Published June 2, 2026 | Version v1
Report Open

ARGIRA Experiment 8: Robustness of Visual–Acoustic Correspondence Across Roughness Thresholds

Authors/Creators

Description

This record contains the materials for ARGIRA Experiment 8, a sensitivity analysis evaluating how the choice of acoustic roughness threshold affects the observed correspondence between image chromaticity and sonified acoustic structure.

 

Using the complete ARGIRA v3 corpus (69 valid images after preprocessing), three sonification pipelines were compared:

 

• Naive deterministic mapping

• OPRS (Order-Preserved Random Spacing)

• RTR (Random Total Randomization)

 

Roughness was computed using multiple cutoff frequencies (500, 750, 1000, 1250, 1500, 1750 and 2000 Hz). Pearson and Spearman correlations between hue dispersion (hue_std) and acoustic roughness were evaluated for each threshold.

 

The principal result is that the ordering observed in previous experiments (Naive > OPRS > RTR) remains stable across a broad interval of thresholds (1000–1750 Hz), indicating that the visual–acoustic correspondence identified in ARGIRA is not an artifact of a single arbitrarily selected roughness cutoff.

 

Included files:

 

• Experiment 8 report (PDF)

• Source code used to generate the analysis

• CSV file containing threshold sensitivity results

 

License: CC BY-NC 4.0

Author: Jose Ranero Garcia

Project: ARGIRA (Artificial Reconstruction of Images through Acoustic Analysis)

---

Files

ARGIRA_Experiment8.pdf

Files (23.5 kB)

Name Size Download all
md5:ec99328b2e00d0871cd11848628de747
8.8 kB Download
md5:c9dc5b0c364f089f7ebaeb1adc1ade6d
12.9 kB Preview Download
md5:5cbcb3829739d3d1bde848bdf8818edd
1.7 kB Preview Download