Presence opens the door.
Panoramic scale and spatial sound can establish place quickly. That sense of arrival is useful, but it is an introduction rather than an outcome.
Research notebook · museum science · instructional design
Beyond the dollhouse.
Spatial capture can preserve a room, reveal its geometry, and offer a new point of view. The research question is harder: when does that freedom help someone understand what the room contains?
02 studies07 viewpointsPanorama + depth + mesh
01 / The question
A room is not yet an interpretation.
A spatial model can produce an immediate sense of presence. Without evidence, story, comparison, and a reason to look closely, that first impression can quickly become the whole experience.
Working hypothesis
Three-dimensional reconstruction works best as a stage for inquiry—not as the lesson itself.
Capture→Orient→Notice→Connect→Understand
02 / Prototype lab
Compare two early spatial studies.
Both experiments place panoramic imagery onto depth or mesh information so the viewer can move beyond a fixed camera point. The rebuilt shell removes the obsolete controls while preserving every recorded viewpoint.
Connected-room study
Gallery geometry across six positions
A linked reconstruction tests movement between camera positions and a detached spatial view of several gallery rooms.
- Input
- Panoramic photographs + reconstructed geometry
- Test
- Orientation, free movement, linked viewpoints
- Constraint
- Geometry can attract attention away from the art and its interpretation
Experimental archival software may require a modern browser with WebGL. The studies are preserved as research artifacts; visual gaps in detached spatial views reveal the limits of the source geometry.
03 / What the tests revealed
The model should make the subject easier to reach.
Orientation must stay legible.
Free movement adds cognitive work. Maps, named viewpoints, guided routes, and dependable reset controls help the learner form a mental model.
Evidence needs an invitation.
Objects, labels, photographs, and architectural details require visible prompts. Hidden polygons and unexplained hotspots make important material easy to miss.
Choice needs structure.
Exploration becomes more productive when learners can pursue a question, compare sources, or follow a short interpretive path without losing the freedom to wander.
Access is part of the model.
A complete experience needs keyboard operation, readable text, motion controls, captions, transcripts, and a meaningful non-spatial path through the same content.
Evaluation comes before novelty.
The useful measure is not whether a model looks impressive. It is whether people can find, interpret, remember, and apply what the museum wants to share.
04 / An evolving toolkit
Different representations answer different questions.
Photographic presence from a fixed point
Excellent image quality, efficient streaming, and faithful visual documentation. Movement occurs between selected viewpoints.
Useful for:gallery records, remote visits, close looking, linked interpretation
Geometry reconstructed from overlapping images
A navigable surface can support measurement, detached views, and object or room-scale inspection. Reflective, transparent, and hidden surfaces remain difficult.
Useful for:objects, architecture, context, conservation reference, spatial relationships
Capture, processing, and navigation as a platform
Systems such as Matterport combine cameras or phone capture with automated alignment, spatial data, a dollhouse view, annotations, and guided tours.
Useful for:repeatable production, remote access, facilities, standard documentation workflows
Views synthesized from a learned scene
NeRF represents a scene with a neural network optimized from posed images, enabling new viewpoints without requiring a conventional textured mesh.
Useful for:view synthesis research, complex appearance, captured scenes
Real-time radiance-field rendering
Optimized 3D Gaussians provide an explicit scene representation and fast, high-quality novel-view rendering. Capture coverage and delivery size still matter.
Useful for:photorealistic objects and places, free-viewpoint presentation, spatial media research
Touch, sound, movement, and spatial cues
Haptics and other sensory channels can make relationships perceptible in new ways, but they require careful testing, accessible alternatives, and a clear learning purpose.
Useful for:tactile form, embodied procedures, access research, guided simulation
05 / Continuing questions
What should be tested next?
- When does a detached spatial view improve orientation, and when does it distract?
- How should a learner move between place, object, primary source, and explanation?
- Can fragile, remote, or inaccessible environments be explored without implying that a model replaces the original?
- Which motion, scale, contrast, caption, transcript, and keyboard controls should be standard?
- What evidence shows that a spatial experience improved recall, interpretation, curiosity, or transfer?
Acknowledgment
These independent experiments were made possible by access to museum spaces and collections. With thanks to the Museum of the Shenandoah Valley and the Smithsonian Institution. Their inclusion here does not imply sponsorship or endorsement.
Sources