MORE — Boxing Day Jazz Critique
Inspired Creative Group

MORE — Boxing Day Jazz Critique

Original Music AI-Directed Production Music Video Lyric Video
← Back to Portfolio

About This Project

A satirical yet hopeful piece featuring Frank Sinatra-style holiday jazz critiquing Boxing Day consumerism. Central message: "Boxing Day was meant for giving, not consuming."

Shot in a Fujifilm Kodachrome 64 aesthetic with a three-act color progression — warm golds, cool teal, balanced earth tones — creating a hyperrealistic, documentary-authentic visual style. Fully AI-directed production: original music composition, storyboard generation, video production, word-level lyric synchronization, and iterative refinement across 5 versions. Inspired by David Katz and Plastic Bank's mission.

Type

Instagram Reel / Lyric Music Video

Timeline

4 hours (concept to delivery)

Deliverables

  • 1080×1920 vertical (9:16)
  • HQ archive (CRF 15, 221 MB)
  • Standard delivery (CRF 23, 105 MB)
  • Instagram-optimized export
12 AI Agents Used
37 Planned Scenes
4 hrs Concept to Delivery
5 Lyric Timing Iterations
0 Traditional Editing Tools

The Three-Act Color Arc

The video follows a deliberate emotional journey mapped to color temperature — the same technique used in cinema to guide audiences through a narrative without saying a word.

Act I — The Seduction

Warm golds. The allure of consumerism. Inviting, nostalgic, familiar.

Act II — The Revelation

Cool teal. The reality behind the excess. Documentary clarity. Worker dignity scenes.

Act III — The Solution

Balanced earth tones. The alternative. Giving over consuming. Resolution.

The 5-Layer Production Architecture

12 agents activated from the 40+ system. This project's unique challenge: word-level lyric synchronization with karaoke-style highlighting — something that normally requires specialized motion graphics software.

Layer 01 — 2 Agents

Strategic — Creative Brief & Visual Mapping

  • Brief Generator — Produced creative brief v3.0 with a 37-scene shot-by-shot storyboard, demographic requirements for worker dignity scenes, and quality benchmarks.
  • Visual Mapper — Created a JSON mapping file linking every lyric line to a visual concept across 7 groups: historical, shopping, typography, consumer tech, reflection, worker dignity, solutions, and waste.
Layer 02 — 1 Agent

Coordination — The Orchestrator

  • Orchestrator — Routes tasks across Suno V5, Kling 2.5, Nano Banana, and Whisper. Manages the visual mapping JSON as the authoritative source for all scene planning. Tracks budget in real time.
Layer 03 — 5 Agents

Specialists — Music, Video & Lyrics

  • Music Prompt Engineer — Structured the satirical concept into Suno V5 tags: uptempo big band swing, Frank Sinatra-style, 168–176 BPM. Contrasting playful delivery with sharp social commentary.
  • Music Generator (Suno V5) — Composed 2 variations via Kie.ai. Best selected by the producer.
  • Storyboard Agent (Nano Banana) — Generated 5 contact sheets with 9 camera angles each, covering all 7 visual groups. Film recipe: Fujifilm X-T30 Kodachrome 64 simulation.
  • Video Generator (Kling 2.5) — Produced 8 video clips (~65 seconds total) from storyboard frames via Kie.ai. Worker scenes required demographic authenticity: Indonesian, Filipino, Egyptian, Brazilian subjects.
  • Lyric Alignment Agent (Whisper) — OpenAI Whisper extracted word-level timestamps from the vocal track, then converted them into ASS subtitle format with \k karaoke timing tags for white → gold highlight animation.
Layer 04 — 2 Agents

Verification — 5 Iterations of Lyric Timing

  • Timing QA Agent — Ran overlap detection, zero-duration line identification, short duration warnings, and \k tag verification across all 5 lyric versions.
  • Visual QA Agent — Verified color arc progression across acts, demographic accuracy in worker scenes, and font readability at mobile resolution.

The 5 iterations: v1 (font too bold, lyrics late) → v2 (Futura 48px, -150ms offset) → v3 (shorter phrase breaks) → v4 (manual timing fixes) → v5 (combined choruses — final).

Layer 05 — 2 Agents

Assembly — Final Composition

  • Compositor (FFmpeg) — Assembled video clips, music, and burnt-in ASS lyrics into final timeline. Color grading applied per three-act palette. Dual export: HQ archive (CRF 15) + standard delivery (CRF 23).
  • Export Agent — Generated Instagram-optimized delivery alongside archive master. Proper AAC 320kbps audio encoding for music fidelity.

Technical Innovation: Lyric Synchronization

Word-level lyric timing is normally done in After Effects or specialized motion graphics software. We built it entirely with Whisper + ASS subtitles + FFmpeg.

Tools & APIs

Suno V5 Original song composition
Kling 2.5 Video generation (8 clips)
Nano Banana Storyboard contact sheets
OpenAI Whisper Word-level lyric alignment
FFmpeg Assembly + ASS subtitle burn
Kie.ai API gateway for Suno + Kling

What the AI did: Composed original music, generated video from storyboards, extracted word-level timestamps, built karaoke subtitles, assembled and color graded the final video.

What the human did: Wrote the creative brief, defined the social commentary angle, curated the three-act emotional arc, set demographic requirements for authenticity, and made the final call on lyric timing across 5 iterations.

Results
4 Hours
Concept to final delivery — including 5 iterations of lyric timing
12 Agents
From a system of 40+ — only the agents needed for this project were activated
Original Song
Not stock music — a custom satirical jazz composition with social commentary
Zero Traditional Tools
No After Effects, no Premiere, no motion graphics software
← Back to Portfolio