Skip to content

Art and animation, written as code. Illustrations, loops, interactive web art, launch-videos and scored films in dozens of styles, identical on every render.

Record complete regression and film playback verification

Alex Greenshpunf21aacc02119
diff --git a/work/S7a/HANDOFF.md b/work/S7a/HANDOFF.md
index 8917b96..e38b02f 100644
--- a/work/S7a/HANDOFF.md
+++ b/work/S7a/HANDOFF.md
@@ -48,14 +48,21 @@ Masking restored and committed:
0 fail, exit 0. TypeScript exit 0. mix-regression-proof.mjs passes; disabling the
support EQ or full guard context each causes its own behavioral regression.
- review-bytes-commit.log and review-direct-commit.log completed with LOADER REVIEW PASS,
- both exit 0. The final combined loader and music-unit suites are still running.
+ both exit 0. The final combined loader suite passed (exit 0). music-unit passed 38 checks in 229 s (exit 0).
Half done or untouched:
- New acceptance with the corrected room/sinc is running (reviewed-mix-acceptance.log).
marimbaCurious passes 100% bars, worst margin 4.129 dB and balance 0.721 dB.
- harpTender/nocturne results and final JSON are pending. Earlier current-pack.json
+ harpTender passed 100% bars, balance 0.647 dB. Nocturne failed at 75% with balance
+ 1.553 dB in reviewed-pack.json. Commit 83319b6 strengthens the general accompaniment
+ response, and final-mix-acceptance.log is rerunning all three on that code. Earlier current-pack.json
and trial results are INVALID final evidence.
-- Final combined loader/music suite, final film proof and final review are pending.
+- Full suite passed again after 83319b6: 332 tests, exit 0. The revised cap-sensitive
+ fixture passes 4 checks; mutation proof fails once for each removed behavior.
+- Final film proof passes (exit 0), with scoredFilm-pack WAV MD5
+ 99c1cdfe77dc3d04fe687216524800e4, Node/browser identity, offline playback and
+ 44.1 kHz device playback of the 48 kHz recorded buffer. No-pack hashes match.
+- Final acceptance and final code review are pending.
- No scoped review fix remains untouched. No score or per-piece numbers changed.
An initial default-install regression used music calibrate, which can regenerate
diff --git a/work/S7a/film-proof.json b/work/S7a/film-proof.json
index 493e6ba..d32359a 100644
--- a/work/S7a/film-proof.json
+++ b/work/S7a/film-proof.json
@@ -31,8 +31,8 @@
},
{
"label": "scoredFilm-pack",
- "md5": "5658bd32a3d2d8276c957d9ccf7c4991",
- "browserPcmMd5": "a7b950001e3595135bd1473e40253a99",
+ "md5": "99c1cdfe77dc3d04fe687216524800e4",
+ "browserPcmMd5": "67155d33629726b944ff6a5ebe488e71",
"frames": 672000,
"recorded": true,
"proof": "Node/browser sample identity",