From d75cdd9c0edfe90cc8c6ef04e7af4ddc1816eb10 Mon Sep 17 00:00:00 2001 From: Claude Date: Sun, 19 Jul 2026 07:04:31 +0000 Subject: [PATCH] =?UTF-8?q?plans/x265:=20record=20PROBE-SPLAT-=CE=BC-HYDRA?= =?UTF-8?q?TION-RHO=20leg=201=20results=20(DROP=20for=20static=20=CE=BC)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The probe queued at the tail of the plan RAN (lance-graph crates/helix/examples/mu_hydration_probe.rs). Leg 1 verdict: KILL→DROP for static spatially-addressed μ-hydration (escalation ≈1.0). The ctrl_naddr contrast proves the mechanism — hydration is addressing-dependent, motion-only; μ-hydration ⊂ motion-hydration. Comma fence green (Steinhaus three-gap = 3, coprime full-perm). Full write-up: lance-graph E-3DGS-MU-HYDRATION-2. Co-Authored-By: Claude Opus 4.8 Claude-Session: https://claude.ai/code/session_01K3RyLEbuNSHxxB3NTTrGki --- .claude/plans/x265-sprite-replay-probe-v1.md | 25 ++++++++++++++++++++ 1 file changed, 25 insertions(+) diff --git a/.claude/plans/x265-sprite-replay-probe-v1.md b/.claude/plans/x265-sprite-replay-probe-v1.md index 425fb7fc..44348a7e 100644 --- a/.claude/plans/x265-sprite-replay-probe-v1.md +++ b/.claude/plans/x265-sprite-replay-probe-v1.md @@ -266,3 +266,28 @@ temporal tautology). "EWA = blasgraph semiring" DROPPED (depth-sorted Porter-Duf runs on a real trained Inria `.ply` (non-φ input via `splat3d::ply`): μ-hydration ratio + semiring-matches-EWA (commutative-⊕ vs sorted-over) + SoA-sweep throughput. PASS→[H] / KILL→DROP bands in the board entry. + +### Results — leg 1 RAN (2026-07-19, `E-3DGS-MU-HYDRATION-2`) + +Leg 1 (the synthetic-structured falsifier, before the real-`.ply` v2) ran on +lance-graph `crates/helix/examples/mu_hydration_probe.rs` (std-only, +`helix::placement` public API, SplitMix64, 5 tests green): + +``` +set=ctrl_naddr rho=1.0000 residual=0.0000 escalation=0.0000 ← φ-phase at GENERATION index: perfect +set=ctrl rho=-0.0198 escalation=0.9978 ← SAME cloud, MORTON spatial address: collapses +set=nonphi rho=-0.0676 escalation=1.0000 ← structured cloud, spatial address: no win +comma_phi_three_gap_distinct=3 comma_coprime_full_perm=1 ← comma fence green (Steinhaus ≤3 + coprime full-perm) +``` + +**Verdict: KILL→DROP for STATIC spatially-addressed μ-hydration** — escalation ≈1.0 +is precisely the KILL band; a spatially-addressed static cloud collapses to the +ordinary float-SoA `splat3d` already ships (no positions-free win). The +`ctrl_naddr` contrast (same data, `rho=1.0` at the generation index vs `rho≈0` at +the Morton address) proves hydration is **addressing-dependent, not data-dependent**: +the win needs a transmitted sequential generation index, which motion has (the +frame/path counter — the `E-X265-HEADTOHEAD-1` ∞-PSNR case) and a static cloud does +not. **μ-hydration ⊂ motion-hydration.** The comma is the decoupler: φ's +incommensurability (three-gap = 3) is exactly what makes the spatial address unable +to invert to `n`. The real-`.ply` `splat3d::ply` leg (v2) would confirm on trained +geometry but the mechanism already forecloses the static win.