Interactive Melody Generation System for Enhancing the Creativity of Musicians

So Hirawata,Noriko Otani

2024-03-06

Abstract:This study proposes a system designed to enumerate the process of collaborative composition among humans, using automatic music composition technology. By integrating multiple Recurrent Neural Network (RNN) models, the system provides an experience akin to collaborating with several composers, thereby fostering diverse creativity. Through dynamic adaptation to the user's creative intentions, based on feedback, the system enhances its capability to generate melodies that align with user preferences and creative needs. The system's effectiveness was evaluated through experiments with composers of varying backgrounds, revealing its potential to facilitate musical creativity and suggesting avenues for further refinement. The study underscores the importance of interaction between the composer and AI, aiming to make music composition more accessible and personalized. This system represents a step towards integrating AI into the creative process, offering a new tool for composition support and collaborative artistic exploration.

Sound,Artificial Intelligence,Human-Computer Interaction,Audio and Speech Processing

What problem does this paper attempt to address?

This paper proposes an interactive melody generation system aimed at enhancing the creativity of music creators by simulating the collaborative composition process between humans. The core technology of the system utilizes multiple Recurrent Neural Network (RNN) models, combined with the Particle Swarm Optimization (PSO) algorithm, to dynamically adjust the parameters of these models based on user feedback, generating melodies that meet the user's creative needs.

Specifically, the system has the following features:
1. **Multi-model Collaborative Composition**: The system includes multiple RNN models, each capable of generating new melodies based on user-input melody fragments. This simulates the process of collaborating with multiple composers.
2. **Dynamic Adaptability**: Users can provide rating feedback on the generated melodies, and the system updates the model parameters using the PSO algorithm, making subsequent melodies more aligned with the user's preferences.
3. **Personalized Creation Support**: Through continuous iteration and learning of user preferences, the system gradually understands the user's creative intentions, thereby generating more personalized melodies.
4. **Interactive Interface Design**: The system provides an intuitive and easy-to-use user interface, allowing users to freely input melody fragments, rhythms, and other information, and to evaluate the generated melodies.

Through experimental evaluation, the system has shown to promote creativity among music creators from different backgrounds. Participants included professionals with various musical backgrounds, who provided positive feedback during the melody creation process using the system. The experimental results indicate that the system not only helps professional musicians explore new creative directions but also stimulates the creativity of non-professional users. Additionally, the study discusses how to further improve the system's functionality, such as adding click sounds to help users better perceive the rhythm.

In summary, by simulating the collaborative composition process between humans, this system provides a novel support tool for music creation, helping music creators gain inspiration and enhance creativity during the creation process.