| 1 | # Palette Comparison Set |
| 2 | |
| 3 | ## Controlled variables |
| 4 | |
| 5 | | Dimension | Value | Status | |
| 6 | |---|---|---| |
| 7 | | Palette | **varies (14 palettes)** | under comparison | |
| 8 | | Rendering | `vector-illustration` | fixed baseline | |
| 9 | | Composition | single-subject hero — see [`image-generator.md`](../../image-generator.md) §4.1 Primitive A | fixed baseline | |
| 10 | |
| 11 | ## Subject |
| 12 | |
| 13 | > A confident professional standing beside an upward growth chart, looking forward — single dominant subject occupying 60-70% of the canvas. This is the "single-subject hero" composition described as [`image-generator.md`](../../image-generator.md) §4.1 Primitive A, used here as a stable baseline because a single dominant subject makes palette color behavior read most clearly. |
| 14 | |
| 15 | ## On HEX values |
| 16 | |
| 17 | **Important**: this set does **not** fix HEX values across all 14 palettes. Each palette uses its own typical HEX triplet. |
| 18 | |
| 19 | Reason: a palette is a *color behavior* (saturation, contrast, temperament, proportion), and the HEX it expects is part of that behavior. Forcing cool-corporate's deep navy through a vivid-launch palette would produce a strange "restrained navy mimicking promotional energy" image — losing the semantic representativeness of the palette. |
| 20 | |
| 21 | This also matches real production: a user who picks vivid-launch will have vivid HEX values in their `design_spec`, not navy. This set simulates each palette's typical production scenario. |
| 22 | |
| 23 | | Palette | Primary | Secondary | Accent | |
| 24 | |---|---|---|---| |
| 25 | | cool-corporate | `#1E3A5F` deep navy | `#F8F9FA` off-white | `#D4AF37` gold | |
| 26 | | warm-earth | `#9A3412` terracotta | `#FEF3C7` cream | `#D4AF37` gold | |
| 27 | | tech-neon | `#0EA5E9` electric blue | `#0A0E27` deep navy | `#06B6D4` cyan | |
| 28 | | editorial-classic | `#0F2C4C` deep navy | `#FAF7F2` warm cream | `#C2410C` burnt orange | |
| 29 | | macaron | (pastel tints) | `#F5F0E8` cream | `#F97316` coral | |
| 30 | | mono-ink | `#1A1A1A` near-black | `#FFFFFF` white | `#E8655A` coral red | |
| 31 | | vivid-launch | `#E11D48` magenta | `#FAFAFA` near-white | `#F97316` orange | |
| 32 | | dark-cinematic | `#14B8A6` teal | `#0A0E27` deep dark | `#D4AF37` amber | |
| 33 | | duotone | `#0F766E` deep teal | `#D97706` amber | (only 2 colors) | |
| 34 | | nature-organic | `#166534` forest green | `#FEF3C7` cream | `#D4AF37` honey | |
| 35 | | jewel-tone | `#047857` deep emerald | `#FEF3C7` deep cream | `#D4AF37` polished gold | |
| 36 | | frost-ice | `#DBEAFE` pale cool blue | `#F8FBFD` near-white | `#3B82F6` steel blue | |
| 37 | | sunset-gradient | `#EC4899` deep pink | `#F97316` warm orange | `#9333EA` purple | |
| 38 | | earthy-dusty | `#C9A38C` muted dusty terracotta | `#E8E0D5` muted oat | `#A8B5A0` muted sage | |
| 39 | |
| 40 | ## Naming convention |
| 41 | |
| 42 | Each filename is `<palette>.png`, matching the corresponding file under `references/image-palettes/`. |
| 43 | |
| 44 | ## How to read this set |
| 45 | |
| 46 | Scan all 14 images side by side. Focus on: |
| 47 | |
| 48 | - **Color temperature** — cool-professional vs warm-human vs neutral-editorial |
| 49 | - **Saturation** — restrained vs vivid vs monochrome |
| 50 | - **Dominant/support/accent balance** — how strongly each semantic role appears |
| 51 | - **Accent weight** — how much visual presence the small accent commands |
| 52 |