https://tube.sethpc.xyz/videos/watch/23e6ef2e-0d4c-41b2-b896-83262c301559 Session 38, iter64. A/B vs iter63: identical config + --deterministic-noise (EulerDiscrete scheduler + one fixed seed for the whole track) instead of Euler Ancestral + per-frame seed. Hypothesis: the per-frame stochastic re-roll WAS the 'flashing combos'; removing it should give smooth morph driven only by the chain + onset B-pulses + linear_scene B-injection. Bitrate halved vs iter63 (10.6 vs 19.7 Mbps) = much less high-frequency frame-to-frame change. Watching for: smooth morph (good) vs converged-static (S35 risk). Fri, 14 Aug 2026 05:03:23 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://tube.sethpc.xyz https://tube.sethpc.xyz/lazy-static/avatars/8e6b7b73-cbe8-40ac-9236-7b0afee31de7.png https://tube.sethpc.xyz/videos/watch/23e6ef2e-0d4c-41b2-b896-83262c301559 All rights reserved, unless otherwise specified in the terms specified at https://tube.sethpc.xyz/about and potential licenses granted by each content's rightholder.