MARingBA: Music-Adaptive Ringtones for Blended Audio Notification Delivery

Title of the Paper

MARingBA: Music-Adaptive Ringtones for Blended Audio Notification Delivery

Paper Information

• Field of Study: Human-Computer Interaction and Music Computing
• Keywords: audio notifications, adaptive interfaces, music computing, blended audio, notification design, beat matching, pitch matching, timbre adjustment, user experience evaluation, music information retrieval

Research Background and Problem

• Problems and Challenges: Traditional audio notifications (e.g., phone ringtones) are often designed to be highly prominent to ensure user attention. However, this can disrupt the music experience and cause discomfort, especially when the notification is not urgent.
  Existing studies have attempted to replace notifications with musical sound effects or embed notifications into pre-designed music, but these approaches often face limitations in terms of user familiarity and adaptability to diverse music genres.

• Research Motivation: As modern users frequently receive notifications while listening to music, designing an audio notification system that seamlessly integrates with background music becomes increasingly important. Such a system could reduce interruptions and enhance user experience.

• Related Work: Many studies have explored interaction design for notifications, focusing on balancing perceptibility and minimizing disruption. However, achieving dynamic adaptation of notifications to music without pre-restricting the range of background music remains an unexplored area.

Solution

• Method and Framework: The authors propose a novel method called "MARingBA," which dynamically adapts audio notifications to the background music users are currently listening to. The core technology leverages music information retrieval techniques and audio processing parameters to adjust notification characteristics, ensuring a harmonious and conflict-free experience.

• Innovations:

  • Introduced a design space for music-adaptive notifications, with parameters including beat matching, pitch matching, timbre adjustment, balance tuning, and frequency filtering.
  • Enabled notifications to dynamically adjust perceptibility based on urgency levels.
  • Developed a prototype system supporting real-time processing, allowing designers to create diverse notifications tailored to various music styles.

• Implementation Steps:

  1. Extract audio features from the background music (e.g., rhythm, pitch, balance).
  2. Adjust the audio characteristics of notifications to match predefined adaptation parameters for the background music (e.g., beat and pitch alignment, timbre or volume fade-in processing).
  3. Run the MARingBA system to generate notification sounds and evaluate the results through user tasks or preference testing.

Research Findings

• Specific Outcomes:

  1. Proposed a model for exploring the design space, including preliminary definitions of adaptive parameters.
  2. Developed the MARingBA system for audio notification design and testing.
  3. Validated the feasibility of the system through two experiments involving designers and general users.

• Advantages:

  • Compared to static audio notifications, MARingBA integrates more naturally into background music, enhancing user preference.
  • Offers multiple dynamic notification adaptation methods to music, supporting the management and adjustment of notification perceptibility and disruption levels.

• Experimental and Evaluation Results:

  • In the first study, music creators successfully designed diverse notifications for high, medium, and low urgency levels, demonstrating the system's creativity and effectiveness.
  • The second user test revealed that dynamic notification adaptation significantly reduces disruption and achieves a balance between high perceptibility and personalized preferences.

• Limitations and Future Directions:

  • Experimental conditions were limited to conventional Western music tonality and rhythm; further exploration is needed for non-mainstream music or more complex musical structures.
  • The current system requires pre-processing of music (approximately one minute), and real-time adaptation capabilities need optimization.
  • The integration of other notification modalities (e.g., visual, haptic) has not been explored.
  • Personalized and automated parameter generation requires further investigation.

Conclusion

This study demonstrates the potential of music-adaptive audio notifications by dynamically adjusting notification experiences based on musical features such as beat and pitch. Future research combining multimodal notifications and personalized optimization could further enhance user experience while reducing digital interruptions.