expM — PARALLEL-CONCAT + RATE-SYNC · first readout

Recipe. Source = 5 clean ctx latent frames as a second token half (frame-aligned RoPE, told pinned .02); target = full 21 frames from pure noise everywhere (PatchForcing init). Knownness sets only the descent rate: T_i = 1 − 0.5·k_i; known tokens land early, x0-project, pin as clean attention references (expH). Training field == inference calendar family (branch1 exact slice t = clamp(1−c/T_i), branch2 AsyncPatch-style bounded per-patch clock jitter). No warp / z_mem anywhere; content rides self-attention only.

Reading (honest). ① expM trails expL_nk by ~1.8dB but the comparison is not matched: expM is at 4500/10k steps (expL at 9500) and expL_nk infers with SDEdit init from z_mem — geometric alignment is free; expM starts from pure noise and moves all content via attention. ② Plücker vs no-pose: +0.12dB — not separated yet at 4500 (zero-init channels engage late); verdict deferred to 10k. ③ Bright spot: expM's unseen-gray is markedly richer (−33…−45 vs −17…−30) — the pure-noise start + rehearsal preserves generation energy; the bottleneck is alignment, not content (mild CameraAnything pattern). Next: matched-step rerun @10k, add expL_uni row (equal information budget), per-block AdaLN camera injection if Plücker channel stays flat.