Colour story
Where this came from.
A six-swatch paint-chip strip supplied by the founder: Coconut White, Cream White, Vanilla Beige, Milk Tea Beige, French Beige and Latte Beige, each labelled with a paint reference code (C1-VW-1 through C1-VW-6) rather than a hex.
SAMPLED. No hex is printed on the strip, only the paint codes, recorded in provenance; the six colour values are read from the pixels. First scored as all six strip colours — the best still lost 17.6%, because the strip itself has no true dark and no two swatches further than 29 points of lightness apart, reading as a tonal blender rather than pieced shapes. The founder, seeing it rendered, 2026-10-12: "coconut and latte is slightly too close to each other. it needs maybe a hint of espresso" — Cream White, the strip's own second-lightest and barely a point of lightness off Coconut White (the ground), was dropped for a genuine dark espresso brown not on the strip, then re-scored across every arrangement with the ground held fixed (still the lightest, still reads as "coconut", not an espresso-ground quilt). The espresso itself then lightened twice more (see its own provenance note) before the founder asked for one further change: "make D or E a light-purple instead of the beige" — Latte Beige dropped for a soft mauve-grey at the same lightness, the last of the six swatches from the original strip to go.
Where each colour came from
-
Fabric A
#E8E4E5sampled, C1-VW-1, Coconut White. Lightest, so it keeps the ground at 50.2% — held fixed through the espresso re-score so this still reads as the light, coconut-and-latte quilt it's named for, not a dark one. -
Fabric B
#6B5742NOT ON THE STRIP. A dark espresso brown, added 2026-10-12 in place of Cream White (the strip's own near-duplicate of the ground) to give this palette a true dark to anchor against, which the strip itself never had. Lightened twice from a first, near-black attempt (#3C2A1E, lightness 18) the founder called "too deep": once to #4E3A2C (24), called "still too dark"; checked this time how far it could go before it stopped doing its job rather than guessing again — contrast against the ground stays above 4.7 even at lightness 37, nowhere near the 1.25 floor — and landed here, lightness 34, comfortably inside that margin. Contrast numbers hold exactly at every step (lost 9.62%, soft 27.74%, unchanged since the first espresso); only the lightness range narrows as it lightens. Takes fabric B, the dark anchor, at 18.2% — "a hint", not the ground, so coconut and latte still carry the quilt. -
Fabric C
#D7CAC6sampled, C1-VW-4, Milk Tea Beige. -
Fabric D
#B9A3B5NOT ON THE STRIP. The founder, 2026-10-12: "make D or E a light-purple instead of the beige. keep it light and neutral like the others." Replaces Latte Beige (C1-VW-6, #C3A998) at the same lightness (68) rather than a lighter or darker one, so it slots in without reopening the rest of the arrangement — a soft, low-saturation mauve-grey, not a true purple, to stay "light and neutral" as asked. Re-checked: contrast actually improves slightly (soft seam share 27.74%→26.56%), so no cost either. -
Fabric E
#B79682sampled, C1-VW-5, French Beige. -
Fabric F
#DDD0CEsampled, C1-VW-3, Vanilla Beige.
What it does to the seams
Checked against all 14 fabric pairs weighted by how much seam length each actually shares in this quilt, against every arrangement of the five strip colours plus the added espresso, ground held fixed. 9.62% of the seam length falls below 1.25 contrast, and 26.56% is soft — down from 17.6%/31.4% on the all-strip version, and no longer reads as tonal. Lightness range 56, against 29 before.
- Seams that will not read
- 9.6%
- Soft seams
- 26.56%
- Lightness range
- 56
- Reads as
- Shapes
- A-F carries 8.83% of the quilt's seams at contrast 1.19 — those seams will not read.
- C-F carries 0.79% of the quilt's seams at contrast 1.06 — those seams will not read.
Contrast is measured on screen colours, not on fabric. It tells you where the piecing will disappear, not which bolt to buy.


