Argira, Accessible Sound Museum

is a sound space. Your position within it determines what you hear. The artworks organize that space.

Accessibility · VoiceOver / TalkBack and exploration modes

Argira is accessible with VoiceOver and TalkBack (in-page support in development). The full, already accessible experience is at Perceptual Map. In the meantime, navigate with Tab · Space to play · ⏹ Stop all always available.

We can't test Argira with every screen reader. If something doesn't work as expected — a button that isn't announced, audio that doesn't respond — you can tell us in the survey. Every report helps us improve.

Argira Station  ·  Jose Ranero García
Peer-reviewed research  ·  ICAD 2026

ARGIRA

Sound Museum · Perceptual Space

A blind person can tell a Monet apart from a Kandinsky by sound alone.

16Works in this demo
45Works in full dataset

Every painting has a unique chromatic structure.
The conversion rule is fixed, but the sound texture that emerges is not designed.

The works are ordered from lowest to highest color variability.
Listen to the gradient: from silence to chaos.

Black/White · SilenceMaximum color · Complexity
Accessibility · Art · Sound

Imagine you couldn't see.
How would you listen to a painting?

Argira turns paintings into sound while respecting their real structure. A painting full of color sounds complex and high-pitched. A dark, quiet one sounds low and simple. It isn't an audio guide describing what you see — it's the structure of the painting translated into sound, with a mathematical correlation of r=0.8869.

  • 👁️ Accessible museums — The sound is generated through a fixed mathematical rule, but the acoustic complexity that emerges isn't programmed: it's a physical consequence of harmonic interaction. It isn't a symbolic mapping like "blue = high-pitched," but a structural transformation that preserves chromatic variability.
  • 🎨 Art for everyone — A child, an adult with no art background, someone who doesn't speak the language — all of them could feel the structure of a painting through their ears.
  • 🔢 Not arbitrary — The conversion rule is fixed (hue → frequency), but the sound texture that emerges isn't designed: it arises from the physics of harmonics. The same painting always produces the same acoustic profile.
  • 🌍 Extensible to other images — Validated across 45 works (peer-validated dataset, Zenodo v10). Extensible to other images, pending specific verification.
Frequently asked questions
Does this already exist?

Yes, sonification systems exist. But Argira began by applying Pearson's correlation (1895) and linear regression to measure the relationship between tone variability and sound complexity. Analyzing 45 works (peer-validated dataset, Zenodo v10), the correlation r = 0.8869 emerged from the data.

What is it actually for?

So that a blind person can tell a Monet apart from a Rothko by sound alone.

How does it work?

The sound is generated through a fixed mathematical rule, but the acoustic complexity that emerges isn't programmed: it's a physical consequence of harmonic interaction. It isn't a symbolic mapping like "blue = high-pitched," but a structural transformation that preserves chromatic variability.

All 16 works, from lowest to highest chromaticity

The correlation between color and sound complexity

Pearson r = 0.8869 · R² = 0.8417 · p < 0.001 · Argira v3.5 (ICAD 2026)

hue_std → Spectral Dimension (Ds)

Each point is a work. The X axis is tone variability (hue_std), the Y axis is sound complexity (Ds). The red line is the linear regression. Data from Argira v3.5 (N=45), the version presented at ICAD 2026 and validated on Zenodo.

16 works shown · Correlation r = 0.8869 calculated on N = 45 works (full dataset, Argira v3.5) · Zenodo v10 ↗

Slow down the sound · Perceive the texture

At normal speed the ear receives the complexity all at once.
Slowing down is like using a magnifying glass on the chromatic texture —
the correlation r ≈ 0.89 is preserved.

Matisse has the highest average saturation in the corpus; the magnifier reveals its harmonic texture.

Woman with a Hat, Henri Matisse, 1905
Work · Magnifier example
Henri Matisse
Woman with a Hat (1905)
Night mode / reduces listening fatigue
Geometric footprint of sound · Argira Inverse — Hz

Analyze your image

Upload any painting or image.
The browser calculates its metrics and generates the acoustic signature, using the same formula as the dataset.

Hearing test

Can you hear color?

Listen to each sound without seeing the images.
Which one sounds like more color, more chaos, more variety?
Choose with the buttons and the paintings will be revealed.
0/5
Correct answers in the crossmodal perception test