light op• Data kinds: image2d → image2d
• Call: import gfx2d; gfx2d.shadow_cast_2d(occluder, x, y, steps=None, softness=0.0) (or opsgfx2d.get("shadow_cast_2d"))
Visibility of every pixel from a point light at `(x, y), as image2d`.
Ray-marches the segment from each pixel to the light, sampling `occluder`
(2-D, `[0, 1], 1 = fully blocking) at steps` points with nearest
lookup, and returns `1 - max(occluder along the ray)`. For a binary
occluder that is exactly 0 or 1 — no tolerance involved.
The sample excludes the pixel itself, so an occluder is lit on its own
light-facing surface and casts behind itself.
`softness > 0` blurs the visibility map by that sigma afterwards, which is
a penumbra-shaped lie rather than a penumbra: a real one widens with
distance from the occluder. It is offered because it looks right and named
so it cannot be mistaken for physics.
`steps` defaults to the image diagonal, which samples about once per
pixel. An occluder thinner than the step spacing can be missed; that is
the honest limit of a fixed-step march and the reason `steps` is an
argument.
• 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.
• gfx2d_scene — py -3.11 examples/gfx2d_scene.py
image2d as input)light)radial_light · light_mask · normal_map_decode · normal_map_shade
*Provenance: gfx2d.py — GFX2D 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.