Multimodal Dataset Normalization and Perceptual Validation for Music-Taste Correspondences

TL;DR

This study develops methods to normalize and validate music-flavor datasets across modalities, demonstrating that computational flavor targets align well with human perception and are transferable across datasets.

Contribution

The paper introduces a pipeline for cross-modal dataset normalization and validates computational flavor targets against human perception, enabling scalable music-flavor research.

Findings

Cross-modal structure is preserved across supervision regimes.

Computational flavor targets significantly align with human ratings.

Synthetic annotations capture sonic seasoning effects.

Abstract

Collecting large, aligned cross-modal datasets for music-flavor research is difficult because perceptual experiments are costly and small by design. We address this bottleneck through two complementary experiments. The first tests whether audio-flavor correlations, feature-importance rankings, and latent-factor structure transfer from an experimental soundtracks collection (257~tracks with human annotations) to a large FMA-derived corpus ($\sim$49,300 segments with synthetic labels). The second validates computational flavor targets -- derived from food chemistry via a reproducible pipeline -- against human perception in an online listener study (49~participants, 20~tracks). Results from both experiments converge: the quantitative transfer analysis confirms that cross-modal structure is preserved across supervision regimes, and the perceptual evaluation shows significant alignment between computational targets and listener ratings (permutation $p<0.0001$, Mantel $r=0.45$, Procrustes $m^2=0.51$). Together, these findings support the conclusion that sonic seasoning effects are present in synthetic FMA annotations. We release datasets and companion code to support reproducible cross-modal AI research.