Affect Responsive Music for Conversational Meditation

Source of Publication

Proceedings of the 31st International Symposium on Electronic/Emerging Art (ISEA2026)

Conference Year

2026

Abstract

This paper presents a conversational music generation system designed to support mindfulness practice through spoken interaction. Users engage with an LLM-based guide that encourages verbal expression of sensory experience, and their utterances are used to estimate affective tendencies as a probability distribution across six emotion categories. Emphasizing recent conversational input, the system maps dominant affective states to descriptive phrases used to construct text prompts for music generation, with each audio segment shaped by the current affective estimate. This sequential structure linking speech, emotion estimation, and music generation offers an auditory environment for sustaining reflective awareness during mindfulness practice.

Document Type

Conference Proceeding

Publication Date

11-6-2026

This document is currently not available here.

Share

COinS