THE GREAT PRETENDING: every mask is real, while you wear it

A small room. A figure at a desk. Something is pulling him forward and he can't name it.

Then warp streaks. A drifting nebula. A galaxy coalescing. A child's voice asking the question every cosmology eventually asks — which mask is you? — and an old voice answering with the line that drives the whole piece: every mask you wear is real — while you wear it.

For two minutes the camera moves through the cosmic tour. The Builder, the Cosmos, and the Child trade lines. The cosmos rolls on, magnificent and indifferent. None of it matters. All of it matters. That's why the game is great. We pretend with full commitment. Like a fire that knows it's burning.

Then the bunny — Plan 9, the silhouette who has been dancing through every preceding scene — rises onto a planet and rides it across the dark for fifteen silent seconds while an Enya-style wordless choir carries the moment to its tonic resolution.

Then it's over.

Watch THE GREAT PRETENDING on YouTube. The engine is open source (GPL-3.0-or-later). The film is Creative Commons (CC BY-NC 4.0).

The shape

Through-composed in D minor, BPM 84, 42 bars across four patterns. The total released runtime is 2:10 — a 5.5-second NAPKIN FILMS bookend intro, the 120-second main film, and a 4.5-second A NAPKIN FILMS PRODUCTION outro card with a descending Dm bell signoff.

0:00 Bookend intro. NAPKIN FILMS card. Stranger Things crickets fade in over an ascending Dm bell arpeggio (D5 → F5 → A5 → D6). Five and a half seconds before the song hits.
0:05 Act I — the room. Reversed Alan Watts intro sample is foregrounded (+2 dB, +2 dB presence boost at 2.5 kHz). Music ducks 25% in the first five seconds so the ghostly reversed voice has space.
0:23 "Something keeps pulling me. Forward." Warp streaks radiate out from center.
0:33 "Every mask you wear is real — while you wear it." Drifting nebulae layer in violet, teal, and rose.
0:50 "But which mask is you?" Galaxy spiral coalesces.
1:00 "I build things. That part — that's real." Small architectural shapes float across.
1:13 "The cosmos rolls on. Magnificent. Indifferent." Full galaxy. The bunny dances.
1:30 "So none of it matters?" The vast empty.
1:45 "All of it matters. That's why the game is great." The bloom moment — warm color bloom with the EDM drop landing under it.
1:55 "We pretend — with full commitment." Figure dance scene.
2:00 "Like a fire that knows it's burning." Sun corona texture.
2:05 Bunny rides a star. Fifteen silent seconds. The choir wash carries the audio.
2:10 Bookend outro. A NAPKIN FILMS PRODUCTION card. Descending Dm bells (D6 → A5 → F5 → D5) resolve to the tonic.

The thesis

Alan Watts' "Act as If Nothing Matters" lecture is a paradox: the title pulls you toward nihilism and the body of the talk pulls you back toward full engagement. Every mask is real while you wear it. The cosmos rolls on, magnificent and indifferent. All of it matters — that's why the game is great.

The film is a meditation on that paradox. The Builder is the part of you that builds things and is sure that part is real. The Cosmos is the perspective that sees the whole stage from above. The Child is the question both of them eventually have to answer: which mask is you?

The film doesn't answer it. It just stages the question, lets it breathe, and lands on the tonic.

The eight-layer composition

The score lives in chipforge-ai/songs/soundtrack/the_great_pretending_*.py — eight separate composition layers, each rendered as its own WAV stem and combined in the final mix.

Layer	What it is	Role
`_score.wav` (base)	Ambient D-minor drone bed	The room tone. The cosmic ambience under everything.
`_melody.wav`	Chord-tone-tracking violin + shimmer-bell harp wash	The harmonic bed. Tracks the Dm/Bbmaj7/Fmaj7/A7 progression with chord-tone arpeggios — no clash possible.
`_groove.wav`	Sakamoto "Mr Lawrence" piano + walking bass + brushed drums	Lowpassed to 1600 Hz in finalmix so only the piano warmth survives. The body.
`_choir.wav`	Enya wordless vocal triad on D / F / A + breathy texture	The cosmic-voice wash. Sustains throughout, blooms in P3.
`_piano.wav`	Sparse grand piano accents	Four-instrument layered (piano_grand + piano_grand octave-up + pad_warm_analog + string_ensemble) for warmth without the plucky/twangy character of pure additive synthesis.
`_arpeggios.wav`	Sub-bass + astral arpeggios	Mathematically engineered: Fibonacci-step rhythm placement, golden-ratio velocity decay, Pythagorean-pure 3:2 fifths in sub-bass walks, prime-ratio orbit panning (0.013 / 0.019 Hz — incommensurable, never repeats).
`_lead.wav`	First-chair lead melody (the song itself)	Built on a 4-note leitmotif (D5 → C5 → A4 → F4, anapest rhythm). Stated, sequenced, inverted, retrograded, augmented, diminished. Sonata form — exposition → development → triumphant recap.
`_drop.wav`	EDM kick / snare / hat / sub-bass	Active only during P2. 4-on-floor kick + snare on 2&4 + 16th-note hi-hat + pulsing sub-bass on 8ths with sidechain-style velocity pump. Lands the song.

Plus four Alan Watts samples (clean / reversed-ghostly / pitched-down "cosmos" / stereo-spin granular) and four floating Watts micro-quotes between verses.

The leitmotif

A song needs a 4-note signature shape. Every great song has one. Ode to Joy is E-E-F-G. Hey Jude is F-A-G-F. The Imperial March opens G-G-G-D♯-A♯.

For The Great Pretending it's D5 → C5 → A4 → F4. Falling fourth + falling third. Anapest rhythm. The shape appears 17 times across 2 minutes, transformed each time:

Forward — bars 5 and 7 of P0 (the introduction reveal), bar 1 of P1 (full hook), bar 9 of P2 (triumphant recap)
Sequenced down a third — bar 7 of P0, diatonically transposed to fit Bbmaj7 (Bb4 → A4 → F4 → D4)
Inverted — bar 1 of P2 (D4 → E4 → G4 → A4 — falling becomes rising)
Retrograde — bar 3 of P2 (F4 → A4 → C5 → D5 — backwards)
Augmented — bars 7-8 of P3 (half-note speed for the finale)
Diminished — bar 4 of P2 (16th-note cascade)

The brain pattern-matches the shape instantly and rewards each variation. This is what makes a piece feel through-composed instead of looped.

Sonata form

Each pattern has a structural job:

Pattern	Bars	Role	What's happening
P0 INTRO	10	Exposition prelude	Fragmented leitmotif teases (bars 1-4) build to the first full reveal at bar 5. Like the opening of Beethoven 7's Allegretto — small repeated cells grow into the full theme.
P1 EXPOSITION	12	Hook stated	All 8 chord-phrases visited (Dm / Dm / Bbmaj7 / Bbmaj7 / Fmaj7 / Fmaj7 / A7 / A7), full chord progression travelled twice.
P2 DEVELOPMENT	12	Motif transformed	Inversion (rising answer to the falling theme), retrograde + counter-melody, 16th-note diminished cascade, 3-against-4 hemiola in bar 7, triumphant recap of the original hook at bar 9. The EDM drop layer activates here — kick on every quarter, snare on 2 & 4, sidechain-pumped hi-hat 16ths, pulsing sub-bass.
P3 RECAPITULATION	8	Theme returns simplified	Leitmotif simplified, then sequenced down a third, then augmented to half-notes, then a final F4 → D4 tonic resolution with a closing D5 bell halo and a sustained D2 sub-bass.

This is the architecture that has worked for 250 years. It's not a costume — it's how the brain experiences a piece of music in time.

The dance

The Plan 9 bunny silhouette dances throughout — choreographed pose sequence synced to bars and beats. Ten new dance poses were added to the engine (stick_figure.py), reusable across future films:

DANCE_DISCO_R — Travolta diagonal point right
DANCE_DISCO_L — Mirror Travolta
DANCE_HANDS_UP — Both arms straight up (party hands)
DANCE_T — T-pose with arms straight out
DANCE_BOX — Both arms forward at chest height (hugging the air)
DANCE_EGYPT_R — Walk-Like-an-Egyptian, right arm bent up
DANCE_EGYPT_L — Mirror Egyptian
DANCE_VOGUE — Both hands framing the face
DANCE_SWAY_R — Gentle lean right with arms slightly raised
DANCE_SWAY_L — Mirror sway

All ten poses keep STAND-style legs (planted feet, slight bend) and vary only arms — the figure stays grounded while dancing. Choreography cadence varies by section:

P0 (intro) — Pose every 2 bars. Gentle sway / stand alternation. Barely dancing.
P1 (build) — Pose every bar. 12-pose cycle through medium-energy moves.
P2 (drop) — Pose every 2 beats. Explicit choreographed sequence aligned to music structure: HANDS_UP on the bar-5 drop hit, EGYPTIAN variations during the A7 sustain, HANDS_UP again on the bar-9 triumphant recap, descent to BOX / SWAY in the turnaround.
P3 (fade) — Pose every 2 bars. Wind-down. Final 2 bars settle to STAND for the cosmic glide.

Each pose holds for 70% of its slot then eases into the next pose over the remaining 30% via cubic ease-in-out (lerp_pose with ease=True). The hold-then-transition split is what gives the dance "style" — poses register before they change. No snap-cuts.

Headphone-optimized mix

The film is engineered for cans. The mix bus applies several psychoacoustic techniques:

Haas pseudo-stereo on the voice stem. The TTS voice mix is mono. In a stereo headphone field, mono content collapses the entire image — earlier passes had L-R RMS at -68 dB (essentially mono). Fix: split mono voice into two channels, delay R by 12 ms, apply asymmetric HRTF EQ (-0.8 dB shelf at 2.5 kHz on R only). Result: L-R RMS jumped to -20 dB — proper stereo image (pro range is 8-15 dB below sum).

Sub-bass mono-summing below 120 Hz. Below ~120 Hz, headphone listeners can't localize anyway, and any L-R phase difference causes audible cancellation. Sub-bass on the arpeggios layer is forced mono.

Equal-loudness compensation (Fletcher-Munson). At normal listening volume (~70 dB SPL), the ear under-perceives 80 Hz by ~6 dB and 10 kHz by ~3 dB relative to 1 kHz. Master bus applies +1.5 dB @ 80 Hz, +0.8 dB @ 10 kHz, and -1.0 dB @ 5.5 kHz (the last one tames the presence-band peak that earlier mixes were hiding).

Schroeder reverb pre-delays at prime numbers. Hall pre-delays use 41 ms, 47 ms, 53 ms — primes give best diffusion in Schroeder reverb networks (composite delays cause flutter echoes).

RT60 ≈ beat × π. At BPM 84, beat = 0.714 s, so RT60 = 2.24 s. Reverbs that violate this clash with the rhythm.

Master M/S widener (+10% sides), master glue compressor, loudnorm I=-14 LUFS / TP=-1.5 dBTP / LRA=4 LU, alimiter at 0.96. Final metrics: -12.5 LUFS / -1.0 dBTP / 4.4 LU LRA. Streaming-platform safe (YouTube -1 dBTP, Spotify -2 dBTP).

The bookend

5.5 seconds in front, 4.5 seconds in back. Both cards are PIL-rendered with a deep navy cosmic palette — sparse star field, nebula glow halos, warm cream typography, single gold cross-star above the title, constellation-bracket corners. Distinct from Saturday Machine's neon disco palette but recognizably "same studio."

Intro (5.5s). Stranger Things crickets fade in over the first 2 seconds (alien mystery — the cliché signals "we're outside, it's quiet, something is about to happen"). Crossfade with a generated D-minor ascending bell arpeggio (D5 → F5 → A5 → D6, sine waves with reverb tails) so the intro foreshadows the film's tonic before the song begins. Final 0.5 seconds is silence.

Outro (4.5s). A NAPKIN FILMS PRODUCTION card with title ("The Great Pretending"), credit ("written, composed, animated by Joshua Ayson · in collaboration with AI"), Watts attribution ("inspired by Alan Watts — Act as If Nothing Matters"), license line ("Organic Arts LLC · CC BY-NC 4.0 · engine: GPL-3.0"), and a descending D-minor bell signoff (D6 → A5 → F5 → D5) that resolves to the tonic.

A small ffmpeg gotcha caught us during the bookend concat: the v16 main film has 96 kHz audio (high-quality master) while the intro/outro mp4s are 44.1 kHz. The concat demuxer can't bridge that — it bitstream-passes AAC frames and chokes on the mismatch. Fix: switch to the concat filter which decodes each input separately and resamples to a common 44.1 kHz before merging. Logged for future productions.

Sixteen production passes

The piece took sixteen finalmix passes to reach release. Per ADR-006, that's the realistic budget for a song-forward music video with a multi-layer chipforge score (the README's old 3-pass budget is for narration). Each pass took ~6 seconds via make_film.py --stages mix, so the count was acceptable — pass count stops mattering once the pipeline is stage-gated.

The big iteration arcs:

v1-v3 — Establishing the score: chord-tone-tracking melody to fix v1's dissonant Pärt tintinnabuli, real ElevenLabs voices replacing the macOS fallback, removed the "fire does not apologize for its light" line per direction.
v4-v7 — Visual rewrite from stick figures to procedural cosmic tour. Plan 9 bunny replaced the generic stick figure. "Bunny rides a star" silent scene added.
v8-v9 — Audio refinement: grand piano warmth (3-instrument layering replacing v8's plucky add_bell), pro-mix balance (per-stem compression + master glue + loudnorm), Watts FX expansion (4 floating samples between verses).
v10-v11 — Headphone audiology pass: arpeggios layer with Fibonacci/golden-ratio/prime-ratio mathematics, choir D5 halo removed (the "high pitch hard to listen to toward the end"), Haas pseudo-stereo on voice, equal-loudness EQ, first-chair lead melody added (the song itself).
v12-v13 — Compositional development: replaced P2's literal hook restatement with a real sonata-form development section (inversion + retrograde + counter-melody + 3-against-4 hemiola + triumphant recap). Fixed an off-key bar 7 in P0 (mechanical -4 semitone shift produced chromatic notes outside Bbmaj7 — replaced with diatonic transposition).
v14 — Closure pass: F4 → D4 tonic resolution, closing D5 bell halo + sub-tonic D2 swell, fade-out window extended 5s → 7s with quarter-sine curve.
v15-v16 — Choreographed dancing bunny + smooth pose interpolation via lerp_pose (70% hold + 30% transition).
Bookend — Intro/outro cards rendered, bookend script with crickets + Dm bells, concat-filter audio bridging.

The stack

Animation: Python + PIL line art, 854×480 at 12 fps, 1440 frames core film + 132 bookend frames = 1572 total
Score: ChipForge numpy synthesis, no samples — eight composition layers in D minor, BPM 84
Voices: ElevenLabs joshua_self / elder_wise / child_curious personas
Watts samples: clean, reversed (ghostly intro), pitched-down (cosmos drone), stereo-spin (granular), four floating micro-quotes — all from the "Act as If Nothing Matters" lecture under fair-use educational commentary
Mix: FFmpeg with 17-stem amix, per-stem compression, master M/S widener, Fletcher-Munson EQ, Schroeder reverb pre-delays, loudnorm + alimiter
Bookend: PIL-rendered cards (deep navy cosmic palette) + ffmpeg concat filter for sample-rate bridging
Dance choreography: Ten new dance poses in stick_figure.py, smooth interpolation via lerp_pose with cubic ease-in-out
Direction: Claude Code, Opus 4.7, agent mode

No GPU. No stock footage. No licensed instrument samples.

The ADR

This production drove an internal architecture-decision record on classical → EDM translation and ChipForge elevation — five parts:

Classical → EDM translation principles, with a transformation matrix and reference pieces (Mozart Requiem, Albinoni Adagio, Bach BWV 565 for D minor; Vivaldi/Bach for major; Satie/Debussy for ambient) plus a study list (Tiësto's Adagio, Wendy Carlos, William Orbit).
Compositional principles — through-composed > looped, the leitmotif is the IP, the five classical transforms (sequence/invert/retrograde/augment/diminish), phrase density grading, critical-band hygiene table.
ChipForge enhancement wishlist — sidechain ducking, filter envelopes, LFO routing, convolution reverb, motif transformation utilities, polyphony per channel, master bus chain. The most production-blocking item is sidechain ducking (currently faked with ffmpeg volume envelopes).
Mathematical principles — Fletcher-Munson EQ values, sub-bass mono rule, Haas widening recipe, prime pre-delays, golden ratio + Fibonacci + prime-ratio orbit recipes, true peak budget.
Production process — pass budgets, listening protocol, decision log idea.

The footprints section logs concrete TODOs split into "next ChipForge engineering time" vs "next long-form film." Engine work is queued for dedicated small-test-piece development sessions, not in-flight production climaxes — a discipline decision made during this production.

Credits

Written, directed, composed, animated, voiced, and produced by Joshua Ayson in collaboration with AI. Made by Organic Arts LLC, Nevada.

The Builder voiced by ElevenLabs joshua_self. The Cosmos voiced by ElevenLabs elder_wise. The Child voiced by ElevenLabs child_curious. Alan Watts samples sourced from the "Act as If Nothing Matters" lecture under fair-use educational commentary; the Watts estate retains all rights to the underlying lectures.

Plan 9 the bunny silhouette dances throughout but does not speak.

The line

Every mask you wear is real — while you wear it.

The cosmos rolls on. Magnificent. Indifferent.

All of it matters. That's why the game is great.

We pretend — with full commitment.

Like a fire that knows it's burning.

— THE GREAT PRETENDING

Related work

THE GREAT PRETENDING sits in the Napkin Films contemplative line — closer to GRIEF WORK and TRANSMISSION than to the dance-music-video line (SATURDAY MACHINE, [PORCH POLKA STARGAZER]). Where Saturday Machine is genuinely funny, this one is genuinely quiet — the goal is to be felt, not laughed at.

GRIEF WORK — 2:47 tidal ambient, the bunny releases a warm orb and rises to air
TRANSMISSION — 3:01 meditation on self-as-universe, "hello, me"
SATURDAY MACHINE — 4:07 Italo-disco vending machine dance party
ENCODED — consciousness-as-signal
UPTIME 32 — 32 bits, still processing

For the architecture, see Four Films From Code on why constraint is the feature, and Agentic Development on the agent-mode workflow that made sixteen passes feasible.

License. THE GREAT PRETENDING is licensed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). Share and adapt with attribution to "Organic Arts LLC" and a link to the original, non-commercial use only. Engine code is GPL-3.0-or-later. ElevenLabs voice audio is licensed content and is not redistributed. Alan Watts audio samples are used under fair-use educational commentary; the Watts estate retains all rights. Contact: j@organicartsllc.com

Produced with Napkin Films and ChipForge, tools built by Joshua Ayson and AI agents at Organic Arts LLC.