Visualizing Music with Machine LearningPublic
Downloadable Contentopen in viewer
The problem of generating art through the use of artificial intelligence is vital to our understanding of novel generation and creativity by machines. Inspired by this problem, our goal for this project was to develop a system that generates music visualizations; these are images that are meant to evoke the overall theme and mood of a song. Through the creation of these music visualizations, we aimed to further explore the idea of creative artificial intelligence through AI generated art and how it compares to human-created art in terms of the overall process, the impact on the viewer, effectiveness in conveying mood and theme, and other qualitative metrics. We successfully developed a system to generate abstract music visualizations comparable in aesthetic to those generated by a human artist. We used these music visualizations to examine their effectiveness in conveying mood and theme. We found that AI can be used to convey the mood and theme of a song with some success through abstract images, as participants in a survey used to determine the effectiveness of the visualizations were correct about 50% of the themes using one method and 50% of the moods correct using an alternative method.
- 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.
- Last modified
- Date created
- Resource type
- Rights statement
- In Collection:
Permanent link to this page: https://digital.wpi.edu/show/vq27zr22w