11 September 2026 · V4 released as a work-in-progress
Three minutes of dance, gardens inspired by Đà Lạt, nightclub light and shared creation. This is the preview we finally chose to share—not a claim that the technology or the film is finished.
I know this is not perfect. After a long, tiring production process, I wanted to share where we had reached rather than keep the work hidden forever. I am proud of the progress, not blind to the mistakes. I will keep exploring video production so future films become better. Thank you for listening to the music and its intention while we learn in public. — Chicoman
The producer spotted characters appearing twice and then disappearing. His viewing estimate is that lip-sync is only about 50% convincing; this is a subjective assessment, not a measured accuracy score. Faces and identities drift, some hands and props deform, and movement or contact can look unnatural. Not every sung phrase is convincingly assigned to the right visible performer.
Music first: the existing Suno song and original master remain the soundtrack. Lead and backing-vocal stems, lyric cues and the vocal-director timeline informed performance planning. The written lyric sheet and the actual sung master are not identical; timings must follow the audio, not assumptions from the text.
Identity and direction: approved Chicoman character references and wardrobe studies guided generation. Earlier Runway and local LTX experiments exposed static framing, face-heavy edits, inconsistent likeness and unreliable synchronization. The new direction called for full-body movement, longer scenes, human warmth, art-making and glamorous, non-explicit adult performance.
Shot generation: H3 Max Reference-to-Video supplied the moving scenes for this cut. Direction favored approximately 15-second scenes, coherent costumes and environments, and people doing something. The three-person scene and eight-person finale were regenerated before R2. Director was explored separately; this release does not establish that interactive steering solves continuity or identity.
Assembly: HyperFrames 0.8.34 assembled 12 scene slots with the original master. R2 is 1344 × 768 at 24 fps, about 180 seconds, with 4,322 decoded frames. Source framing was retained without digital zoom. Generated scene audio was muted; the original song remains the sole soundtrack. Earlier renders and online releases were preserved.
DeepSeek vision through OpenRouter inspected all 4,322 decoded frames in batches. Deterministic checks found no black frames, exact duplicate frames or freeze runs meeting the three-frame threshold. The vision scan returned 26 flags, grouped around a glowing forearm, tray-holding hands and the last two frames. A separate Gemini audiovisual review and playback inspection supplemented the scan. Yet the producer still found duplicate/disappearing people and poor lip-sync. Checking every frame did not mean understanding every temporal error.
Frame-by-frame inspection alone does not prove stable character count across time, exact real-person likeness, correct mouth-to-voice ownership or phoneme-level synchronization. It also cannot certify pronunciation, cultural nuance, natural body motion, believable acts of kindness or emotional impact. Future tests should track each performer across a whole shot, compare identities against approved references, evaluate visible mouths against isolated vocals, and include native-language and human editorial review. Other defects may remain: this list is open, not exhaustive.
The final DeepSeek frame scan recorded US$0.70765. This is one QA pass, not the total film budget. Generation, retries, other reviews, subscriptions and human time are not reconciled into a verified total here.
The YouTube preview is unlisted and disclosed as AI-generated. YouTube reported no copyright issues at upload; that is not a legal clearance or an artistic-quality verdict. Publishing its link here makes the preview discoverable from this site. Older versions remain part of the story below.
Source: the original lyric sheet embedded in the archived Suno MP3, also shared by multiple song variants. Original wording is preserved; formatting and performance directives are omitted. These are written lyrics, not a certified verbatim or time-aligned transcript of V4. The master repeats, overlaps and omits passages, including the written “Vision becomes decision” block. Meanings and intentions below are editorial readings, not additional sung words. English annotations are included on every language version; localized annotations follow where applicable.
Je vois plus loin, Je choisis mon chemin. Une idée dans la main, Un monde pour demain.
(I see further; I choose my path; an idea in my hand, a world for tomorrow. Intention: imagination becomes a deliberate direction.)
Deovia… Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
Je rêve, donc je crée. Je crée, donc je deviens.
(I dream, therefore I create. I create, therefore I become. Intention: creativity shapes who we become.)
Every thought plants a seed, Every choice shapes what I see. Where attention flows, life grows, Where intention leads, I go.
(Intention: attention, thoughts and choices are seeds for the life we build.)
I give the dream a name, I give the spark a flame. Creation changes what I see, What I create creates me.
(Intention: turn an inner spark into action; what we make changes us in return.)
Deovia… Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
See it clear, choose with love. Plant it deep, rise above.
(Intention: choose with love and nurture the choice with commitment.)
Deovia… Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
Dream to form, form to flow. Give it life, let it grow.
(Intention: let ideas take form, move and grow.)
Creo el mundo que imagino, Cada paso abre camino. Siembro amor, florece vida, Cada luz encuentra salida.
(I create the world I imagine; each step opens a path. I sow love, life blossoms; every light finds a way out. Intention: love and small actions create possibilities.)
Sueño grande, vivo presente, Corazón claro, mente consciente.
(I dream big and live in the present, with a clear heart and a conscious mind. Intention: combine vision with awareness now.)
Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
Lo que creo, crea en mí. Lo que siembro, vive aquí.
(What I create creates within me; what I sow lives here. Intention: creation transforms the creator from within.)
Une pensée devient vision, Une vision devient action. L'action devient création, La création, évolution.
(A thought becomes vision; a vision becomes action. Action becomes creation; creation, evolution. Intention: follow an idea through to meaningful change.)
Je donne au rêve une voix, Et le futur chante déjà.
(I give the dream a voice, and the future is already singing. Intention: make a possibility tangible by expressing it.)
Vision becomes decision. Decision becomes creation. Creation becomes liberation. Liberation, elevation.
(Intention: decision and creation can lead to freedom and growth. This written passage is absent from the V4 master.)
Tikkun Olam, cœur en lumière. Love in motion, everywhere.
(Tikkun Olam evokes repairing the world; cœur en lumière means a heart in light. Intention: put love into action and contribute to a kinder world.)
Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
Gieo yêu thương, nở bình an, Tâm sáng soi bước dịu dàng.
(Sow love, let peace blossom; a bright heart lights gentle steps. Intention: cultivate peace through everyday kindness.)
Ước mơ gọi, tim đáp lời, Từng ý đẹp nở thành đời.
(Dreams call, the heart answers; each beautiful intention blossoms into life. Intention: respond to hope with action.)
Tôi trao đời một ánh sáng, Đời trao lại ngàn mùa sang.
(I offer life a light; life gives back a thousand radiant seasons. Intention: generosity can begin a cycle of flourishing. This is a poetic reading of “ngàn mùa sang,” not a literal guarantee of reward.)
Deovia… Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
See it clear, choose with love. Plant it deep, rise above.
(Intention: choose with love and nurture the choice with commitment.)
Deovia… Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
What I choose, chooses me. What I create, sets me free.
(Intention: our choices shape us, and creative action can bring freedom.)
Je vois. Je crois. Je crée. Creo. Camino. Viviré.
(I see. I believe. I create. I create. I walk. I will live. Intention: carry vision into an actively lived life.)
Tikkun Olam…
(Tikkun Olam: repairing the world. Intention: reconnect personal creation with care for others.)
One small seed can shape tomorrow.
(Intention: even a small beginning can influence the future.)
Deovia…
(Intention: Deovia is the recurring name/refrain, not a phrase with a literal English translation.)
What I create, creates me.
(Intention: the closing thesis—what we create also creates us.)
Chicoman creation journal · Work in public
One song. Four languages. A real dance take. An evolving collaboration between human taste and generative tools.
This page documents the entire making of the music video—from the first ChatGPT conversation and rapid Suno experiments to lyric refinement, the first public preview and the exact-face canonical-cast Preview V3.
Production journal · September 8, 2026
The first cuts proved parts of the pipeline could work. They also showed what was missing: room to move, recognizable faces, and a reason to watch again.
The director asked for more sensual, glamorous styling to draw viewers toward the song's message. The technical sheets felt too rigid, so the next direction pairs expressive posture with individual stage wardrobes. Attraction is part of the brief, alongside artistry, kindness and human connection. Broader reach and replay value are intentions, not measured results.
A lingerie-inspired image attempt was rejected by the image service and produced no deliverable. With the director's permission, the brief changed to opaque stage fashion. The rejected attempt is recorded, not presented as a success.
Six new portraits are published for review, not locked for video production. Amara's existing look is retained: her new portrait remains pending after generation attempts were rejected. Each profile preserves its earlier references. Chicoman remains unchanged while we await his photographs. The exact model variant is not exposed by the built-in generation tool.
Explore the cast's visual evolutionSeven fictional adult performers now have new multi-angle neutral sheets. New stage-costume studies are available alongside them; some wardrobe sheets remain pending after image-service safety rejections. The original references remain visible; every new sheet is a candidate, not an approved replacement. Chicoman’s new photographs are still pending. The decision: approve faces and costumes before spending another frame on V4.
Review the seven character sheetsThe studies are beginning to meet inside the edit: a continuous creation scene and a solo dance now share a 38-second passage of the original song. A separate full-body vocal study explores longer performance without cutting at every pause. These are internal reviews, not finished releases.
Putting scenes together exposed a new weakness: the pottery activity did not clearly finish before the dance began. The next experiment gives the artists a shared moment of completion, rather than hiding the gap with an object-only insert. Lip-sync, Chicoman’s likeness, and the whole-team ending remain unresolved. Automated checks support judgment; they do not replace it.
Longer, human-led scenes: full-body dance, beautiful stagewear, warm scenery, shared art and acts of kindness. No digital zoom or empty-stage filler. Keep the original song intact and build toward an ending with the team together.
V4 remains in development. These studies are not a release or a certification of lip-sync. Earlier uploads stay online as part of the story.
Explore the version archiveOfficial AI-hybrid music video · Preview V3
Preview V3 keeps the exact real Chicoman capture as the continuous base underneath and outside the accepted 01:32.933–01:38.800 selective lip-sync window, adds five accepted canonical-character motion inserts and two canonical image studies, and preserves the untouched Déoviá master.
Exact-face canonical-cast preview · complete 03:00 picture
Five accepted canonical-character motion windows · two canonical image studies · one accepted selective Chicoman lip-sync window
Spend compute where motion carries meaning.
The animatic decides which shots deserve premium movement. Stillness, pans, dissolves and transformation hold the visual rhythm; complex dance, partner work and camera travel become rare events instead of three uninterrupted minutes of undifferentiated spectacle.
Thirty-seven-shot hybrid animatic · retained as the visual assembly and camera-planning authority.
Download v0.2Real dance take + generated pose keys · retained as motion and synchronization evidence.
Download v0.1 ↓Déoviá · Chicoman
Place performers on the full-song timeline, shape their entrances and test voice ideas. Your version stays in your browser and never changes the official production.
Public full mix · local-only edits · sanitized JSON exchange
v0.8 · Motion lab
Six approved persona-neutral studies established the motion language. Preview v1 promotes CREATE, PLATFORM, REPAIR and MOBILIZE to accepted silent Runway Seedance 2.5 inserts; SHARE and TEACH remain deterministic local studies.
00:16.292–00:19.750 · 83f @ 24 fps
A slow push and growing light turn an act of making into visible momentum.
01:11.458–01:14.375 · 70f @ 24 fps
Depth, light and a measured reveal test how access can read as an invitation instead of a product shot.
01:42.083–01:46.417 · 104f @ 24 fps
A broken surface gathers energy along its seam, making restoration feel active rather than decorative.
02:06.375–02:10.208 · 92f @ 24 fps
A measured glide follows a dish and seed-light passing from hand to hand, turning reciprocity into visible action without assigning a singer.
02:16.917–02:19.750 · 68f @ 24 fps
A close camera move follows hands routing a cable, making knowledge transfer physical without claiming lip-sync.
02:33.792–02:36.708 · 70f @ 24 fps
A living landscape gains forward pressure, carrying the final call to action without assigning a mouth.
What this is
Four silent persona-neutral Seedance 2.5 inserts · SHARE and TEACH remain local · no lip-sync claim
Spend
1,100 Runway credits · 1,000 accepted + 100 unused fallbacks · no top-up
Next gate
Review Preview V3 as a whole, then advance only identity-safe selective lip-sync.
Version archive · v0.1–Preview V3
The archive keeps each milestone visible without pretending every internal review reel is cleared for publication. What changed—and what we learned from it—is the real work.
A real Chicoman dance take was aligned to the untouched 03:00.072 master, establishing physical timing before the multi-character world existed.
DecisionKeep the first cut as motion and synchronization evidence—not as final casting or lip-sync.
Open public artifact ↗
Thirty-seven shots created the first complete hybrid spine. The separate H2 pilot then proved the exact six-second camera geometry for $0.
DecisionTest the camera and the edit before paying for a photoreal frame.
Open public artifact ↗
Owner playback corrected the Spanish and French handoffs, isolated the one-word ‘évolution’ voice and removed Linh Vũ from language blocks she should not own.
DecisionFollow the rendered performance and explicit listening notes—not assumptions about a voice.
Forty-three shots tested unique imagery, honest unassigned gaps and a clean separation between who sings and who may appear on screen.
DecisionMeaning and editorial truth come before spectacle.
All 23 owner-timed intervals, 42 ownership changes and 10 overlaps became the timing authority while screen presence remained an independent choice.
DecisionA singer can own a phrase without appearing in every frame of it.
CREATE, PLATFORM, REPAIR, SHARE, TEACH and MOBILIZE replaced passive holds with images of people and systems acting on the lyric.
DecisionShow consequences and invitations to act—not merely attractive portraits.
The map expanded to 59 unique slots and 4,322 exact frames at 24 fps, with six action sequences, reproducible renders and zero missing image needs.
DecisionThe full reel stays private while 22 slots remain review-open, consent-gated or blocking-only; v0.8 begins with approved neutral frames.
The complete three-minute picture combines the untouched master with four accepted silent, persona-neutral Runway Seedance 2.5 inserts at exact editorial windows. SHARE and TEACH remain deterministic local studies.
DecisionPublish one honest reviewable preview, disclose the 1,100-credit experiment and its unused fallbacks, and make no lip-sync or identity-transfer claim.
Open public artifact ↗The complete three-minute cut keeps the exact real Chicoman capture as its base, including underneath the accepted selective lip-sync overlay, and adds seven closed-cast character windows without introducing any new persona.
DecisionUse generated Chicoman only for the accepted 01:32.933–01:38.800 lip-sync window; keep the exact real capture underneath and everywhere else, discard candidate audio, and make no full-song lip-sync claim.
Open public artifact ↗Production law · Camera first
A cinematic reveal needs changing perspective, foreground parallax and a motivated arrival—not an image enlarged until it gets soft. Each paid shot now passes a resolution-aware camera gate before it can be quoted or generated.
Crane, dolly or aerial movement changes perspective through three depth planes. The lens stays fixed; a tiny post trim is optional.
84° and 63° hero compositions carry architecture, formations and spectacle. A 4K master defaults to zero crop.
63° and 47° shots hold the action nearer the center. Stabilization and all finishing moves share one 1.08× emergency cap.
OpenRouter currently exposes 2.5 only to 720p, so it serves close movement tests and inserts—not final wide choreography.
Eye level → five metres high
The signature ensemble reveal begins 1.8 metres from the lead, reverses eight metres while craning upward, and finishes on the complete V formation. Seed, dancers and silver-city architecture move at different speeds, creating a real air-zoom sensation without scaling the frame.
This geometric previs plays the exact 00:42.760–00:48.808 master passage. Six role slots expand into the V formation while a fixed 63° camera reverses eight metres and rises from 1.55 to five metres.
Honest label: this proves timing, blocking and camera geometry—not photoreal motion, final identities, anatomy or lip-sync.
Capability truth checked August 24, 2026 against the public OpenRouter video catalog. The router refreshes that catalog before paid work; if Seedance 2.5 later gains 1080p, it can graduate to the tighter 1080 profile instead of being permanently hard-coded to today’s limit.
Where it began
The first breakthrough was not a model. It was realizing that abundance had changed the creative job from generating everything to choosing what deserves amplification.
“I made roughly 200 songs in minutes. Some were actually good. What now?”
The genres were deliberately unstable: Vina house, dembow, country, blues, Dominican rap, deep house and spiritual focus music colliding at generation speed.
“Generation is abundant. Selection matters. Taste matters. Coherence matters.”
The answer became a tournament: fast audition, ruthless shortlist, then invest disproportionately in one song and one visual identity worth remembering.
The production contract
Intent → Specification → Candidates → Evaluation → Refinement → Artifact
The original orchestrator brief asked for cheap exploration first, explicit model and cost tracking, self-critique, a bounded preview, and a human stop before full production.
The full trail
Every stage leaves evidence: a conversation, an audio decision, a timestamp, a rejected frame, a visual rule or a render that can be watched and challenged.
A burst of rapid Suno experiments mixed Vina house, Dominican dembow, country, blues, rap and spiritual focus music. The volume was exciting, but it exposed the real bottleneck: not generation—selection.
CompleteThe first ChatGPT discussion reframed the chaos as a tournament: audition quickly, preserve the strange combinations that work, and spend real attention only on the strongest survivor.
CompleteFrench, English, Spanish and Vietnamese lyrics gathered around one thesis: attention becomes intention, intention becomes action, and what we create changes us in return.
CompleteThe title became Déoviá: French elegance, Vietnamese tonal movement and an open-ended invented word for a borderless creative world. Chicoman remains the artist behind it.
CompleteThe three-minute master arrived with twelve aligned stems. Bass, vocal clarity, structure and lyric timing were reviewed without pretending every generated pronunciation or written line survived unchanged.
CompleteA 03:14.98 portrait HDR dance take became the physical source. It was sampled into 390 analysis frames; thirty-six moments and four identity anchors were then re-extracted at full resolution.
CompleteSoft poses, identity drift, fused hands and one memorable three-hand render were rejected. The accepted direction became grounded, athletic and masculine, with modest natural muscle and readable weight.
CompleteThe real take, approved pose keys and untouched master were assembled into a full three-minute proof film. It is a production spine—not the final cast, lip-sync or visual universe.
Watch nowThree pixel-aligned Chicoman looks now prove the morph system: one identity, one grounded pose and one body silhouette, with tailoring, material and light changing on the beat.
CompleteThe first cast-direction frame introduces five fictional adult Vietnamese fashion dancers in a close forward formation. Individual identities, choreography anchors and motion tests come next.
CompleteChoreography references were reduced to mechanics, levels and camera relationships. Three new sheets now separate Chicoman power, lead-dancer articulation and magnetic crew accents without copying the source performers.
CompleteNine approved frames, the untouched master excerpt and restrained editorial motion now test the ensemble reveal and four-stage Chicoman transformation across the exact 00:42.760–01:12.760 proof window.
Watch nowA $200 owner ceiling now sits above a stricter $180 automatic stop. Every paid request is quoted, reserved, reconciled and recorded before another one can begin; the API key never enters the site or project files.
CompleteFour original adult Dominican identity candidates cost $0.14 total. Model A is now locked for Marea and Amara; each has a story function, creative motive and continuity ID before outfits, poses or animation multiply.
CompleteEvery hero request now declares its lens, height, physical trajectory, parallax, focus and end frame before it can spend. Seedance 2.0 earns the 4K wides; lower-resolution shots are framed tighter and nearly final.
CompleteEight linked profiles now preserve stable IDs, intentions, movement and camera rules, explicit approval-state continuity assets and IMDb-style appearance credits—plus machine-readable manifests for future AI agents.
LiveA local 6.048-second geometric camera previs tests the first ensemble reveal against the exact master. Its first opening composition cropped performers, so it was rejected and corrected before any paid generation or identity transfer.
Watch pilotLyrics lab
The timecoded editorial transcription follows the generated performance as it actually appears—not just the written sheet it began from.
“Je rêve, donc je crée.”
Dream becomes creation.
“Creation changes what I see.”
The central feedback loop.
“Creo el mundo que imagino.”
Imagination enters motion.
“Gieo yêu thương, nở bình an.”
Plant love; peace blooms.
“One small seed can shape tomorrow.”
The closing seed image.
“What I create, creates me.”
The song resolves into itself.
Song analysis · Exact master + aligned guides
Déoviá works because its philosophy is not merely described: the lyric repeatedly turns thought into choice, choice into movement, and movement back into identity. The strongest line is also the whole dramatic engine—what the creator makes eventually remakes the creator.
180.072 s
untouched 48 kHz master
93.96
measured BPM · ~94
B♭ minor
estimated song key
−14.5 LUFS-I
−1.6 dBTP · no gain
Why it works
Where it needs discipline
Emotional architecture
A private thought receives a name, a direction and a future.
Attention, choice and action turn the philosophy into cause and effect.
Déoviá stops being an idea and becomes a physical, repeatable invitation.
Creation becomes embodied confidence: the imagined world is actively chosen.
French process language expands into Tikkun Olam: personal creation acquires an ethical horizon.
Love, peace, multilingual returns and the seed image resolve the thesis: creating the world also recreates the creator.
Keep the master. Refine through contrast, not more volume.
This is a future test, not a claim about the current master: keep one smooth spiritual dance spine. Try reggae and dembow in the offbeat pocket, country or blues color as brief texture, and a recurring accordion motif as a human voice. A melodic rap cadence may add edge without forcing a separate rap section into a song whose present strength is continuity.
Tempo was remeasured from the isolated drum guide; the estimated B♭-minor key was checked across the master and aligned vocal guides. Loudness and duration come from the untouched master. These guide stems are for analysis only—the final soundtrack remains the master.
Movement laboratory
Chicoman’s continuous dance take established weight, timing and personality. AI did not invent the body language from zero; it was asked to clarify selected moments.
Grounded weight, broad stance, strong lines, purposeful wrists and fists.
Modest natural muscle—athletic enough to carry the move, never inflated.
Extra limbs, fused hands, identity drift and decorative softness do not get rescued.
Phase 2 · In progress
The new direction combines Chicoman’s outfit morphs, an original adult Vietnamese fashion-dance ensemble, lyric-shaped environments and effects that visibly respond to movement. These are visual-bible studies—not claims of final casting or finished video.
Oxblood architectural tailoring with a long asymmetric movement line.
Matte-black dance techwear traced by restrained cyan energy seams.
Ivory structure and warm-gold piping for the final transformation.
Matte titanium-weave peacekeeper uniform with restrained cyan-gold light paths.
Weaponless charcoal wrap tailoring with a calm technological halo and forearm light arcs.
Pearl-gold finale tailoring with low-glow circuitry and calm architectural symmetry.
A cold-platinum full-body morph—skin, hair and wardrobe become one continuous material.
Cast identity pass 01 · linked profiles
The group study is becoming a reusable cast bible. These are original fictional adult Vietnamese performers, ages 23–26—not real models, endorsements or final casting.
Each identity gets one movement specialty and one visual responsibility. That separation makes choreography easier to read and gives identity-reference tools a smaller, testable job.
@linhvu · Flagship singer-dancer
Center lead · vocal close-ups · quiet editorial authority
@dieuan · Formation + partner work
Forward groove · clean diagonals · Chicoman interaction
@hamy · House + isolation
Footwork accents · torso control · motion relay
Cast & credits · IMDb-style graph
Identity lab · Model A locked
These are original fictional adult Dominican characters, not real performers or celebrity copies. Identity comes first: choose one face, then generate turnarounds, movement and wardrobe from that lock.
Artist-producer · rhythmic challenger
Prove that joy, discipline and freedom can occupy the same beat. He moves from observer to challenger to Chicoman’s finale collaborator.
Model A locked · continuity sheet developed.
Download full resolution ↓
Alternative gravity study · not selected
Download full resolution ↓Singer-dancer · emerging Caribbean-pop auteur
Embody “Creo el mundo que imagino” as self-authorship. Her sensuality comes from athletic control, eye-line and creative agency—not from being somebody else’s muse.
Model A locked · fuller natural curvy-athletic continuity developed.
Download full resolution ↓
Alternative silhouette study · not selected
Download full resolution ↓$200
owner ceiling
$180
automatic stop
$0.14
spent so far
4
paid identity images
To direct an edit, send the visible ID plus one change—for example: “DR-02-A · sharper eye makeup, keep her exact face.” Chicoman’s real identity photos remain local until external reference transfer is explicitly approved.
Movement pass 01
A public BN Dance Team reference and Chicoman’s local HD copy were sampled for levels, pivots, relays and camera energy. No source frames or performer identities appear in these generated keys.
Traveling knee drive · planted chest lock · diagonal sweep · pivot rebound
Rising starburst · low anchor · diagonal relay · asymmetric knee frame
Hip-led isolation · torso-wave relay · low triangle · over-shoulder pivot
No-spend proof · v0.1
This is an honest static concept animatic: nine approved images, cinematic pans and zooms, 0.5-second dissolves and the exact Déoviá master excerpt. It proves the edit plan—not generated dance motion, lip-sync or final effects.
00:42.760–01:12.760 · exactly 30.000 seconds
Static-image animatic · original master audio · no paid generation
Face, anatomy, movement and camera stay locked. Seven looks transform; the person does not.
The 30-second static animatic now tests the ensemble reveal and four-stage outfit escalation against the real master.
Lock one identity per character before side views, outfits or video. Premium motion still waits for an accepted draft and a named animatic window.
Toolchain
AI-assisted is not a single button. It is a chain of specialized tools joined by naming, timestamps, review rules and human decisions.
Rapid musical mutation, arrangement exploration, generated lead vocal and stem separation.
Creative dialogue, orchestration contract, lyric mapping, media analysis, shot planning and production code.
Audio probing, timeline alignment, HDR-to-SDR conversion, frame extraction and final assembly.
Built-in Codex image generation created continuity and movement studies without using OpenRouter credit. A separate $0.14 OpenRouter pilot proved the guarded external API workflow.
Chicoman chooses the song, meaning, name, movement, visual standards, edits and what deserves publication.
The public journal, media delivery and a durable home for the evolving process.
What comes next
Preview V3 is now the exact-face review authority. The next pass is selective lip-sync that must preserve Chicoman’s identity before any result may enter the edit.
This is an AI-assisted work. Chicoman owns the direction, selection and final edit. Generated music, vocals and imagery are disclosed as part of the story, not hidden behind it.
This page is also the working prototype for Deovia.com: a future official Deovia Studio for music, videos, films, games, apps, communities and the transparent creative systems connecting them.
“What I create,
creates me.”
Creation journal started August 23, 2026 · Last updated August 31, 2026
