radiosity op• Data kinds: table × table × table → table
• Call: import fullseye as fs; fs.ledger.render_optscene_gi(scene, camera, lights, bounces=None, patch_mm=2.0, outer_mm=None, tol=1e-06, n_quad=3, supersample=1, study_bounces=8, study_tol=0.01) (to call the implementation directly, import radiosity3d; radiosity3d.render_optscene_gi(scene, camera, lights, bounces=None, patch_mm=2.0, outer_mm=None, tol=1e-06, n_quad=3, supersample=1, study_bounces=8, study_tol=0.01); from the registry, opsoptics.get("render_optscene_gi"))
Inspection render with interreflection: the photometry of render_optscene plus n bounces (or the converged sum) between Lambertian patches.
The same scene (optscene.scene_*), camera (optical_camera) and lights (illumdesign.light_source results) as optscene.render_optscene. The direct part is computed with the very same irradiance function, so bounces=0 reproduces render_optscene exactly (a test gate). On top of it the surfaces are cut into patches (radiosity_patches, at most *patch_mm* inside pockets and *outer_mm* (default *patch_mm*) elsewhere on a side), their view factors computed with occlusion (form_factor_matrix), and the light reflected between them added:
• bounces=None — converged, by progressive shooting (radiosity_progressive_3d, stop at residual energy tol);
• bounces=k — the Neumann partial sum: light that bounced at most k times between surfaces before reaching the camera (0 = direct only, 1 = one interreflection).
Each pixel's interreflected irradiance is gathered from the patch radiosities with the point-to-polygon closed form at the pixel's own surface point (final gather, one visibility ray per patch), so it is smooth inside a patch. Pixel radiance = ρ/π (E_direct + E_interreflected) in the same linear units as render_optscene. Light intensity enters linearly (the geometry is cached by a SHA-256 of the scene's shapes, camera and patch size — materials and light intensity are read on every call, so a calibration loop that varies albedo or intensity re-solves only the radiosity).
*study_bounces* K > 0 adds a bounce study: the images after 0..K bounces, the increment of each (max abs change / image peak), the observed ratio of successive increments g (the geometric analogue of an observed order of convergence), a GCI-style estimate 1.25 · Δ_k g/(1 − g) of what the remaining bounces still add, whether k is in the asymptotic range (g moved by at most 5 % at k and at k − 1; before it the estimate can be optimistic), the actual error against the converged image, and bounces_needed = the first k in the asymptotic range whose GCI estimate is below *study_tol*.
Returns a dict (table) with kind = "gi_render", image (H, W, 3), direct (H, W, 3, identical to render_optscene), difference (H, W, 3) = image − direct (the interreflection map), indirect_share (H, W), bounces, converged, n_shots, unshot_history, n_patches, patch_radiosity (N, 3), bounce_study (list of rows, or None), bounces_needed and notes.
Raises ValueError: an empty scene, or an unsupported one (see radiosity_patches:
Lambertian boxes, box-minus-box pockets, finite planes; an infinite plane only black); no lights; *bounces* not None or an integer in [0, 1000]; bad *patch_mm*, *tol*, *n_quad*, *supersample*, *study_bounces* or *study_tol*; an all-zero image (as render_optscene).
Every optics op validates its input before computing (nothing slips through silently):
• Units are baked into the argument name — _mm / _um / _deg / _mrad. Confusing mm with µm does not crash; it yields a plausible-looking wrong answer, so the name prevents it. Nothing here guesses the unit from the magnitude.
• Strings raise ValueError — float('50') succeeds, so an unparsed configuration value would slip through as a length (measured: thin_lens('50', '200') returned a plausible 66.667 mm). bool is refused too, as the implicit promotion True == 1.
• complex / masked arrays raise ValueError (real-valued slots only; silently dropping the imaginary part or peeling off the mask is refused). NaN/Inf raises ValueError on every input.
• Division by zero and its relatives are refused by name: focal length 0, radius of curvature 0, refractive index <= 0, a fully opaque aperture (all zeros, so the normalisation is 0/0), a PSF whose sum is <= 0, a Stokes vector with S0 = 0, and an object sitting at the front focal point (the image is at infinity).
• Only two ops return a non-finite value, and both state it as a contract: depth_of_field returns far_mm = inf beyond the hyperfocal distance (that is what the hyperfocal distance means), and gaussian_beam returns wavefront_radius_mm = inf at the waist (the radius of curvature of a plane wavefront). Both also return a finite companion (far_is_infinite / curvature_per_mm). Any other silent NaN/Inf is detected internally and raises ValueError — "float64 overflowed" and "the answer is infinite" are different claims, so the first is never returned wearing the face of the second.
• Size caps: generated grids are capped by optics.MAX_GRID (4096); supplied fields/PSFs/apertures by optics.MAX_FIELD_ELEMENTS (2^24); ABCD element chains by optics.MAX_SYSTEM_ELEMENTS (1024); Zernike by MAX_ZERNIKE_TERMS (512) / MAX_ZERNIKE_ORDER (40) / MAX_ZERNIKE_BASIS (2^25). This closes, fail-closed, the paths where a small argument triggers a huge internal allocation (measured: n_max=40 × 4096² needs 108 GB).
• Physically impossible states are refused too: a Stokes vector with degree of polarisation > 1, negative transmittance, negative intensity, and invalid Zernike indices such as n-|m| odd.
• Sample-data catalog (download URLs / licences) — 2-D uses skimage.data (BSD/public domain) plus synthetic images; 3-D lists download URLs for real data sources (Stanford, PDS, …).
• Operator provenance and references — the sources of the research/methods this op family came from.
• The canonical algorithm (author, year) and its uses are named in the family usage guide above.
• poc_pocket_interreflection — py -3.11 examples/poc_pocket_interreflection.py
table as input)abcd_matrix · wavefront_stats · sfr_from_edge · paraxial_trace · seidel_coefficients · spot_stats · tolerance_analysis · wavefront_from_opd
radiosity)form_factor_parallel_rects · form_factor_perpendicular_rects · form_factor_polygons · radiosity_patches · form_factor_matrix · radiosity_progressive_3d
*Provenance: radiosity3d.py — OPTICS operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.