How to Use the Canvas Workflow? NadouPro Complete Guide to Director-Level Control

Most AI video tools give you only one input box—write a prompt and wait for the result. NadouPro’s canvas workflow is different: before the model runs, you can control type, style, light, camera movement, color, and focal length. The prompt describes the scene content; the parameters define the visual logic. Together, they deliver director-level control.

The Practical Role of Light in Generation

Light is not a style preference; it is a causal variable. Choose the wrong lighting preset, and a golden sunset lavender field may turn into a cheap-looking image lit like a parking lot. The NadouPro camera parameter system provides multiple lighting presets, each telling the model different information about the physical world:

• Backlight: the light source is behind the subject, giving the facial outline a warm halo, suitable for romantic sunset scenes

• Overhead soft light: light falls from above, placing the eyes in the shadow of the brow bone, suitable for large spaces, churches, and institutional-feeling scenes

• Side light: a large area of soft light enters from the side, leaving half the face in light and half in shadow, suitable for dramatic scenes such as interrogations and confrontations

• Natural window light: only natural light from outside the window is accepted, with soft light falling off over distance, suitable for lived-in indoor scenes

• Practical light source: only visible light sources within the frame are accepted, such as candles or explosion firelight, with no hidden fill light added; suitable for atmospheric scenes such as candlelight and battlefields

• Silhouette: the subject stands against a bright background, the face is unreadable, and only the outline remains—the story is told through form

The core principle is this: each lighting preset tells the model physical spatial information about the scene. When it is matched correctly, the model knows where the light comes from and what the space feels like.

The Complete Parameter Panel of the Canvas Workflow

In the canvas video node, the following parameters can be set:

Parameter	Options
Type	General, Action, Epic, Drama, Comedy, Horror, Film Noir
Color scheme	Natural clean color, bleached warm tone, ultra-saturated neon, teal-orange epic, sodium-light decay, cold steel, bleach bypass, black and white
Camera movement style	Classic still frame, silent machine, one-shot take, epic scale, intimate observer, impossible camera movement, documentary snapshot, raw chaos, dreamlike flow
Light	Backlight, overhead soft light, side light, natural window light, practical light source, silhouette
Camera body	16mm raw film, 35mm fine-grain film, clean digital
Lens	Clinical sharpness, macro, anamorphic widescreen, warm glow, vintage haze
Focal length	8mm, 14mm, 35mm, 50mm, 75mm
Aperture	f/1.4 wide aperture, f/4 medium aperture, f/11 deep focus

None of these parameters is "style beautification." Each answers a decision the model needs to make. Type answers "what kind of film is this"; light answers "where does the light come from"; color answers "the emotional temperature"; camera movement answers "how the camera participates in the narrative"; focal length and aperture answer "the physical texture of the image."

Canvas Agent Entry: Conversational Parameter Adjustment

The Agent entry in the upper-right corner of the canvas supports conversational creation. You can tell Agent in natural language, "make it a warm dusk tone, pull the focal length to 50mm, and change the camera movement to a handheld feel," and the parameter panel will update in real time. Agent can see all character, scene, and prop assets on the canvas, and directly references @ tags when generating prompts. When splitting shots, it also automatically prefills the parameters for each shot.

One key point: after Agent generates a prompt, it places it in the input box for you to review and does not automatically trigger generation. The final click is always yours.

Asset Space: Unified Asset Management

In the canvas workflow, characters, scenes, and props are all managed as assets in the asset space. Created assets are automatically brought into every subsequent shot, so there is no need to upload them again each time. Assets across projects can also be called.

Three Practical Cases

War Scene: Oppression and Immediacy

A 15-second sea landing scene. Parameter settings: action type, bleach bypass color, raw chaos camera movement, practical light source, 35mm fine-grain film, clinical sharpness lens, 14mm wide angle, f/4 medium aperture.

Bleach bypass gives the image hard contrast and a silvery metallic feel—cool grays, muddy blacks, and hot explosion flashes as accents. Raw chaos camera movement pulls the camera down to ground zero, with handheld shaking matching explosion reactions. Practical light source means only the visible explosion firelight within the frame provides illumination, with no hidden fill light added. The 14mm wide angle covers the full horizontal width of the dense group of soldiers on the deck and the gray sea.

Rally Chase: Handheld Motion

A six-shot escalating rally sequence—drift, struggle, lift-off, landing. Parameter settings: action type, cold steel color, raw chaos camera movement, practical light source, 35mm fine-grain film, anamorphic widescreen lens, 14mm, f/4.

Cold steel color immerses the entire image in cold steel blue, with only the headlights cutting through the dusk. The anamorphic widescreen lens gives horizontal stretch and oval bokeh, while the 14mm focal length stretches the visual width of the forest road.

Dark Fantasy Battle: Epic Scale

A 15-second paladin duel against a giant beast. Parameter settings: epic type, teal-orange epic color, epic scale camera movement, backlight, 35mm fine-grain film, anamorphic widescreen lens, 35mm, f/4.

Teal-orange epic creates a contrast between cool creature bioluminescent shadows and warm morning light. Epic scale camera movement uses IMAX-level low-angle orbiting and crane-style composition, making the giant beast read as truly giant. Backlight places the low morning sun behind the subject and the giant beast, and during the final strike, the sky beam becomes the key light source.

Prompts Describe Events; Parameters Define the World

The core logic of the canvas workflow: the prompt describes what happens, while the parameters describe what kind of world it happens in. When both are in place, the generated video does not feel like it was made by AI. Writing only a prompt and ignoring parameters is equivalent to handing all visual decisions to the model to guess—and the result is usually an in-between state that "doesn’t look like anything."

FAQ

Q: What is the difference between the canvas workflow and ordinary video generation?

Ordinary video generation starts from a prompt, and the model interprets it on its own. The canvas workflow provides visual parameters such as type, light, camera movement, and lens before the model runs, so generation executes preset decisions rather than interpreting a text description.

Q: Can Agent generate videos directly?

No. Agent is responsible for adjusting parameters and writing prompts, but the final click to generate is always yours.

Q: What is the difference between focal length and lens?

Focal length controls the field of view and the sense of spatial compression; lens controls optical texture—how highlights spread, how edges render, and how depth of field reads.

Q: What focal length should be used for portraits?

50mm or 75mm. 50mm renders space naturally and is the most versatile; 75mm introduces compression, brings the background closer, and is suitable for close-ups.

Q: What aperture should be used for landscape scenes?

f/11 deep focus, keeping both foreground and background clear so the spatial relationship between subject and environment is immediately readable.

Q: How does Agent know what assets are on the canvas?

Agent can see all characters, scenes, and props in the asset space, and automatically references them with @ tags when generating prompts, without needing manual specification.

For more features, visit NadouPro: NadouPro

Contains AI-generated information that may be incorrect. Please verify independently.