Novice Users' Evaluation of Two Multi-track Music Machines for AI-Assisted Music Composition: Usability, User Experience and Acceptance
Abstract
The maturity of creative AI systems in the arts raises new questions regarding their integration into creative practices. The music field is no exception, and is seeing a rise in new creative AI tools, notably in music composition. We study how these systems and their design impact user's adoption. Specifically, we conducted a user study with 98 novice participants evaluating usability, user experience, and technology acceptance for two computer-assisted composition (CAC) systems: MMM-Cubase v2 and Calliope. Findings show both systems are easy to control and use, with Calliope being easier to use and more immersive. 76.9% (MMM-Cubase v2) and 72.9% (Calliope) of users report positive predicted future use, while the novel and efficient workflow contributes to lower barriers to music-making. Depth of control and model transparency remain outstanding issues while users highlight concerns over loss of musical diversity and skill learning.
All Study Forms : Details on the measurement instruments used for evaluation.
Generated Music Files - MMM-Cubase v2 : Music generated (audio files) by the participants using MMM-Cubase v2.
Generated Music Files - Calliope : Music generated (MIDI files) by the participants using Calliope.
Video Demo
MMM-Cubase v2 Plugin Demo
In this video, we demonstrate the generative capabilities and interaction design of the MMM-Cubase v2 plugin, a research project in partnership with Steinberg Media Technologies GmbH.
Literature Map
Poster
[Coming soon]
