Dark on light — the two-tone route

Everything else in this document adds light. A multiply layer takes it away (ADR-0106), and that is the whole route to a dark figure on a light ground.

system = "fragment_field"     # the LIGHT GROUND: flat, bright, full-frame
name   = "Two Tone"

[palette]
stops = [
  { at = 0.0, color = "#fff4e0" },
  { at = 1.0, color = "#ffd9a0" },
]

[params]
color_span   = "0"            # one coordinate = one flat tone across the frame
color_center = "0.2"
glow         = "1.1"

[layer]                       # the DARK FIGURE: it subtracts, it does not add
system = "swarm"
join   = "over"
blend  = "multiply"
mix    = "1"

[layer.params]
brightness = "0.15"           # <- the darkening operand. LOW is DARK.
size       = "5.0"

The rule under it is one sentence: multiply darkens by the layer’s colour, within the layer’s coverage. Both halves matter, and the second is the one that decides which scene you reach for.

brightness runs backwards here, and nothing warns

In every additive preset you have ever written, brightness up is more visible. In a multiply layer it is the darkening operand: the blend un-premultiplies the layer’s colour and multiplies the ground by it, so brightness = 0 is an opaque black mark and a high brightness clamps to white and does nothing at all. This is a parameter whose meaning is inverted by a setting in a different table, and no gate or warning can see it.

Which scene: it is a question of footprint, not of capability

layer scene	coverage	what a multiply slot gives you
fragment_field, reaction_diffusion, shape_field	every pixel (alpha = occlude, 1 by default)	the whole frame is darkened — a graded wash, a banded gradient, a figure and its surround
swarm, emitter	inside each mark only (alpha = the mark’s geometric falloff)	discrete dark marks on an untouched light ground

Both routes reach a dark figure, and the particle one reaches darker: a frozen swarm at brightness = 0 takes a light chain from display luma 174.1 to 0.9, where the field route bottoms out at 18.9. Pick by the shape you want, not by what you think can darken.

If you have read ADR-0106 or design-backlog 0069, they say a particle layer cannot darken at all. That is wrong and this table is the correction. The claim was that a particle’s alpha is its brightness; in fact swarm.rs emits vec4(color * g, g) where g is the mark’s geometric falloff and is independent of its colour, and layer_blend.rs un-premultiplies before taking the mode. core/tests/layer.rs carries the measurement.

The light ground must come from the chain, never from bg_*

The tempting version of the recipe above is a bright backdrop with a dark layer over it. It does not work, and the reason is structural rather than a tuning problem: the backdrop is not in the chain’s input (post.rs) — it is composited underneath the finished chain, so no blend mode ever sees it.

Both ways it can go:

what you set	what you get
occlude = 1 (default), any blend	the frame is byte-identical to the same preset over a black backdrop. The backdrop is not darkened, it is gone — coverage held it out
occlude = 0	the backdrop is added after the blend, so it is a floor: the same multiply layer that reaches luma 18.9 over black reaches only 171.3 over a lit sky (the sky alone reads 196.9)

So bg_bright cannot be the white paper of a two-tone graphic. Make the ground a flat, bright, full-frame scene in the chain — color_span = 0 pins a field to one tone — and spend the [layer] slot on the figure.

Two costs worth knowing before you commit to this

• It uses the preset’s only [layer]. A preset carries one layer table, so a two-tone graphic cannot also carry a second figure as counterpoint.
• “Black” is dark grey unless the marks are. The field route’s floor is luma ~18.5 with bloom and the tonemap both live; the particle route at brightness = 0 gets to 0.9. If you need a true black, that difference is the lever.

---