio Perception

Advisor

• Doyle, James Kevin

Year

• 2022

Date created

• 2022-04-28

Resource type

• Major Qualifying Project

Major

• Psychological Science

Rights statement

• In Copyright