You've just finished a 24-shot storyboard for a short film. You step back and look at it. Something feels off. The shots are technically fine. Framing is reasonable, continuity holds up, the story gets from A to B. But it feels flat. There's no rhythm to it. No moments that land.
We've been there. More times than we'd like to admit. And the problem almost always turns out to be the same thing: shots that exist because "the scene needed a shot here," not because the shot itself does something specific for the story.
Where this framework comes from
The framework we're breaking down here comes from the open-source Shanyin Director Master, a script-to-storyboard methodology created by award-winning AI filmmaker Shanyin. It organizes directing decisions into two practical tools: a dual-layer model for evaluating individual shots, and a six-dimensional framework for setting a project's overall visual tone.
We're not going to cover everything in the repository. It's 2,500+ lines of directing methodology, covering everything from dialogue scene handling to shot timing statistics. What we're pulling out here are the two ideas we've found most useful when actually sitting down to design storyboards.
The dual-layer model
The core idea is simple. Every shot in a storyboard should answer two questions at once:
What does this shot do for the story? This is the structural layer. It's about function. Does this shot establish a location? Introduce a character? Advance the plot? Foreshadow something? Create a breathing moment between tense scenes?
The structural functions include: conveying information, establishing characters, advancing plot, planting or paying off foreshadowing, presenting theme, creating symmetry or callbacks, and providing emotional transitions.
What effect does this shot create for the viewer? This is the execution layer. It's about craft. Does this shot make the audience uneasy? Direct their attention to a specific detail? Distort their sense of time? Use a visual metaphor to communicate something that dialogue can't?
The execution effects include: creating tension or pressure, controlling the viewer's gaze, building atmosphere, manipulating time perception, deploying visual metaphor, using rhythm shifts to signal narrative layer changes, and meeting or breaking audience expectations.
The test is whether you can describe a shot's purpose in one sentence that covers both layers. Here's what that looks like:
Good: "Back-to-back framing shows the couple's emotional distance while hiding their faces from the audience."
The structural layer (emotional distance between characters) and execution layer (hiding faces to create audience curiosity) are both present.
Bad: "Shows the relationship changing."
Too abstract. Which relationship? Changing how? What is the camera actually doing to communicate this?
This sounds obvious when you read it on a page. But go look at your last storyboard and try to articulate both layers for each shot. We did this with a 30-shot board we'd been happy with, and found that 8 shots only had a structural purpose ("we need to show them arriving") with no execution intent at all. Those were the flat spots.
Dual-layer in practice: Parasite's staircase
In Bong Joon-ho's Parasite, there's a recurring visual motif: stairs. The Kim family climbs up to the Park house. Later, they descend back down to their semi-basement. The staircase that connects the two worlds is shown repeatedly.
If you break down any single staircase shot through the dual-layer model:
Structural layer: The shot transitions the story between two locations (rich house above, poor neighborhood below). It moves the plot forward.
Execution layer: The vertical composition — always framed so the audience sees the height difference — creates a visceral sense of class hierarchy. The viewer feels the distance even if they're not consciously thinking about class.
Bong didn't just need "a shot of them walking home." He needed vertical movement that makes you feel the gravity of class. That's the difference between a shot that has structural purpose only and one that has both layers working.
Six dimensions of directing
The dual-layer model works for evaluating individual shots. But how do you set the overall visual direction for a project? This is where the six-dimensional framework comes in.
Instead of describing a project's style with a single genre label ("it's a thriller"), this framework breaks the directing approach into six independent dimensions. Most projects will touch 2-3 of these. Some touch 4. Almost none use all 6.
A. Emotional tone is the film's temperature. Is it the quiet realism of Kore-eda's Shoplifters? The neon-soaked melancholy of Wong Kar-wai? The surreal absurdism of Yorgos Lanthimos? This dimension determines your default color palette, lighting approach, and pacing.
B. Genre conventions are the rules your audience expects. A thriller needs information asymmetry (the audience knows more than the character). Horror needs controlled revelation of the threat. Comedy needs timing. The question isn't just "what genre is this?" but "which conventions do we follow, and which do we deliberately break?"
C. Physical action defines where the kinetic energy comes from. A martial arts film has a different vocabulary than a war film. A disaster film uses scale. A fight scene in a John Wick movie keeps the camera wide so you can see the choreography. A fight scene in a Bourne movie uses shaky close-ups for chaos. Same genre, opposite execution.
D. Thematic relationships determine how you visually represent the connections between characters. Romance uses proximity and shared frame. Family drama uses domestic spaces. Coming-of-age stories track how the frame changes around the protagonist as they grow. This dimension is about showing, not telling.
E. Narrative form is structure. A mockumentary looks different from a road movie, which looks different from a musical. The form itself imposes visual constraints and opportunities. A single-location film needs to extract maximum visual variety from limited space. A time-loop film needs visual markers that shift subtly each iteration.
F. Social perspective asks how the film relates to reality. Social realism (Ken Loach) uses documentary-adjacent techniques. Historical period pieces need visual accuracy. Political satire often uses absurd visual contrasts.
Take Parasite again. It operates on three dimensions: B (genre: thriller/dark comedy hybrid), D (thematic relationships: two families mirroring each other), and F (social perspective: class as architecture). The other three dimensions aren't absent, but they're not driving the visual decisions.
When you map a project across these six dimensions before designing any shots, you end up with a vocabulary for the visual choices you'll make. "The emotional tone is restrained realism, the genre leans thriller, the thematic focus is on parent-child distance." That sentence alone tells you more about how to shoot the film than "it's a family drama" ever could.
The 70/30 rule
One more idea from the framework that we've found especially practical. In any given scene, roughly 70% of your shots should be conventional. Standard coverage. Clear compositions. Shots that keep the story moving without calling attention to themselves.
The remaining 30% is where you put the creative work. Visual metaphors. Unexpected angles. Callbacks to earlier compositions. These are your memory-making shots, the ones the audience will remember after the film ends.
The mistake most storyboards make is one of two extremes. Either every shot is conventional (safe but forgettable) or the storyboarder tries to make every shot visually interesting (exhausting and incoherent). The 70/30 split gives you permission to be boring most of the time, which makes the creative moments actually land.
The specific creative techniques to deploy at those 30% moments:
- Visual metaphor: using composition or objects to represent abstract ideas (the Parasite stairs for class)
- Perspective flip: suddenly showing the scene from an unexpected viewpoint
- Structural callback: echoing an earlier composition to create meaning through repetition
- Sensory counterpoint: the visuals and audio tell different stories (cheerful music over a sad scene)
- Information withholding: the camera deliberately doesn't show something the audience wants to see
- Visual escalation: a slow zoom, a tightening frame, a color shift that builds toward a climax
You don't need all six in one project. Pick 2-3 that match your project's dimensions, and deploy them only at narrative peaks.
Auditing your own storyboard
If you want to apply this to something you're working on, here's the process we use:
Go through your storyboard shot by shot. For each one, write down a single sentence that covers both layers. "This wide shot establishes the apartment's emptiness while making the viewer feel the character's isolation through the negative space." If you can't write that sentence — if you can only describe the structural function ("it shows the apartment") — that shot needs rethinking.
Then zoom out and check your creative distribution. Count how many shots in each scene are doing something visually beyond standard coverage. If the answer is zero, the scene will feel flat. If the answer is "all of them," the scene will feel overwrought. You're aiming for roughly 3 out of every 10.
Finally, check your dimensions. Can you name 2-3 of the six dimensions that are driving your visual decisions? If you can't, your storyboard might not have a coherent visual identity. It might be a collection of individually fine shots that don't add up to a style.
Where we use this
Our storyboard agent evaluates narrative purpose for every shot it generates. When it designs a storyboard sequence, it tags each shot with both a structural function and an execution intent, then checks the distribution across the sequence. You can see the full list of references and directors that inform our approach on our References & Open Source page.
If you want to try designing a storyboard with this framework in mind, give it a shot. The agent won't explain the theory to you while you work — it just applies it. But now that you know what's happening under the hood, you might notice the difference.
The framework discussed in this article comes from the Shanyin Director Master, an open-source directing methodology by AI filmmaker Shanyin. All illustrations generated using Genkee AI.