How to achieve true 1:1, distortion-free mapping in multi-projector systems

When people say “projection mapping,” they often mean “make it fit the surface.” But in museums, architectural visualization, simulators, showrooms, and any environment with text, UI, drawings, or fine lines, “it fits” is not enough. The real requirement becomes:

1:1 pixel mapping (pixel-to-pixel): the content pixels arrive at the display canvas without being silently rescaled, re-sampled, or softened by hidden steps in the pipeline.

This article frames 1:1 mapping as a responsibility boundary problem inside complex display systems, and compares two common approaches:

• GeoBox hardware processing (FPGA-based technical layer)
• Other camera-calibration software (PC/GPU-driven mapping layer) 

Why “1:1” fails in real projects

1:1 rarely fails because someone forgot a menu option. It fails because multiple layers each do “a little scaling”:

• OS or GPU scaling overrides your “custom” output
• The media server outputs a “standard” raster that does not match the effective blended canvas
• Warping is applied on top of a mismatched raster, forcing interpolation
• Overlap (edge blending) changes the effective canvas width, so the “math” is wrong from the start

If your system is doing any of the above, your “mapping” may still look aligned, but it will not be pixel-pure.

1:1 is a Technical Layer decision

At GeoBox, we treat complex display as a layered system. 1:1 is achieved only when the responsibility boundaries are explicit.

Layer model (simple but practical)

Content and Compute Layer
Media server, PC, player, real-time engine, content generation.

Technical Layer (signal orchestration)
Resolution enforcement, pixel allocation, deterministic processing, geometry correction, blending, synchronization.

Display and Optics Layer
Projectors, lenses, surface, mounting tolerances, environmental drift.

GeoBox lives primarily in the Couche technique as an external hardware node, separating “content creation” from “display topology.”

Software based solution often extends the Compute Layer pushing geometry, blending and mapping into the PC/GPU domain.

The difference is not a feature checklist. It is where the system takes ownership of pixel accuracy.

GeoBox 1:1 (Distortion-Free) Workflow

A content-preparation view, with a clear Technical Layer responsibility boundary

GeoBox makes distortion-free “1:1” projection mapping straightforward by moving the complexity downstream into a dedicated Couche technique. The content side only needs to deliver one single canvas that matches the aspect ratio of the combined display area. Everything else (how that canvas is implemented across multiple projectors) is handled in the processing layer.

• Define one simple content contract: the combined display aspect ratio
  Start from the final target surface (the combined display area). Measure or define its aspect ratio (e.g., 5:1). This ratio becomes the only requirement for content creation.
• Create content in the same aspect ratio (resolution can vary)
  As long as your content is produced in the same aspect ratio as the combined display area, the system will not introduce geometric stretching (no horizontal/vertical distortion). The input resolution can change across different players or media sources, but the aspect ratio remains consistent.
• Enforce predictable fill behavior with GeoBox “Full Screen” + geometric alignment
  Set GeoBox to the selectable “Full Screen” mode so the content always fills the output frame in a consistent, controlled way. Then use GeoBox’s basic yet powerful geometric alignment to reshape each output image to the target boundary without changing the content’s aspect ratio.
  This is the key reason the workflow stays simple: the system scales uniformly (aspect-correct), rather than letting different layers apply non-uniform scaling that causes distortion.
• Validate “no distortion” on-site, then lock it as a repeatable profile
  Use a simple geometry test pattern (circles, squares, grids). If circles remain circles and grids remain proportionate across the full canvas, you’ve verified distortion-free behavior. Save the setup as a recallable profile so the same aspect-correct mapping can be restored instantly after maintenance, signal changes, or power cycles.

Technical honesty (one sentence that prevents confusion):
“Distortion-free 1:1” here means no aspect-ratio distortion (no stretching/squashing). Image sharpness still depends on pixel density (how many pixels you feed versus how large the wall is).