Constants and thresholds
Every numeric constant and threshold in the tracing pipeline, in one place, with where it
comes from. This is the companion table README.md and the .html pages
refer to.
Each row gives the constant's current location as file.rs:line inside crates/, and a
one-line statement of what it controls. The basis column classifies how the value was
arrived at, using four labels:
- derived — an exact consequence of another fact (a unit conversion, a parameter count, a geometric identity). Changing the fact changes the constant.
- measured — chosen or confirmed by running the corpus and recording a number (a sweep, an A/B test, a regression that killed an earlier value).
- motivated — the existence and rough shape of the constant is explained in the source (why a floor or a cap is needed at all), but the specific number is not swept or derived.
- none — no derivation is stated anywhere in the source. This is the list of constants most likely to have been fitted to one example rather than measured across the corpus; a reader tuning the pipeline should treat these as the first candidates to re-derive.
Full context, measured numbers and quoted doc comments for every row below live in that
stage's own reference document — this table only summarizes. Paths are relative to
crates/.
01 — Intake (01-intake.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
REF_EXTENT |
128.0 px | inkvec-cli/src/lib.rs:90-99 |
intake size every pixel-denominated constant was tuned at | measured |
INTAKE_SCALE_FLOOR |
1.5 | inkvec-cli/src/lib.rs:150-151 |
below this, --intake-scale leaves the input alone |
measured |
INTAKE_SCALE_CAP |
8.0 | inkvec-cli/src/lib.rs:153-154 |
ceiling on how much --intake-scale discards |
motivated |
--max-dim default |
2048 px | inkvec-cli/src/args.rs:113-116 |
ceiling on traced (not emitted) size | motivated |
--time-budget split |
0.6 merge / 0.25 boundary-solve | inkvec-cli/src/args.rs:117-120 |
advisory wall-clock split between the two most expensive stages | none |
MAX_FACTOR (pixel_grid) |
32 | inkvec-cli/src/alpha.rs:222-228 |
largest replication factor the unblock pre-pass tries | motivated |
smallest (pixel_grid) |
64 px | inkvec-cli/src/alpha.rs:231-234 |
floor below which unblocking is not attempted | motivated |
| block-constant tolerance | 1/512 per channel | inkvec-cli/src/alpha.rs:219-223 |
how exactly a block must match to count as replication | motivated |
MARGIN (choose_matte) |
10.0 (CIEDE2000) | inkvec-cli/src/alpha.rs |
closeness to a matte candidate to count as "swallowed" | none |
SWALLOWED |
0.33 | inkvec-cli/src/alpha.rs:271-306 |
share of drawn-and-translucent mass a matte may swallow | motivated |
DRAWN |
0.5 | inkvec-cli/src/alpha.rs:297-307 |
alpha above which a pixel counts as silhouette, not glow | motivated |
DRAWN_FLOOR |
0.05 | inkvec-cli/src/alpha.rs |
alpha below which a pixel is ignored entirely | none |
SOFT_SHARE |
0.05 | inkvec-cli/src/alpha.rs |
glow share that keeps white without the candidate ladder | none |
FLAT_ALPHA |
0.02 | inkvec-cli/src/alpha.rs |
neighbour-alpha spread counted as "flat" translucency | none |
| CI-gate literal | 0.33 | inkvec-cli/src/alpha.rs:457 |
warns when white would swallow more than this, without --cutout |
duplicates SWALLOWED as a separate literal — see 01-intake.md Open questions |
DEGRADED_RESIDUAL / --sr-threshold |
0.5 | inkvec-sr/src/detect.rs:1-36 |
interior-residual threshold above which --sr auto cleans |
measured (30 icons, 5 conditions) |
--sr-scale default |
2 | inkvec-sr/src/detect.rs |
output scale of the SR pre-pass | none |
| interior-residual normalisation | sum / 9n |
inkvec-sr/src/detect.rs:97-104 |
matches the reference Python implementation | derived (deliberate match, not a bug) |
02 — Coverage (02-coverage.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
DEFAULT_SIGMA_MODEL |
0.05 px | inkvec-trace/src/coverage.rs:39,59-60 |
floor added in quadrature to noise-derived positional sigma | measured (analytic circles, tests/subpixel.rs:107-167) |
MAX_SIGMA |
4.0 px | inkvec-trace/src/coverage.rs:106 |
ceiling on positional uncertainty when the gradient vanishes | none |
| plateau fraction | 0.001 | inkvec-trace/src/coverage.rs:200-202 |
how much of the luminance extreme defines fg/bg |
motivated |
| extreme-band tolerance | 0.15 | inkvec-trace/src/coverage.rs |
which pixels near each extreme are averaged into fg/bg |
none |
MAD_TO_SIGMA |
0.6745 | inkvec-trace/src/coverage.rs:174-177 |
MAD-to-Gaussian-sigma conversion | derived (standard statistical constant) |
| Laplacian noise gain | sqrt(20), computed from LAPLACIAN_KERNEL |
inkvec-trace/src/coverage.rs:172,229 |
corrects the 4-neighbour Laplacian's noise gain | derived; a hardcoded sqrt(6) was a bug, fixed 2026-09-08 |
| noise floor | 0.5/255 | inkvec-trace/src/coverage.rs:1007 |
minimum estimate_noise can return |
derived (palette divides by it, coverage.rs:1004-1011) |
confidence_penalty floor |
saturation.max(0.05) |
inkvec-trace/src/coverage.rs |
prevents unbounded penalty near-zero saturation | none |
confidence_penalty cap |
8.0 | inkvec-trace/src/coverage.rs |
ceiling on sigma inflation from low saturation | none |
EDGE_FLOOR |
2/255 | inkvec-trace/src/coverage.rs:339 |
minimum first difference counted as a real edge | motivated |
MAX_W |
64.0 | inkvec-trace/src/coverage.rs:341-342 |
clamp on a single edge-width observation | motivated |
MIN_W |
0.25 | inkvec-trace/src/coverage.rs |
lower clamp on a single edge-width observation | none |
| minimum edge observations | 16 | inkvec-trace/src/coverage.rs:375-377 |
below this, intake_scale returns 1.0 |
none |
OVERSAMPLE_TOL |
3.0 (8-bit levels) | inkvec-trace/src/coverage.rs:436-440 |
round-trip error threshold for oversample_factor |
measured; only safe downstream of intake_scale, never as a gate alone |
min. size for oversample_factor |
16x16, sw/sh >= 8 |
inkvec-trace/src/coverage.rs |
avoids measuring on too little data | none |
03 — Palette (03-palette.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
SAME_INK_DE00 |
1.5 (CIEDE2000) | inkvec-trace/src/color.rs:285-302 |
perceptual floor: colours this close are one ink | measured (swept 1.0 -> 0.4145, 1.5 -> 0.4124 on the screen set) |
SOFT_SAME_INK_DE00 |
5.0 | inkvec-trace/src/color.rs:427-439 |
same-ink floor on soft/oversampled intake | measured (incorpo mark, 4x upscale) |
SOFT_NOISE_SIGMAS |
3.0 | inkvec-trace/src/color.rs:417-425 |
noise-merge threshold, gated on soft-intake evidence only | measured (costs 10.9% objective if run unconditionally) |
NOISE_SIGMAS |
0.0 | inkvec-trace/src/color.rs:617-634 |
clean-intake default noise-merge threshold, deliberately off | measured (costs 0.4328 -> 0.4451 on the screen set if always on) |
SOFT_INTAKE_EDGE |
1.75 px | inkvec-trace/src/color.rs:406-415 |
gates whether soft-intake handling runs at all | measured (980-raster corpus edge-width survey) |
DEFAULT_MERGE_DISTANCE |
0.035 (OKLab) | inkvec-trace/src/color.rs:100-125 |
palette merge radius | measured (swept 0.055/0.040/0.035/0.030 on the full set; 0.035 best) |
MIN_INK_WEIGHT |
0.004 | inkvec-trace/src/color.rs:127-133 |
minimum accumulated weight to count as a real ink | motivated |
BLEND_IMMUNE_WEIGHT |
0.03 | inkvec-trace/src/color.rs:135-147 |
image share above which a colour is never dismissed as anti-aliasing | motivated; not read anywhere in current extract_palette_mdl — see 03-palette.md Open questions |
BLEND_INTERIOR_FRACTION |
0.25 | inkvec-trace/src/color.rs:149-164 |
interior-fraction threshold, anti-aliasing vs. ink | measured (swept against conflicting optima on two corpora) |
BLEND_STRADDLE_FRACTION |
0.5 | inkvec-trace/src/color.rs:166-177 |
straddle-fraction threshold | motivated |
STRADDLE_STEP |
0.12 | inkvec-trace/src/color.rs:178-181 |
quantisation-noise floor for straddle detection | motivated |
JND_FLOOR |
0.012 (OKLab) | inkvec-trace/src/color.rs:611-615 |
below this, two colours are never treated as separate inks | none |
PARAMS_PER_INK |
3.0 | inkvec-trace/src/color.rs:279 |
one parameter per OKLab channel | derived |
STAT_PIXELS |
65536 | inkvec-trace/src/color.rs:692-705 |
cap on per-candidate statistical pass cost | derived (matches the 128px tuning point) |
INKVEC_BLEND_TMIN (env) |
0.04 | inkvec-trace/src/color.rs:535-538 |
interior-mixture band on the A-B colour axis | none |
04 — Regions (04-regions.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
min_region shipped default |
2 px | inkvec-cli/src/args.rs:64, inkvec-trace/src/lib.rs:210 |
smallest component kept before despeckle/carve | measured (a 9px colour-mode floor was tried and measured worse) |
SADDLE_SIGMAS |
3.0 | inkvec-trace/src/regions.rs:13 |
sigma a corner's coverage must clear 0.5 by before a saddle resolves | motivated (a standard "three sigma" bar, not swept) |
| absorption thinness gate | interior < area/5 | inkvec-trace/src/regions.rs |
candidate slivers for blend absorption | motivated |
| absorption dominant-neighbour gate | top neighbours >= 4/5 of foreign contacts | inkvec-trace/src/regions.rs |
which components are candidate slivers | motivated |
| absorption pass threshold | >= 4/5 pixels within tolerance | inkvec-trace/src/regions.rs |
whether a qualifying component is absorbed | motivated |
tol (both blend passes) |
max(3*sigma_noise, 0.025) sRGB |
inkvec-trace/src/regions.rs:278,449 |
how close a pixel must be to a mixture hypothesis | measured floor (~6.4/255), no stated reason for that specific floor |
reassign_blend_pixels::ROUNDS |
4 | inkvec-trace/src/regions.rs |
erosion-like rounds in per-pixel reassignment | none |
| reassign "beats own label" margin | residual < 0.5 * own residual | inkvec-trace/src/regions.rs |
how much better a blend hypothesis must be to move a pixel | motivated |
05 — Gradients (05-gradients.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
PARAMS_FLAT |
3.0 | inkvec-trace/src/gradient.rs |
flat-fill description length | derived |
PARAMS_LINEAR |
10.0 | inkvec-trace/src/gradient.rs:50-51 |
linear-gradient description length | derived |
PARAMS_RADIAL |
9.0 | inkvec-trace/src/gradient.rs:52-53 |
circular-radial description length | derived |
PARAMS_RADIAL_ELLIPTIC |
11.0 | inkvec-trace/src/gradient.rs:54-55 |
elliptical-radial description length | derived |
PARAMS_STOP |
4.0 | inkvec-trace/src/gradient.rs:56-57 |
cost of each interior stop | derived |
MAX_MID_STOPS |
2 | inkvec-trace/src/gradient.rs:58-59 |
most interior stops fitted | measured (corpus stop-count survey) |
MIN_GRADIENT_PIXELS |
16 | inkvec-trace/src/gradient.rs:62 |
fewest interior pixels before a gradient is attempted | motivated |
BIMODAL_MARGIN |
0.85 | inkvec-trace/src/gradient.rs:64-68 |
ramp-vs-step decision threshold; overridable via INKVEC_BIMODAL |
motivated |
MIN_VISIBLE_CONTRAST |
1.5/255 | inkvec-trace/src/gradient.rs:70-72 |
floor on visible contrast for any gradient candidate | motivated |
MIN_RAMP_SUPPORT |
0.10 | inkvec-trace/src/gradient.rs:73-79 |
least fraction of samples a gradient must visibly shade | motivated (concrete regressions, value not derived) |
QUANT_HALF_STEP |
0.5/255 | inkvec-trace/src/gradient.rs:81-82 |
residual dead zone from 8-bit quantisation | derived |
MIN_SHARED_BOUNDARY |
3 | inkvec-trace/src/gradient.rs:2054 |
shortest shared border before a union is considered | none |
MAX_FIT_SAMPLES |
4096 | inkvec-trace/src/gradient.rs:2056-2063 |
cap on samples one fill evaluation walks | measured (runtime: 5ms vs 46s on a 512px image) |
FIT_PIXELS_CAP |
65536 | inkvec-trace/src/gradient.rs:2067 |
pixels any fill fit sees, uniformly subsampled | motivated (cross-referenced to MAX_FIT_SAMPLES) |
CENTRE_SEARCH_SAMPLES |
1024 | inkvec-trace/src/gradient.rs:2080-2084 |
radial-centre hill-climb sample count | measured (same output as 4096, 4x cheaper) |
CARVE_RESIDUAL |
0.06 | inkvec-trace/src/gradient.rs:2515-2519 |
floor on the carve candidate threshold | motivated |
CARVE_MAX |
64 | inkvec-trace/src/gradient.rs:2520-2521 |
most features carved from one image | motivated |
bic_lambda(n) |
0.5 * ln(n) | inkvec-trace/src/gradient.rs:292-298 |
fill-selection lambda | derived (Bayesian information criterion) |
MAX_SLIVER_MISFIT |
4.0 | inkvec-trace/src/gradient/stops.rs:276 (doc at :13-15) |
rejects a candidate interior stop when the smaller sample-count side of its split holds under 1/10 of the fitted samples and its median residual exceeds 4x the larger side's (floored at 1/255) | motivated; reachable from both merge_gradient_bands_with_ink and fit_fill via the shared fit_pixels -> fit_samples -> fit_mid_stops path; no stated derivation for 4.0 itself |
06 — Planar map (06-planar-map.md)
build itself has no free numeric constants — the topology decision is exact integer
equality, not a threshold. The one constant this stage's tests exercise indirectly belongs
to the upstream saddle merge:
| name | value | file:line | controls | basis |
|---|---|---|---|---|
SADDLE_SIGMAS |
3.0 | inkvec-trace/src/regions.rs:13 |
see 04-regions.md; reused unmodified here | motivated |
07 — Sub-pixel (07-subpixel.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
MIN_UNMIX_CONTRAST |
0.02 | inkvec-trace/src/planar.rs:337 |
floor on unmixing contrast below which a point does not move | none |
CORNER_COS |
0.5 (60 deg) | inkvec-trace/src/planar.rs:367 |
turning angle above which the tangent window narrows to 1 point | motivated (geometric bound) |
INKVEC_SUBPX_WIN (env) |
default 1, range 1..=8 | inkvec-trace/src/planar.rs:353-362 |
width of the tangent-estimation window | measured (widening to 2 improved dE00 but cost DISTS and caused a face-order regression; left at 1) |
DEFAULT_SIGMA_MODEL |
0.05 px | inkvec-trace/src/coverage.rs:39 |
see 02-coverage.md | measured |
CONTRAST_REF |
0.25 | inkvec-trace/src/planar.rs:688 |
reference contrast for simplify_faint's inflation |
none |
MAX_INFLATION |
4.0 | inkvec-trace/src/planar.rs:689 |
cap on simplify_faint's sigma multiplier |
none |
JUNCTION_MAX_MOVE |
1.5 px | inkvec-trace/src/planar.rs:753 |
rejects an intersection solution beyond this move | measured (0.5px tried first, made results worse) |
JUNCTION_MIN_CONDITION |
0.02 | inkvec-trace/src/planar.rs:757 |
minimum eigenvalue ratio admitted for a junction intersection | derived (equivalent to rejecting crossings below ~16 degrees) |
junction_fit_points() |
6 | inkvec-trace/src/planar.rs:733 |
interior points for near-junction extrapolation | none |
junction_skip() |
1 | inkvec-trace/src/planar.rs:738 |
points nearest the junction excluded from the tangent fit | motivated |
junction_curvature_points() |
16 | inkvec-trace/src/planar.rs:742 |
points used to test for significant curvature | none |
MIN_QUADRATIC_POINTS |
5 | inkvec-trace/src/planar.rs:885 |
minimum points before a quadratic tangent model is tried | none |
CURVATURE_SIGNIFICANCE |
3.0 | inkvec-trace/src/planar.rs:886 |
sigma threshold for preferring a quadratic tangent | motivated ("3-sigma" convention) |
TAPER_DEGREES |
55.0 | inkvec-trace/src/planar.rs:1203 |
branch angle admitted for taper testing | motivated (deliberately loose; see 07-subpixel.md) |
TAPER_MAX_MOVE |
8.0 px | inkvec-trace/src/planar.rs:1208 |
largest move a taper estimate may make | motivated |
TAPER_MAX_CONSUMED |
0.35 | inkvec-trace/src/planar/junctions.rs:387 |
fraction of the shortest incident boundary a taper move may consume | none at the current declaration — an earlier revision's derivation comment was lost when this constant moved into planar/junctions.rs during the module split |
TAPER_MAX_SIGMA |
1.0 px | inkvec-trace/src/planar.rs:1224 |
largest standard error a taper estimate may carry | motivated |
TRIM_MIN_POINTS |
4 | inkvec-trace/src/planar.rs:1112 |
fewest points an edge keeps after trimming | motivated |
TRIM_LOOK |
3 | inkvec-trace/src/planar.rs:1116 |
how far ahead to look when deciding an edge's direction | motivated |
MIN_WIDTH / MAX_WIDTH |
0.05 / 6.0 px | inkvec-trace/src/planar/taper.rs:98,102 |
taper sample admission band | motivated |
MIN_SAMPLES |
4 | inkvec-trace/src/planar/taper.rs:105 |
fewest taper samples trusted | none |
MAX_DEFECT |
0.15 px | inkvec-trace/src/planar/taper.rs:115 |
largest tangent-circle residual admitted | measured (calibrated against three worked cases) |
08 — Boundary solve (08-boundary-solve.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
MAX_STEP |
0.35 px | inkvec-trace/src/boundary_opt.rs |
largest per-point displacement in one CG step | none |
MAX_TOTAL |
1.0 px | inkvec-trace/src/boundary_opt.rs:88-90 |
total leash from the point's starting position | motivated |
K_KINK |
0.05 | inkvec-trace/src/boundary_opt.rs |
kink weight, fraction of the data term's initial value | motivated (scaling rule derived, value not swept) |
K_ANCHOR |
0.10 | inkvec-trace/src/boundary_opt.rs |
anchor weight, fraction of the data term's initial value | motivated |
JUNCTION_ANCHOR |
4.0 | inkvec-trace/src/boundary_opt.rs |
multiplier on w_anchor at a junction point |
motivated |
MIN_CONTRAST |
2.0/255 | inkvec-trace/src/boundary_opt.rs |
pixel usable-evidence floor for the data term | none |
EPS (in priors) |
1e-4 | inkvec-trace/src/boundary_opt.rs |
floor inside the kink term's square root | motivated |
INKVEC_BOPT_ITERS (env) |
48 | inkvec-trace/src/boundary_opt.rs |
iteration cap | measured (24 was found unconverged) |
INKVEC_BOPT_MS (env) |
1200 ms | inkvec-trace/src/boundary_opt.rs:1002-1003 |
time budget | measured (60s budget gave the same result) |
| degenerate-box guard | (x1-x0)*(y1-y0) > 64 |
inkvec-trace/src/boundary_opt.rs:406-407 |
skips a segment whose bbox would touch too many spatial-hash cells | motivated |
| fold-guard floor | scale > 0.1 |
inkvec-trace/src/boundary_opt.rs |
how far the solve halves the accepted displacement before giving up | none |
| backtracking factor | step *= 0.4, up to 6 tries |
inkvec-trace/src/boundary_opt.rs |
line-search backoff | none |
| relative-improvement stop | rel < 1e-4 |
inkvec-trace/src/boundary_opt.rs |
early stop once a step buys almost nothing | none |
09 — Decode (09-decode.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
LEAK_GATE |
0.05 | inkvec-trace/src/decode.rs |
diagnostic threshold on leak (not gated on) |
none |
MIN_VERTS / MAX_VERTS |
3 / 16 | inkvec-trace/src/decode.rs |
vertex-count range a candidate order may propose | none |
MAX_DEV |
2.0 px | inkvec-trace/src/decode.rs |
worst-case chord deviation before a face is "something else" | none |
CURVE_BIAS_PX |
0.35 px | inkvec-trace/src/decode.rs |
mean-offset threshold in is_polygonal's curve test |
none |
MIN_GAIN |
0.5 | inkvec-trace/src/decode.rs |
residual-cut fraction of sse0 required for a decode |
measured (screen set: 10 worse/7 better at 1.0, better on every axis at 0.5) |
EVIDENCE_OVERRIDE |
0.0 (off) | inkvec-trace/src/decode.rs |
residual-ratio threshold to skip the post-solve shape recheck | measured trade-off (0.5 fixes one case, costs 0.6% objective; default declines the trade) |
CURVE_MIN_SAMPLES |
8 | inkvec-trace/src/decode.rs |
fewest ring samples before the curve test applies | none |
THIN_PX |
2.5 px | inkvec-trace/src/decode.rs |
width above which a face is not attempted | measured (conditioning cliff location) |
PARAMS_PER_RIBBON |
10.0 | inkvec-trace/src/decode.rs |
parameter charge for one pooled ribbon | derived |
SHARE_MAX_PX |
1.75 px | inkvec-trace/src/decode.rs |
widest ribbon share_widths will pool |
motivated |
MAX_RING_POINTS |
512 | inkvec-trace/src/decode.rs |
largest ring ring_of will accept |
motivated |
PIXELS_PER_UNKNOWN |
4 | inkvec-trace/src/decode.rs |
boundary-cut band pixels required per free coordinate | motivated |
MAX_BBOX_PIXELS |
20,000 | inkvec-trace/src/decode.rs |
largest face bounding box attempted | none |
GN_ITERS |
14 | inkvec-trace/src/decode.rs |
Gauss-Newton iteration cap | none |
FD_STEP |
0.01 px | inkvec-trace/src/decode.rs |
finite-difference step for the coverage derivative | none |
MAX_STEP |
0.35 px | inkvec-trace/src/decode.rs |
per-iteration vertex step clamp; shared value with boundary_opt::MAX_STEP |
motivated |
MAX_TOTAL |
1.0 px | inkvec-trace/src/decode.rs |
cumulative leash from a vertex's starting position | measured (uncapped: one emoji frame drifted to a 13-sided polygon) |
DECODED_SIGMA |
0.05 px | inkvec-trace/src/decode.rs |
sigma given to written-back samples | motivated (matches coverage::DEFAULT_SIGMA_MODEL) |
SAMPLE_PX |
1.0 px | inkvec-trace/src/decode.rs |
spacing of written-back samples along a decoded edge | motivated |
| GN damping schedule | mu0 1e-3, x4/reject, /3/accept, floor 1e-7, cap 1e9 | inkvec-trace/src/decode.rs |
Levenberg step damping | motivated (standard schedule shape) |
ridge in varpro |
1e-6 * trace / k |
inkvec-trace/src/decode.rs |
regularises a fill column with no pixel support | motivated |
| turning-corner angle floor | 0.4 rad | inkvec-trace/src/decode.rs |
minimum turning angle counted as a corner | none |
10 — Symmetry (10-symmetry.md)
No numeric thresholds. Every test is exact equality on integer pixel coordinates or exact
half-integer lattice points (inkvec-trace/src/symmetry.rs:19-21,156-163,243-245,250-252)
— a deliberate design choice, not an omission.
11 — Curve fitting (11-fitting.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
PARAMS_LINE |
2.0 | inkvec-fit/src/lib.rs:41 |
line parameter cost | derived |
PRUNE_SLACK |
4.0 | inkvec-fit/src/lib.rs:45, multimodel.rs:150 |
safety factor on scan cut-off | motivated |
CORNER_CHAMFER |
1.0 px | inkvec-fit/src/lib.rs:457 |
corner-adjustment chamfer allowance | derived (one level-set sampling step) |
CORNER_TURN_MIN |
pi/6 (30 deg) | inkvec-fit/src/lib.rs:460 |
when a vertex meeting is treated as a corner | none |
CORNER_DEGREES |
45.0 | inkvec-fit/src/lib.rs:722 |
corner-vs-smooth-join threshold | none |
| max-shift factor | 3.0x max(sigma), floor 0.25 | inkvec-fit/src/lib.rs:800 |
corner intersection displacement cap | motivated |
PARAMS_CUBIC |
6.0 | inkvec-fit/src/multimodel.rs:102 |
cubic parameter cost | derived |
G1_BREAK_DEGREES |
10.0 | inkvec-fit/src/multimodel.rs:137 |
tangent break below which a join is nearly free | none; swept empirically |
MAX_ARM |
1.0 | inkvec-fit/src/multimodel.rs:142, merge.rs:117 |
largest admissible control arm, as a fraction of chord | derived |
MAX_RESIDUAL_SAMPLES |
32 | inkvec-fit/src/multimodel.rs:147 |
cubic residual evaluation points (O(1) cap) | none |
PRUNE_PATIENCE |
8 | inkvec-fit/src/multimodel.rs:158 |
over-budget candidates before scan stops | measured, value not stated |
DP_MAX_POINTS |
768 | inkvec-fit/src/multimodel.rs:163 |
decimation threshold | none |
TANGENT_WINDOW_MAX |
16 | inkvec-fit/src/multimodel.rs:166 |
widest one-sided tangent window | none |
NEWTON_STEPS |
3 | inkvec-fit/src/multimodel.rs:170 |
Newton steps for arc-length projection | none |
FREE_MAX_SWING |
75.0 deg | inkvec-fit/src/multimodel.rs:1103 |
how far a free cubic's tangent may depart from the estimate | none |
DIRECTION_SAMPLES |
8 | inkvec-fit/src/multimodel.rs:1491 |
monotone-sweep samples for arc validity | none (asserted) |
MAX_ASPECT |
12.0 | inkvec-fit/src/multimodel.rs:1666 |
most elongated ellipse worth fitting | none |
MIN_POINTS (ellipse) |
24 | inkvec-fit/src/multimodel.rs:1669 |
minimum points to try an ellipse | none |
LENGTH_STRIDE |
16 | inkvec-fit/src/multimodel.rs:1676 |
ellipse candidate-length sampling | none; doc comment ("every fourth length") does not match the value |
| bow-penalty factor | 4.0 | inkvec-fit/src/multimodel.rs:1301 |
arc-vs-line residual ratio that triggers the bow penalty | none |
PARAMS_ARC |
5.0 | inkvec-fit/src/curves.rs:158 |
circular arc cost | derived |
PARAMS_ELLIPTICAL_ARC |
7.0 | inkvec-fit/src/curves.rs:165 |
elliptical arc cost | derived |
MAX_ARC_DEGREES |
120.0 | inkvec-fit/src/primitives.rs:50 |
longest single-arc sweep | derived (conditioning argument) |
MAX_REDUCED_CHI2 |
4.0 | inkvec-fit/src/primitives.rs:1081 |
primitive acceptance gate | derived (tau^2 at default tau=2) |
PARAMS_CIRCLE / PARAMS_ELLIPSE / PARAMS_ROUND_RECT / PARAMS_RECT |
3.0 / 5.0 / 6.0 / 4.0 | inkvec-fit/src/primitives.rs:32-39 |
primitive parameter costs | derived |
BREAK_PARAMS |
2.0 | inkvec-fit/src/merge.rs:53 |
join a free cubic no longer meets smoothly; overridable via INKVEC_MERGE_BREAK |
derived |
MAX_SPAN |
96 | inkvec-fit/src/merge.rs:57 |
longest merge run attempted | none |
MAX_RUN |
4 | inkvec-fit/src/merge.rs:65 |
segments a merge run may absorb | none (was overridable, no longer swept) |
MAX_ROUNDS |
6 | inkvec-fit/src/merge.rs:75 |
merge sweep passes | motivated |
SEARCH_DEGREES |
100.0 | inkvec-fit/src/merge.rs:111 |
free-cubic angle search width | motivated (a 60-degree clamp put the optimum outside the search) |
SHARPEN_MAX_CHORD |
2.5 | inkvec-fit/src/merge.rs:472 |
chamfer-cubic chord ceiling | motivated (chamfer ~1px/side) |
SHARPEN_MAX_EDGE |
8.0 | inkvec-fit/src/merge.rs:477 |
short-edge-with-chamfers ceiling | none |
SHARPEN_MIN_TURN |
pi/6 (30 deg) | inkvec-fit/src/merge.rs:479 |
corner-vs-smooth threshold | none; duplicates CORNER_TURN_MIN |
PARAMS_AXIS_LINE |
1.0 | inkvec-fit/src/merge.rs:646 |
axis-snapped line cost | derived |
MAX_AXIS_DEV_SIGMA |
3.0 | inkvec-fit/src/merge.rs:651 |
per-sample axis-snap deviation cap | none |
PARAMS_SMOOTH_CUBIC |
4.0 | inkvec-fit/src/merge.rs:831 |
S-shorthand cost |
derived |
| G1 pre-filter angle | 20 deg | inkvec-fit/src/merge.rs:902 |
when a smooth-join candidate is worth the exact refit | none |
FLATTEN |
16 | inkvec-fit/src/simple.rs:51 |
flattening resolution for the self-crossing test | motivated (below render-visibility floor) |
MAX_REPAIRS |
8 | inkvec-fit/src/simple.rs:55 |
span-cap halvings before falling back | derived (2^8 covers any contour produced) |
EPS (endpoint coincidence) |
1e-6 px | inkvec-fit/src/simple.rs:59 |
adjacency exemption tolerance | derived |
12 — Repair (12-repair.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
ROUNDS |
10 | inkvec-cli/src/rings.rs:64 |
max halving rounds of the outer repair loop | none; one comment refers to "15 rounds" (rings.rs:222), inconsistent with this value |
MERGE_BUDGET |
96 segments | inkvec-cli/src/rings.rs:164 |
per-boundary merge affordability (4x this is the global ceiling) | measured (~8ms/segment) |
RING_SAMPLES |
4 | inkvec-cli/src/rings.rs:456 |
interior samples per curved segment for area/containment | none |
| crossing-pair limit | 32 | inkvec-cli/src/rings.rs:96,260 |
max crossing pairs reported per ring per call | none |
| explosion thresholds | c > 32 && c > 4*f |
inkvec-cli/src/rings.rs:230 |
when a capped refit is discarded for the unconstrained fit | measured (bulma case); the pair (32, 4) not separately justified |
| cap initial value | polys[k].len().max(2) |
inkvec-cli/src/rings.rs:69 |
starting span cap per edge | derived |
| halving rule | (cap[k]/2).max(1) |
inkvec-cli/src/rings.rs:126,131 |
tightening schedule, keeps repair logarithmic | derived |
BLK |
16 | inkvec-cli/src/rings.rs:227 |
bounding-box block size for the all-pairs prune | motivated (speed only) |
FLATTEN, MAX_REPAIRS and EPS are shared with inkvec-fit/src/simple.rs — see 11-fitting.md.
13 — Emit (13-emit.md)
| name | value | file:line | controls | basis |
|---|---|---|---|---|
EMIT_DECIMALS |
2 | inkvec-cli/src/emit.rs:54 |
coordinate decimal places | derived — see 13-emit.md, "Coordinate precision" |
MIN_RING_AREA |
0.25 px² | inkvec-cli/src/emit.rs:57 |
smallest ring area worth emitting | motivated |
S-shorthand tolerance (fmt_ring) |
10^-decimals (0.01 px default) | inkvec-cli/src/emit.rs:193 |
rounded-space reflection test | derived |
S-shorthand tolerance (fmt_fitted) |
5e-4, raw-space | inkvec-cli/src/emit.rs:77 |
stroke-path reflection test | none |
| ring segment minimum | 2 | inkvec-cli/src/emit.rs:252 |
a ring is discarded below this | none |
| path point minimum | 3 | inkvec-cli/src/emit.rs:133 |
fmt_path early return |
derived (degenerate-polygon guard) |
| arc rotation precision | 3 decimals | inkvec-cli/src/emit.rs:105,235,291 |
arc phi formatting |
none |
| rounded-rect radius floor | 1e-4 | inkvec-cli/src/emit.rs:305,367 |
below this, written as a plain rect | none |
| ellipse rotation floor | 1e-3 | inkvec-cli/src/emit.rs:344 |
below this, no transform written |
none |
| alpha-ramp endpoint precision | 2 decimals | inkvec-cli/src/emit.rs:625 |
independent of EMIT_DECIMALS |
none |
| opacity precision | 3 decimals | inkvec-cli/src/emit.rs:626,660,905 |
fill-opacity/stop-opacity |
none |
INKVEC_EMIT_DECIMALS (env) |
overrides EMIT_DECIMALS |
inkvec-cli/src/emit.rs:126 |
the mechanism used to isolate rounding from segment price in the 7.2% measurement | measured |
| background-match precision | 2 decimals (hard-coded) | inkvec-cli/src/post.rs:42-52 |
--no-background element matching |
none; coupled to EMIT_DECIMALS but not derived from it — breaks silently if the two diverge |
| margin viewBox precision | 2 decimals | inkvec-cli/src/post.rs:147 |
--margin growth |
none |
ci_gate.py ratio limit |
5% relative | bench/ci_gate.py:33 |
regression gate on parameter count vs. artist | measured (project's compactness regression budget) |