Explore

3D Scanning: Key Notes & Considerations

Puppet Size and Pose Considerations

A Pose and T Pose depends on the needs

The aesthetic workflow of the sculpting artists should be prioritized. While a T-pose is generally more intuitive for scanning and best suited for rigging, an A-pose works just as well. If the A-pose feels more natural or convenient for the sculptor, feel free to use it.

Armpit textures can be projected or captured separately if needed, and we can convert an A-pose into a T-pose later in the software.

2. Size of the puppets.

It’s hard to define an exact ideal size for scanning, as it depends heavily on the design and surface texture. As a general guideline, appendages or parts thinner than 3mm are difficult to capture accurately.

Regarding to the question. dolls around 70 cm tall (like Gretel) are perfectly fine. Larger dolls often help in capturing finer details. What matters most is the surface texture: smooth, uniform surfaces (like on the right side of the example image) have less visual information and are harder to scan, especially when small. In contrast, more textured or detailed surfaces (like on the left) are easier to capture, even at smaller sizes. High-contrast in colors also help improve scan quality.

⁠

Some parts can be difficult to scan, so any additional information you can provide would be very helpful.

Hair Scanning

Please let us know how the character’s hair will be designed. Realistic hair (as shown on the left) cannot be scanned accurately, but it can be recreated later in 3D software. For the best scanning results, the hair should be pulled tightly and neatly back during the scan.

Thicker, clay-like hair is easier to scan and can also be refined or recreated afterward.

⁠

Clothing and Appendages

Thin details like laces or fine patterns can be tricky to scan and are often best recreated in 3D software. If your character includes these, just let us know so we can plan and share what we might need to ensure the best possible result.

⁠

Lighting in 3D environment

How does lighting work in a 3D environment?

In Blender, lighting behaves much like it does in the real world. It influences how objects appear, how shadows are cast, and how materials respond to light. To achieve the desired look, the lighting setup must consider not just the light sources themselves, but also how objects are shaded and how they interact with the light.

Key factors that affect this interaction include:

Surface materials: Properties like roughness, metallic, and specular determine how light reflects or scatters across a surface.

Object normals and geometry: The shape and direction of surfaces affect how light hits and bounces off them.

Render engine (Eevee vs. Cycles): Cycles offers more physically accurate lighting, while Eevee is faster but may require more tweaking for realism.In Blender, lighting behaves much like it does in the real world. It influences how objects appear, how shadows are cast, and how materials respond to light. To achieve the desired look, the lighting setup must consider not just the light sources themselves, but also how objects are shaded and how they interact with the light.

Light sources and reflections: see the table below

Lighting Options

Lighting Options

Light Type

Usage

Common Info

Point Light

Emits light equally in all directions from a single point (like a bulb)

Use with falloff settings (Inverse Square is realistic)

Sun Light

Simulates distant sunlight; consistent direction and intensity

Rotate to set direction; size affects shadow softness

Spot Light

Emits a cone of light, like a flashlight or stage spotlight, useful for dramatic lighting.

Control angle, blend, and falloff to shape the beam. Customized Gobo Light texture possible.

Area Light

Emulates light from a flat surface, like a softbox, great for indoor or studio lighting. But more expensive for render.

Size and shape affect softness and spread

Ambient Light

Not a separate light, but controlled via World settings

Use HDRI environments for realism; adjust strength in World tab

Mesh Light

Emissive materials on mesh objects

Set Emission shader in Material; enable “Multiple Importance Sampling”

HDRI (Environment)

Image-based lighting from spherical images. super realistic reflections and lighting, very easy global illumination.

Load in World Shader using Environment Texture node

Volume Light

Light interacting with particles or fog. great for god rays but require high render time.

Use with World Volume or cube with Principled Volume shader

There are no rows in this table

⁠

During the scanning process, how is lighting handled?

Base on the scanning techniques, Lighting requirements varies:

Photogrammetry Scans:

Lighting: Use soft, diffuse lighting to minimize harsh shadows and highlights. Shadows can confuse the photogrammetry software and lead to errors in the mesh reconstruction. Avoid direct light sources or strong directional lighting.

Surface Texture: Surfaces with high texture and contrast (e.g., rough, detailed, patterned) work much better for photogrammetry.

Smooth or low-texture surfaces are challenging to capture accurately because they lack visual features for the software to track between images.

Gaussian Splatting (GS) to Mesh

Compared to photogrammetry, Gaussian Splatting is a much newer technique that offers several advantages. It provides better handling of lighting interactions, enables highly detailed captures, and can even perform well on low-texture surfaces, which are typically difficult for traditional photogrammetry.

Lighting setup requirements are relatively simple, and light-object interactions can be captured effectively: even on transparent or reflective surfaces.

Though Gaussian Splatting preserve full Spherical Harmonics lighting info in Point cloud data, while converting GS data to Mesh might compress Spherical Harmonics lighting info, causing quality lost.

More Info on Gaussian Spatting Technique

Overview of Gaussian Splatting (GS)

Overview of Gaussian Splatting (GS)

Physical build Vs. 3D creation

Which types of sets or elements are more relevant to build physically? Which ones would benefit more from being created directly in 3D?

In general, large-scale environments and effects benefit the most from being built in 3D. However, objects that require physical interaction can be more challenging and time-consuming to create and manage in 3D.

Physical vs 3D Sets Comparison

Physical vs 3D Sets Comparison

Aspect

Physical Sets

3D Sets (Digital)

Realism

Very high (natural lighting, real materials, tactile)

Can be highly realistic with time, skill, and render power

Actor Interaction

Direct and seamless

Requires green screen or tracking;

interaction is harder

⁠

Setup & Build Time

Time-consuming; depends on materials and construction

Faster for concept work; longer if high detail is needed

Cost

Expensive for large or complex builds

Expensive for high-end renders or simulation-heavy scenes

Flexibility & Revisions

Limited once built

High — elements can be tweaked, duplicated, re-lit easily

Lighting Control

Limited to physical lights and setups

Fully controllable and adjustable

Environment Scale

Limited by space and budget

Unlimited — ideal for vast or imaginary worlds

Camera Movement Freedom

Physical constraints apply

Full freedom (fly-throughs, impossible angles)

Post-Production Integration

Natural, especially with actors

Requires compositing, matchmoving, and cleanup

Best For

Close-ups, actor interaction, tangible realism

Backgrounds, large-scale environments, VFX-heavy scenes

There are no rows in this table

⁠

3D Sets Difficulties

3D Sets Difficulties

Name

Column 2

Notes

Interaction

Clipping Artifacts

When a character holds or touches an object, parts may intersect or pass through each other if not carefully animated—like fingers clipping through a cup. This demands detailed adjustments to hand and object positioning.

Loss of Real-World Detail

Soft or elastic objects (like pillows or fabric) are harder to replicate accurately. Subtle deformations from touch or weight are often lost, making the interaction feel less convincing compared to real-world physics.

There are no rows in this table

⁠

Freedom of manipulation

- How much flexibility does 3D give us in terms of modifying objects or environments?

- Can we easily change the color of an object, a wall or a set element?

- Are there technical limitations we should be aware of?

Modifications to objects or environments should remain as flexible as possible throughout the project. In general, changes are much easier to implement when planned early in the workflow. As the project progresses, technical and creative constraints tend to build up, making later adjustments more complex and time-consuming.

Changes to elements like colors or design features are typically manageable, but more specific information is needed to provide clearer guidance and direction.

Scanning constraints

1. Is there a minimum or maximum size for an object to be scanned effectively?

2. Are there any limits regarding the types of surfaces, shapes or patterns that can be scanned? For example: transparent, reflective, very dark surfaces, complex geometries or repetitive patterns.

3. In general, what are the main technical limitations we should keep in mind when preparing the set and props?

1. Static Objects / Environments: There are fewer constraints on maximum size for static objects or environments. Larger scenes can be scanned or captured using a 360-degree camera or a wide-angle lens, which should provide sufficient data for generating the scene within a 3D environment.

2. Surface Types & Technique Considerations: The effectiveness of scanning techniques depends on the surface types and the project requirements. A thorough round of testing is recommended to tailor the workflow for optimal results.

3. General Notes: Transparent and reflective surfaces should be avoided when using photogrammetry, as well as very dark or low-texture materials. Complex geometries are typically manageable, but tests are necessary to help identify and plan around any limitations.

4. Technical Limitations: The primary technical bottleneck is converting scans into animation-ready, rig-ready meshes, which is often the most time-consuming part of the process. Previsualization and testing are essential to streamline the workflow and avoid delays.

5. Additional Note: Using a motion capture suit for animation should not present any issues, provided that the character is successfully rigged and skinned.

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.