WHY RADIANCE
Horticulture Lighting Simulator, Powered by Radiance
Luminous Photonics uses Radiance to calculate how light moves through modeled horticultural environments, then transforms those transport results into spatial, spectral, and plant-level insight.
PHYSICAL FOUNDATION
The simulator needs an engine that treats geometry, sources, materials, visibility, and interreflection as one transport problem. Radiance provides that physical foundation and a toolchain that can be composed around a specialized horticultural application. 12
One of the most rigorously validated lighting simulation engines. 1
Radiance has been compared with lighting calculations, scale models, and real spaces. Its documented method separates direct, specular-indirect, and diffuse-indirect contributions within a physically based, light-backwards ray-tracing system. 12
Radiance is distributed as source code under a license that permits use, redistribution, and modification. That openness makes the transport foundation inspectable and keeps the simulator from depending on a sealed lighting engine. 4
Radiance is a collection of programs designed to work in concert. Scene conversion, transport, filtering, analysis, and domain-specific interfaces can remain distinct, which is exactly the boundary the simulator needs. 3
Radiance's published validation history supports the core lighting engine. It does not by itself validate Luminous Photonics fixture hardware, plant optics, electrical input, or prototype equivalence.
TRANSPORT PROVENANCE
Radiance performs optical transport. Luminous Photonics defines the domain inputs, retains transport artifacts, and turns those artifacts into horticultural views and metrics.
Room surfaces, fixture bodies, emitting faces, materials, plant geometry when required, and receiver definitions establish the declared scene.
Radiance evaluates visibility and direct and indirect transport through the declared geometry, source, and material system.
Plant-free canopy-plane samples or plant-inclusive, banded leaf-surface samples preserve the transport result before presentation.
The application aggregates receiver data, applies declared spectral profiles and units, and prepares spatial and plant-level result structures.
Heatmaps, raw-sample scatter views, fixture overlays, uniformity summaries, and leaf-level modeled photon flux make the result inspectable.
TWO-LAYER ARCHITECTURE
The simulator separates the plant-free baseline from plant-inclusive receiver transport. This keeps canopy-plane PPFD artifacts independent of plant geometry while allowing a second layer to model transport at leaf surfaces.
LAYER 01
Plant geometry: not included
LAYER 02
Plant geometry: included
| Layer | Scene | Receiver | Transport basis | Outputs |
|---|---|---|---|---|
| Baseline PPFD and uniformity | Room and fixtures | Horizontal canopy grid | Scalar PAR PPFD | Heatmaps, metrics, fixture overlays |
| FSPM receiver transport | Room, fixtures, and plants | Leaf-surface receiver samples | Multispectral B / G / O / R / FR | Incident flux, modeled spectral absorption, reflection, transmission |
Baseline PPFD artifacts remain plant-free. The FSPM scene contains plant geometry and declared leaf optical properties.
Digitized curves from Kang and Zhen (2025) are integrated into five transport bands: B, G, O, R, and FR. Profile-weighted absorptance, transmittance, and reflectance coefficients define the Rex leaf optical profile for each band. 6
The pipeline describes Luminous Photonics' model construction. Reference 6 is the source of the digitized Rex curves, not a validation of the simulator implementation.
COMPLETE FIXTURE TRANSPORT
Module housings, heat sinks, frames, and alignment links are non-emitting geometry. Emitting faces remain separate from body geometry so source output is not double-counted and body geometry does not self-occlude an incorrectly combined emitter.
Compared systems receive the same declared anodized-aluminum material proxy. Fixture bodies participate in direct visibility and shadowing, while reflected contributions remain part of the transport calculation. Radiance supports detailed fixture geometry as part of electric-lighting models. 5
void metal fixture_body_anodized_aluminum
0
0
5 0.70 0.70 0.70 0.90 0.10SIMULATION OUTPUTS
Each view answers a different question about the same modeled system. These 30 ft × 50 ft examples are output demonstrations, not results from the separate 12 ft × 12 ft quality-setting study.
PRELIMINARY NUMERICAL SENSITIVITY STUDY
12 ft × 12 ft room, SMD mode, targeting 1000 µmol/m²/s mean PPFD
Three Radiance parameter presets were evaluated against the same application scenario. Standard is the final production preset for repeated five-band simulations.
| Preset | -ab | -ad | -as | -aa | -ar | -dj | -ds | -dt | -dc | -dr | -lr | -lw |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| StandardPRODUCTION | 3 | 512 | 128 | 0.22 | 48 | 0.35 | 0.40 | 0.08 | 0.50 | 1 | 6 | 2e-4 |
| Quality | 5 | 2048 | 512 | 0.12 | 96 | 0.65 | 0.20 | 0.03 | 0.85 | 3 | 12 | 5e-5 |
| Rigorous | 6 | 4096 | 1024 | 0.08 | 128 | 0.70 | 0.15 | 0.02 | 0.90 | 4 | 16 | 2e-5 |
| Preset | Application-derived input | Mean PPFD | CV |
|---|---|---|---|
| Standard | 4713.1 W | 1000.00 | 1.60% |
| Quality | 5228.7 W | 1000.20 | 1.47% |
| Rigorous | 5173.4 W | 1000.32 | 1.38% |
Coefficient of variation 0 to 1.70%
Application-derived input 0 to 5400 W
Spatial uniformity remained stable across the three presets, while the application-derived input required to reach the target was approximately 9–10% lower under Standard than under Quality and Rigorous.
Standard provides a practical production balance for repeated five-band simulations because the higher presets substantially increase runtime due to higher sampling and bounce settings.
Application-derived electrical input is not published as a validated simulator metric because its preset sensitivity remains unresolved and no physical lighting-system prototype is available for validation.
MODEL SCOPE
The model calculates spatial and spectral transport within the declared scene, source, material, and receiver assumptions.
Plant absorption, reflection, and transmission are modeled from the declared five-band Rex optical profile. They are transport outputs, not direct biological measurements.
The simulator makes no claim of physically validated fixture electrical performance or prototype equivalence.
COMMUNITY
This work has been invited to be presented to the Radiance community at the University of Washington.
HORTICULTURE LIGHTING SIMULATOR
The engine uses precomputed playback so you can run simulations for pre-selected configurations almost instantly. Give it a try and let us know what you think!