# ✅ The five screens, end to end — what each one is actually worth

**Status:** ✅ measured, 2026-08-29. One page, because the answers for these five
screens are otherwise spread across a dozen documents and none of them says how
good the result *is*.

Rendered with `sylpheed-cli screen render` (adding `--all --black` for the two
splash halves) and correlated against the committed live captures.

| screen | build | elements drawn | corr vs capture | alignment |
|---|---|---|---|---|
| title | 4 | 15 / 24 | **+0.9500** | dy=0 dx=0 |
| main menu | 5 | 11 / 16 | **+0.9460** | dy=0 dx=0 |
| `EXTRAS` | 6 | 13 / 18 | **+0.9440** | dy=0 dx=0 |
| publisher splash | 10 | 2 / 3 | **+0.9600** | dy=0 dx=0 |
| developer splash | 11 | 6 / 7 | **+0.9643** | dy=0 dx=0 |

Every one aligns at **exactly zero offset** — searched ±2 px in both axes — so
placement and scale are right and the residual is tone and detail, not geometry.

## Every omission is accounted for

The counts are not slack; each undrawn element has a reason already documented:

| screen | not drawn | why |
|---|---|---|
| title | `ptlogo1`×2, `ptlogo2`×2 | `kind & 0x4` **ghost instances** — the draw capture shows one quad per wordmark |
| | `pteff00.prm`, `pteff02.prm` | `.prm` primitives, off without `--primitives` |
| | `ptloop01.rat`, `ptloop02.rat` | `loop*` animations, off without `--animated`; at rest both sit off-screen |
| | `pteff04.t32` | the **`8AX` name mismatch** — the art *is* on screen, drawn via `ptbase2` ([8AX](structures/ui-8ax-fullres-background.md)) |
| main menu, `EXTRAS` | 2 `.prm` + 2 `loop*` + `pteff05.t32` | same three reasons |
| both splashes | `palogo_eff0.prm` | `.prm` primitive |

4 + 2 + 2 + 1 = 9 on the title; 2 + 2 + 1 = 5 on each menu; 1 on each splash.
**Nothing is unexplained.**

## What the residual is, per screen

Ranked by what a port would notice:

1. **Tone.** The capture is darker than the composite by a gamma of ≈1.34–1.49
   ([tone curve](structures/ui-render-tone-curve.md)) — the game's own display
   ramp, not an emulator artefact. This is the largest single contributor to the
   0.94–0.96 rather than higher.
2. **`8AX` resolution.** The background is drawn from the 640×360 `ptbase`
   upscaled 2×, where the game uses the native 1280×720 `8AX`
   ([measured](structures/ui-8ax-fullres-background.md), two screens at 68 % of
   ceiling).
3. **Paint-order ties on `EXTRAS` only.** One drawable overlapping tie
   (`ptframe3`×`ptframe4`), and the capture is consistent with our order
   ([check](structures/ui-paint-order-derived-check.md)). The other four screens
   use orders measured from the game.
4. **Rotation is decoded but not rendered** — and does not affect any of these
   five at rest ([rotation](structures/ui-keyframe-rotation.md)).

## Reach

* These are **static composites at the resting pose** against **single frames**.
  Nothing here says the animation is right; Q1's interpolation law is settled but
  a multi-keyframe group's absolute timing is not
  ([time unit](ui-keyframe-time-unit.md)).
* Correlation on a whole frame is dominated by large areas. It is a sanity figure,
  not a per-element check — the per-element work is in the documents linked above.