ARGIRA – Experiment 9: Generalization of Hue Dispersion–Roughness Relationships Across Independent Image Corpora

This repository contains data, analysis scripts, and results for Experiment 9 of the ARGIRA project. The experiment evaluates whether the relationship between chromatic dispersion (hue_std) and spectral roughness observed in the original ARGIRA v3 corpus generalizes to two independent image collections: alternative paintings and landscapes/objects from the Metropolitan Museum of Art Open Access collection. Three sonification pipelines are compared: • Naive – deterministic frequency mapping proportional to hue.• OPRS – Order-Preserved Random Spacing.• RTR – Random Total Reassignment. Results show robust generalization of the Naive pipeline across all corpora (r ≈ 0.54–0.62). RTR consistently produces the weakest correlations, while OPRS exhibits a corpus-dependent effect, remaining significantly separated from RTR in some corpora but not in others. The repository includes raw results, bootstrap confidence intervals, analysis scripts, scatter plots, and documentation necessary to reproduce the analyses.