BEAUTY IN A PASSING MOMENT — REVIEW02 FumiTours / Fumi Videos 26 October 2026 (Australia/Sydney) Separate production credits. Do not add these to the film as an end screen. Script: existing Steve-approved Beauty in a Passing Moment / ON-J02 script, unchanged. 148 spoken words. The reflection on mujō and impermanence preserves the original caution that Japanese ideas of beauty have different histories and one word cannot stand for them all. Voice: Fumi-Voice-Proof03-Candidate02-Slightly-Deeper, the sole current approved Fumi voice target. Fresh Qwen 3 TTS 1.7B narration via Fal using the saved canonical embedding and recipe; approved half-semitone treatment applied once, no tempo change, echo or added reverb. The exact master remains unchanged. This finished performance is pending Steve's listening approval. Pictures: existing approved illustrative onsen, garden, doorway, welcome, and closing imagery. Recovered from the original complete three-film source package libfile_8eceb93689e481918013e4a24f1562b8. Five retained motion sources were generated with MiniMax H3 Max via Fal. Two retained artworks (garden and doorway) came from the original approved source collection. No new pictures were generated for this revision. These are illustrative settings, not named or independently documented real hotels, people or properties. Original source custody references: approved bath source libfile_c26444f3119c8191a3e72e6178a516ca and Mugen source libfile_f1faa88d796c8191b146d2397aba1be6. Sound: original designed water ambience, ElevenLabs Sound Effects v2 via Fal, request 01a1623c-5b25-7e73-8f23-afd1ea44df8a. Original piano excerpt retained: Frédéric Chopin, Nocturne Op.9 No.2, from Musopen’s Complete Chopin Collection, published under CC0. https://archive.org/details/musopen-chopin Track: NocturneOp9No2.mp3; source MD5 346a115438975aa7c9ae076ca877a955. The selected excerpt starts at 96 seconds. Piano stays softly behind Fumi. Collection metadata and font licences are included in the editable source. The old mixed soundtrack and superseded voice are not used. Typography: DejaVu Sans, Nimbus Roman and Noto Sans CJK JP. Font licence files are retained in the editable package under sources/fonts. Existing artwork titles and branding are retained; captions have explicit two-line spacing. Editing, sound mixing and finishing: FFmpeg, Pillow, NumPy and SciPy. Speech recognition for wording/timing checks: ElevenLabs via Fal. Original cultural-research references are retained separately in records/Cultural-Source-References.json; no new factual claim was added. Publication: no site or repository changes. Website writer should publish only after Steve approves this exact revised film. Older films and source archives remain preserved. No on-screen credit/attribution page.