Home / Research / Layered Reality Project
Featured Research Project
Leonardo painted 30+ invisible layers, some just 1–2 micrometers thick — thinner than microfilm. By aggregating thousands of photographs taken under varying conditions, we’re attempting to computationally separate these layers for the first time.
Could Leonardo have encoded sequential animation frames within his sfumato layers?
The optical barrier: Human depth of field spans several centimeters at typical viewing distances, meaning all layers within a painting’s 40-micrometer thickness would appear simultaneously in focus. Focus-dependent layer separation is physiologically impossible within a painting’s thickness.
This research reveals compelling alternative approaches: laser crystal technology, VR implementations, and DIY demonstrations can effectively illustrate the concept, while Leonardo’s documented optical knowledge and potential neurodivergence provide fascinating context for his innovative techniques.
Focus rack technique, layered artists, and the optical barrier
Rack focus shifts the focal plane during a continuous shot, moving attention between subjects at different depths. Cinematographers use wide apertures (f/1.4–f/2.8) and longer focal lengths. Notable examples: “The Graduate,” “Casino Royale,” “Breaking Bad.”
David Spriggs, Nobuhiro Nakanishi, Yosman Botero, and Xiaowan Xia all create stacked transparent layers — but none use focus-shifting. All existing layered art creates depth via parallax or lenticular redirection.
At 40 cm viewing distance with 0.6 diopter DOF, the sharp focus zone spans ~8–10 cm. The Mona Lisa’s total paint is 40 µm. Objects within millimeter-scale separations fall inside a single focus zone.
~270 camera obscura diagrams. First to compare eye function to camera obscura. Documented persistence of vision. Studied Aristotle, Euclid, Ptolemy, Bacon, and Alhazen. He had all the conceptual building blocks.
| Viewing Distance | Min Layer Spacing | Optimal Spacing |
|---|---|---|
| 20–30 cm (close) | 8–10 mm | 15–25 mm |
| 30–50 cm (comfortable) | 15–25 mm | 25–40 mm |
| 50–75 cm (normal) | 25–40 mm | 40–60 mm |
Leonardo’s glazes are thinner than microfilm emulsion
| Property | Microfilm | Mona Lisa Paint |
|---|---|---|
| Total thickness | 68–142 µm | 30–40 µm |
| Functional layer | 5–15 µm (emulsion) | 1–2 µm (individual glaze) |
| Number of layers | 1–2 | Up to 30 |
| Information carrier | Silver halide grains | Pigment particles |
At 1–2 micrometers, Leonardo’s individual layers represent precision “even by today’s standards” according to Philippe Walter (Louvre/CNRS, 2010). Theoretical information capacity: ~150 gigabytes equivalent across 30 addressable layers.
Every photograph captures unique data no scientific instrument has recorded
Yes! Historical photographs show the painting under different conservation states and lighting conditions no longer present.
Higher resolution is better, but even modest smartphone photos contain useful data. The aggregation of many imperfect images reveals more than a few perfect ones.
Reflections from the protective glass actually encode useful information about lighting angles. Crowd shadows too. Submit everything.
Images used exclusively for non-commercial research. You retain ownership. All contributions credited (or anonymized, your choice) in publications.
3D subsurface laser engraving as proof of concept
Nd:YAG laser focuses inside K9 optical crystal, creating microscopic fractures (“bubblegrams”) that scatter light. Millions of points form complete 3D images while the surface stays smooth. Point accuracy: <100 µm. Max depth: 6 inches.
Distinct layers at different depths are possible. Software slices 3D designs into thousands of Z-depth paths. Shallow layers first, then deeper. No documented focus-rack animation in crystal yet — this would be a first.
| Vendor | Specialty | Price Range |
|---|---|---|
| Crystal Sensations | Complex 3D, up to 6” depth | $500–$2,500+ |
| Bathsheba Grossman | Fine art & scientific visualization | Custom quote |
| ArtPix 3D | Consumer custom crystals | $49–$399+ |
Build a layered Mona Lisa you can actually focus-rack through
Fly through the painting’s layers in virtual reality
Image Texture → Principled BSDF + Transparent BSDF → Mix Shader → Material Output. Use Empty as animated focus target with DOF keyframes. F-Stop 0.5–2.8 for shallow DOF.
Surface Type: Transparent. Post-processing DOF with Gaussian (mobile) or Bokeh (quality) mode. Focus control script tracks distance to target and overrides focusDistance.
OpenXR + XR Interaction Toolkit. Android XR via Unity 6+. Traditional DOF is not recommended for VR (motion sickness) — use controller-based focus, hand gestures, or eye tracking instead.
Boolean modifier workflow in Blender: animate cutting plane through mesh. Slice it Up 2 plugin for After Effects. Runway Gen-3 for AI-enhanced slice animations.
What the Louvre has done — and what’s still needed
| Method | Year | Key Findings |
|---|---|---|
| 3D Laser Scanning (NRC Canada) | 2004 | 12 mm convex warp, craquelure network, incised preparatory drawing |
| Multispectral (Lumiere Technology) | 2004–06 | Spolvero, hidden underdrawings beneath lead white |
| X-ray Fluorescence (Philippe Walter) | 2010 | Sfumato layers 1–2 µm, total paint ~40 µm |
| Synchrotron X-ray Diffraction (ESRF) | 2023 | Experimental paint formulations including plumbonacrite |
| Technique | Resolution | Capability |
|---|---|---|
| Macro X-ray fluorescence | 100–150 µm | Elemental mapping, subsurface layers |
| Infrared reflectography | ~100 µm | Underdrawings, compositional changes |
| Optical coherence tomography | 1–10 µm | Cross-sectional imaging |
| Multispectral (Cotte method) | ~25 µm | Layer-by-layer analysis |
Status: No complete layer-by-layer separation imaging has been achieved or proposed for the Mona Lisa. The C2RMF continues incremental analysis. 500 years of aging — varnish replacements, pigment degradation, lead soap formation, and craquelure — complicate interpretation.
“The Old Guitarist”: hidden mother with child. “Woman Ironing”: complete hidden portrait. “The Crouching Woman”: hidden landscape. All revealed via X-ray analysis.
~1/3 of early paintings contain hidden images from canvas reuse. “Patch of Grass” revealed hidden peasant woman portrait.
Confirmed: spolvero underdrawing, hidden hairpin. Controversial: Cotte claims “different woman” beneath (disputed by Martin Kemp/Oxford, not endorsed by Louvre).
2005: completely different initial composition. 2019: MA-XRF revealed hidden infant Christ and winged angel. Palm print detected near Mary’s eye.
The academic case for ADHD and its implications for innovation
Marco Catani (King’s College London), Brain vol. 142, June 2019. Chronic procrastination from childhood, “none of his projects had been finished” after 20 years for the Duke. Polyphasic sleep. Left-handedness with right-hemisphere language dominance (<5% of population). Mirror writing suggestive of dyslexia.
Hyperfocus explains patience for 20–40 microscopic glaze layers. Divergent thinking produced sfumato. Mind wandering fueled cross-domain connections: water flow to hair, bird wings to arches. 16+ years refining the Mona Lisa as a paradox of incomplete projects.
Saper Vedere: Leonardo declared “knowing how to see” as his great theme. Research shows art students with dyslexia have superior mental imagery and 3D visualization. His notebooks are “richer in pictures than words” — dyslexia may have “channeled Leonardo’s focus into visual thinking.”
Key measurements and workflow reference
| Parameter | Value | Source |
|---|---|---|
| Mona Lisa total paint thickness | ~40 µm | Walter et al., 2010 |
| Individual glaze layers | 1–2 µm | Walter et al., 2010 |
| Number of layers | Up to 30 | Cotte / Lumiere Technology |
| Human depth of field | 0.5–0.9 diopters | PMC research |
| Min focus plane separation | 15–40 mm | Calculated from DOF |
| Microfilm emulsion | 5–15 µm | Library of Congress |
| Laser crystal accuracy | <100 µm | SSLE specifications |
| OCT best resolution | 1.2 µm | Cheung & Liang, 2015 |
The core hypothesis — focus-dependent animation within a painting’s thickness — faces a fundamental barrier: human depth of field spans centimeters while Leonardo’s layers occupy micrometers.
However, the investigation yields valuable alternative frameworks. Laser crystal technology creates genuine multi-layer portraits. DIY transparency demonstrations allow real focus-rack experiences. VR simulates layer-by-layer reveals. And the crowdsourcing approach represents a novel methodology for aggregating visual information about artworks at unprecedented scale.
Leonardo’s documented understanding of optics, persistence of vision, and accommodation — combined with his potential neurodivergent pattern recognition — makes him a uniquely plausible candidate for conceiving layered techniques ahead of his time. Whether or not hidden animation exists, his sfumato represents information encoding at scales that continue to challenge modern analysis.
Your vacation photo of the Mona Lisa contains data no scientific instrument has recorded. Help us build the largest crowdsourced analysis of a single artwork in history.