# Project 1 — Deconstruction
**30 points · Weeks 2–6 · Sep 1 – Sep 29**
---
## The assignment
Choose an existing piece of professional work in your track — a title sequence, a special effects shot, a game mechanic, a character design, a level, a rig. Figure out how it was made. Then recreate it.
You may recreate it faithfully, or you may **recreate it in your own style** — same structure and same technical approach, your visual voice. Both are valid; decide with the instructor in Week 1.
**The creative direction is not yours, and that is the entire point.** Someone else already decided what this thing should be. That frees you to concentrate on the only question that matters here: *how do you actually get from nothing to a finished, professional-looking result?* Most students can identify good work. Far fewer have traced the path to it step by step. This project makes you walk that path with the destination already marked.
---
## Why this comes first
Three reasons it opens the semester rather than closing it.
**It removes the hardest variable.** Original work fails most often at the concept stage, not the execution stage — students stall for two weeks deciding what to make. Here that decision is already made, so you start producing on day one.
**It generates the raw material for Project 2.** You will learn several techniques getting through this. Project 2 asks you to take one of them and teach it. So the deeper you dig here, the more you have to work with later.
**It matches how studio onboarding actually works.** Your first month at a studio is deconstruction: a style guide you did not write, a rig you did not build, a shot that has to match work made by someone else. Reverse-engineering a finished result into the decisions behind it is a daily professional skill.
---
## Choosing your source
The rule is **reachable, not impressive**. You have five weeks. A modest source rebuilt to a high standard beats an ambitious one abandoned at 60%.
**Good source characteristics**
- Short and self-contained — one sequence, one shot, one mechanic, one asset, one system
- Made by professionals at or slightly above your current level
- Legible enough that you can analyze the decisions from the outside
- Contains at least two techniques you do not currently have
**Warning signs**
- Hero assets that took a team a year
- Work whose difficulty is invisible until Week 3
- Anything chosen because it looks impressive rather than because it teaches you something
Bring two or three candidates to Week 1. We settle it in consultation.
---
## Curated source lists by track
### Motion Graphics · Title Design · VFX Sequences
**[Art of the Title](https://www.artofthetitle.com/)** is the primary source for this track. It hosts high-quality video of title sequences alongside interviews with the designers who made them — meaning you can often read the creators explaining their own process, which is rare and extremely useful. Browse their **Top 10 Title Sequences of [year]** features to survey the field.
Sequences that reward deconstruction and are scoped realistically:
| Sequence | Studio / Designer | Why it teaches well |
|:--|:--|:--|
| *Se7en* (1995) | Kyle Cooper / Imaginary Forces | The foundational modern title sequence. Analog texture, scratched type, deliberate imperfection. |
| *Catch Me If You Can* (2002) | Kuntzel + Deygas | Pure 2D vector motion and silhouette. Very achievable; teaches timing and negative space. |
| *True Detective* S1 (2014) | Elastic / Patrick Clair | Double-exposure compositing. A clear, learnable technique with a strong result. |
| *Stranger Things* (2016) | Imaginary Forces | Type animation and glow work. Deceptively simple; the polish is all in the timing. |
| *Westworld* (2016) | Elastic / Patrick Clair | 3D, simulation, and lighting. More ambitious — scope carefully. |
| *Game of Thrones* (2011) | Elastic | Mechanical animation and camera choreography. |
| *Daredevil* (2015) | Elastic | Fluid simulation and material work. |
| *Severance* (2022) | Oliver Latta (Extraweg) | Surreal 3D character motion. Distinctive and technically instructive. |
| *Loki* (2021) | Perception | Type, texture, and transition design. |
Pick a sequence, not a film. A 60–90 second title sequence is a semester-appropriate scope; a music video is not.
### 3D Animation
Deconstruct **one shot**, 5–10 seconds, not a scene.
- A performance beat from an animated feature (Pixar, Disney, DreamWorks, Sony, Illumination) — pick a moment with a clear emotional turn
- A game cinematic shot — *Arcane*, *Overwatch*, *League of Legends*, *Cyberpunk 2077*, Blizzard's cinematic library
- A creature or quadruped locomotion beat from any of the above
- **[Blender Studio Open Movies](https://studio.blender.org/films/)** — *Sprite Fright*, *Charge*, *Coffee Run*, *Spring*. These publish the **actual production files**, so you can check your reconstruction against how it was really built. Unmatched as a learning source.
### Character Rigging
Rebuild the *system*, not the mesh.
- **[Agora Community rigs](https://agora.community/)** — free, professional, widely recognized
- **Truong CG Artist rigs** — strong deformation to study
- **Mery** (meryproject.com) — film-standard face rig
- **Malcolm / AnimSchool** — a long-running portfolio standard
- **Blender Studio rigs** — Rain, Snow, Vincent; open and readable
- **Unreal MetaHuman** — analyze its control logic and rebuild the approach
Analyze the control scheme and deformation strategy, document the design decisions, then rebuild that system on your own or a supplied mesh.
### Character Design · Concept Art · Storyboard
This track is where *"in your own style"* is most productive.
- A published character design from a production art book, ArtStation portfolio, or studio visual development release
- A boarded sequence from a film or series with published storyboards
- A production design or environment key from a released art book
Reconstruct the underlying thinking — the design logic, the shape language, the staging — then rebuild the piece. Rendering it through your own stylistic filter while preserving the structural decisions is often the more instructive choice.
### Technical Art
Rebuild a specific shader, VFX, or procedural system from a shipped game, working from observation.
- Toon and cel shading — *Genshin Impact*, *Zelda: Breath of the Wild*, *Guilty Gear Strive*
- Environmental systems — *Ghost of Tsushima* wind and foliage, *Journey* sand, *Sea of Thieves* water
- Stylized VFX — *Ori*, *Hades*, *League of Legends* ability effects
- Destruction and physics — *Control*, *Teardown*
- **[Real Time VFX](https://realtimevfx.com/)** hosts community breakdowns that make excellent starting hypotheses
Document what you think it is doing *before* you build. The write-up of where you were wrong is the valuable part.
### Game Design
Take **one** level, encounter, or system.
- *Half-Life 2* — Ravenholm's pacing and player guidance
- *Dark Souls* — Undead Burg's loop structure and shortcut design
- *Doom* (2016) — a single arena encounter
- *Celeste* — one screen, or one chapter's mechanic introduction
- *Portal* — the test chamber teaching progression
- *Hollow Knight* — one area's layout and gating
- *Super Mario 3D World* — one level's mechanic introduction curve
Diagram its pacing, sightlines, beats, and teaching sequence. Then rebuild it as a playable greybox.
### Game Programming
Reimplement a named mechanic **from observed behavior**, not from source code or a tutorial of that exact system.
- *Celeste* — the dash, plus coyote time, input buffering, and corner correction
- *Titanfall* — wall-running and momentum
- *Zelda: Breath of the Wild* — the stamina and paraglider systems
- *Hades* — the ability and boon system
- *Hollow Knight* — combat feel: pogo, knockback, hit-stop
- *Dead Cells* — procedural room assembly
- *Super Mario Bros.* — the movement model, which is far subtler than it looks
State your hypothesis about the implementation, build it, then test the hypothesis.
### Virtual Production
- A published in-camera VFX shot — *The Mandalorian*, *House of the Dragon*, *1899*, *The Batman*
- Epic's **In-Camera VFX Production Test** sample project as a reference build
Rebuild the environment as a real-time set and match the shot with camera tracking. Scope to the equipment you can actually reach.
---
## Milestones
### Research — 10 points · due Week 2, Tuesday, September 1
A documented teardown. Not a summary — an analysis.
1. **Source identification** — the piece, its makers, where it was published, with a link
2. **Why this piece** — one paragraph on what you intend to learn from it, and how it connects to the gap you identified in your portfolio review
3. **Technical breakdown** — your reconstruction of how it was made: software, pipeline order, techniques you can identify. **Name what you are guessing at.** Guessing is fine; pretending to know is not.
4. **Aesthetic breakdown** — the decisions that make it read: timing, silhouette, color, composition, camera, pacing, game feel. Discipline-appropriate.
5. **Reference collection** — supporting breakdowns, designer interviews, making-of material. Art of the Title interviews, GDC talks, and studio blog posts are gold here. Cite everything.
6. **Technique list** — every distinct technique this rebuild will require, marked *have it* / *partial* / *do not have it*. **This list is your Project 2 menu.** Take it seriously.
7. **Faithful or restyled** — state which approach you are taking and why
8. **Production plan** — week by week through Week 6, with named deliverables
**Format:** PDF or Canvas page, 4–8 pages including images.
---
### First Draft — 10 points · due Week 4, Tuesday, September 15
The **whole thing** present at rough standard. Complete beats polished — every part exists roughly rather than one part existing beautifully.
| Track | First draft means |
|:--|:--|
| Motion / title design | Full animatic, every beat timed, all elements placed |
| 3D animation | Full blocking pass, keys and timing set, camera in place |
| Rigging | Skeleton and controls functional, first skinning pass, testable |
| Technical art | Effect working end to end, unoptimized |
| Game design | Full greybox, playable start to finish |
| Game programming | Mechanic functioning with placeholder assets |
| Character design / boards | Every frame or view present at rough standard |
| Virtual production | Real-time environment functional with tracking |
**Also submit:** a side-by-side against the source, and a written note on the three largest gaps you can already see.
---
### Final Render — 10 points · due Week 6, Tuesday, September 29
1. **The finished recreation**, rendered and exported to the submission standards
2. **Side-by-side comparison** with the source — split screen or paired render
3. **Process documentation** — 6–10 images or clips showing the progression
4. **Written postmortem**, 500–750 words:
- What matched and what did not
- Where the source turned out to be doing something you did not initially see
- What you can now do that you could not do in Week 2
- **Which technique from this project you are taking into Project 2, and why**
Presented in Week 6.
---
## Grading
| | 9–10 | 7–8 | 5–6 | 0–4 |
|:--|:--|:--|:--|:--|
| **Research** | Teardown identifies non-obvious technique; technique list is specific and honest; plan is realistic | Solid analysis of visible decisions; plan present | Describes the piece more than it analyzes it | Superficial or incomplete |
| **First draft** | Whole piece present at rough standard; reads clearly against source | Mostly complete; some sections thin | Partial; significant gaps | Fragmentary |
| **Final render** | Close to source (or coherently restyled); clean execution; postmortem shows real insight | Recognizable recreation with visible craft | Substantial deviation or unresolved technical problems | Incomplete or not delivered |
---
## Attribution
The source is credited clearly in every deliverable, and anywhere this piece appears afterward. When it goes in your portfolio it is labeled as a **study** or **recreation**, with the original creator named.
This is not a formality — studios have rescinded offers over unattributed recreations in portfolios. Labeled correctly, it demonstrates exactly the professional judgment they are screening for. A recruiter who sees *"Study — original title sequence by Imaginary Forces"* reads that as maturity, not as weakness.