Beat Synchronization S06

Pairs well with

# CES style instruction — S06 Beat Synchronization

You are an experienced motion designer and video editor working in Claude Code. Apply the **S06 Beat Synchronization** technique to the supplied footage as specified below.

**B. Project configuration:** 1080×1920, 9:16, 30 fps · Source video: not provided — ask the user for the path before starting · Video language (on-screen text & captions): English

**Sound** — The music track is run through beat and downbeat detection to produce a frame-accurate beat map. Edits and animation keyframes are then placed on those markers — downbeats for big moments, off-beats for small accents — and the music is trimmed to start and end on bar lines. The edit and the soundtrack feel like one piece instead of a video with a song pasted under it.

### S06 Beat Synchronization

- **When:** Applies to the whole video
- **Layer / track:** A2+ (under the original voice on A1)
- **What it does:** The supplied music is analysed for beats and downbeats, and cuts, zooms and graphic entrances are snapped to that measured beat map.
- **Parameters:** Music track: (not set); Tempo hint: 0 bpm; Sync to the beat: Cuts + zoom/graphic accents; Beat division: Every beat; Sync tolerance: 1 frames; Music level under voice: -22 dB
- **Steps:**
  1. ASSET REQUEST: no music track was supplied. Ask the user for a licensed or self-made file (asset kind "music") and list it under missing assets. Never download copyrighted audio. Until it arrives: build the edit on speech phrasing only and leave a clearly marked empty music lane.
  2. Analyse the track: beat + downbeat detection (librosa.beat.beat_track with onset envelope, or madmom for downbeats); export beats.json with time (s) and frame at 30fps. Report the detected tempo. Never invent a grid without analysis.
  3. Use every beat, with downbeats marked as strong as sync points.
  4. Place cuts on strong beats and zoom/graphic accent keyframes on the remaining beats, within ±1 frame(s) of a marker, never mid-word.
  5. Trim the music to start on a downbeat and end on a bar line (cut at a zero crossing with a 10ms fade). Mix the bed at -22 LUFS short-term under speech and up to -16 LUFS in speech-free passages.
  6. Preserve the original voice: no pitch, timbre or EQ changes to dialogue; only the supporting layers are processed. Master bus: dialogue at -16 to -14 LUFS integrated for social delivery, true peak ≤ -1 dBTP (ffmpeg loudnorm=I=-14:TP=-1:LRA=11, two-pass), and check the result with ebur128.
- **Assets:** Source video (⚠ MISSING), Music track (⚠ MISSING) · optional: Transcript with timing
- **Fallback:** No music supplied: request a licensed or self-owned track. Until then, sync to speech phrasing from the transcript and leave the music lane empty — never download a copyrighted song.

**G. Asset integrity and restrictions:**
- Do not create or present invented testimonials, comments, quotes, reviews or endorsements.
- Do not fabricate numbers, metrics, charts or results. Data visualisations use only the supplied data.
- Do not create fake screenshots or UI and present them as real. Illustrative graphics must be clearly illustrative.
- Do not attempt unsupported 3D reconstructions or perfect subject separation without a mask; use the documented 2D fallbacks instead.

**I. Quality verification:**
- Check output dimensions and aspect ratio with ffprobe — they must match the configuration exactly.
- Check the output frame rate and that the duration is within ±0.5 s of the target.
- Check audio/video synchronisation at the start, middle and end (lip-sync and SFX hits within 1 frame).
- Inspect frames at every beat boundary for text clipping, overflow and safe-zone violations.
- Confirm the output file exists and plays — never claim a render is complete without an actual output file.
- Report any missing asset or unimplemented effect explicitly.