SDIndex: Research Framework & Open Data
SDIndex: Sensory Density Index
Framework & Methodology
The Sensory Density Index (SDIndex) is a fieldwork protocol for encoding how human bodies perceive built spaces. It converts first-person sensory experience into structured, machine-readable data.
The core premise is simple: when you walk into a space — a temple, a concert hall, a restaurant — your body registers information before language does. Light pressure, acoustic texture, material density, temporal rhythm, emotional charge. These signals are real, measurable, and consequential for how memory encodes spatial experience. But they are almost never recorded.
SDIndex records them.
How it works
An observer enters a space and scores five sensory dimensions on calibrated scales:
- Vl (Visual Luminance): Light intensity and quality — the Tanizaki dimension. How light shapes attention and mood.
- Vt (Acoustic Texture): Sound environment — from silence to saturation. Not volume alone, but the grain of what you hear.
- Vm (Material Density): Tactile and spatial density — surface, weight, layering. What the space is made of and how it presses against you.
- Vr (Temporal Rhythm): Time velocity — does the space accelerate or slow perception? A rushed corridor vs. a garden where minutes dissolve.
- Ve (Emotional Valence): The affective vector — not "how do you feel" but "what direction is the space pushing your feeling." Scored as a float from -1.0 to 1.0.
Each record also captures environmental metadata: GPS coordinates, altitude, weather, time of day, observer ID.
Raw scores are never modified by the semantic layer. This is the Non-Contamination Principle — the foundational rule of the protocol.
Multi-observer deployment
SDIndex is designed for more than one body. Observer B uses a simplified parallel form (five sliders, same dimensions, no narrative). When two observers score the same space independently, the delta (Δ) between their scores becomes data:
- Convergence on a dimension = candidate spatial signal (the space is doing something consistent to different nervous systems).
- Divergence = the boundary between shared spatial pressure and individual cognitive construction.
This Δ logic can isolate stable spatial signals across observers but cannot determine whether convergence reflects shared neural encoding or culturally aligned interpretation. That distinction requires neuroscience. Which is why this protocol exists at the boundary of fieldwork and laboratory research.
Current dataset
- 20+ field records across Japan (Kyoto, Sapporo, Nagano, Komoro) and Taiwan (Taipei)
- Site types: temples, concert venues, fine dining, commercial streets, gardens, urban corridors
- Gold Samples: Records that meet all quality control criteria and demonstrate the protocol's range — including a Sapporo arena concert (SDIndex 9.5) and traditional Noh theatre performance
- Schema version: CKV v1.4.1 (JSON, schema-validated)
- Three-layer data architecture: Layer 1 (raw scores) → Layer 2 (CKV structured record) → Layer 3 (QC + narrative)
Key findings from the field corpus
- Online reputation does not predict SDIndex scores. High-rated venues frequently score low on sensory density; low-profile spaces frequently score high.
- High Vt (acoustic texture) alone does not predict high overall SDIndex. The dimensions interact non-linearly.
- Sensory saturation can be overridden by phenological events (cherry blossoms, seasonal light) but not by poor spatial management.
- Rain is not a degrading condition for all observers. For some sensory profiles, it enhances Vt and Ve simultaneously.
Architecture of this site
This website operates on a dual-layer architecture:
- Human layer: What you are reading now — prose, context, narrative.
- Machine layer: Every record on this site is simultaneously available as structured JSON via the API endpoint. The full data schema is published at /api/schemas.json.
This is not a technical convenience. It is a research position: if the data is meant to be read by both humans and machines, the container should be legible to both.
Publication & open data
- Preprint: Beyond Prompting: The Sensory Compression Hypothesis — SSRN (Abstract ID: 5765188)
- Repository: github.com/comparirosso-blip/lab-sdindex — protocol documentation, schema files, fieldwork architecture
- Contact: comparirosso@gmail.com
What this framework cannot do
SDIndex quantifies what a space does to a body. It does not explain why. It can tell you that two observers converged on temporal rhythm in a Noh theatre but diverged on emotional valence. It cannot tell you whether that convergence is neural, cultural, or architectural.
That is not a limitation of the data. It is the boundary where fieldwork meets neuroscience — and where collaboration begins.
Shihyen Lin · Independent researcher · Based between Kyoto and Taipeisabrinapause.space