Table of Contents
A vanity needs a room view and a close-up of the sink. An armchair is available in several fabrics and frame options. Additional image sizes are required for the online store. Different tasks, one common goal: Each product should be clearly recognizable, and the images should form a cohesive series.
For a small number of designs, the work steps can be managed individually. However, as the number of materials, frames, views, and formats increases, so does the number of assignments. This is where CGI automation can help. Its value lies in the ability to consistently execute coordinated decisions. Which decisions these are must be clarified in advance.
What Actually Changes in a Variant
A list of materials is not yet a design concept. A bouclé upholstery fabric derives its character from its texture and the shadows between the loops. On smooth leather, sheen and wrinkles stand out differently. With wood, the direction and scale of the grain contribute to the piece’s overall effect. A color value alone does not describe these characteristics.
To make products comparable across a series, the camera position and framing often need to remain consistent. That does not mean every surface looks convincing under the same settings. A dark material can lose detail, while a light one may look too flat in the same area. Selected variants are therefore used to test how far the shared lighting concept can be applied.
Product logic is just as important. Not every cover is available with every frame. These rules are based on the brand’s product line. If they are taken into account too late, images of combinations that do not actually exist will be generated.
The same sofa, two covers. Comparisons like this show which material properties change and which visual choices should be retained for a series.
What tasks a script can perform
A script executes predefined instructions. In a CGI production, for example, it can set material assignments, call up preconfigured cameras, or start render jobs. After rendering, additional scripts can name the images, export them in agreed-upon sizes, and assign them to the appropriate variants.
From wireframe to material variant: for this demonstration, we rendered Studio Seitz’s Helm armchair in Blender using the existing 3D model and its textures.
One model, three coordinated material combinations: in this sequence, the script controls material assignments, camera movement and image rendering. Shape and lighting remain consistent. Oak with Calla, walnut with Macadamia and ash with Asphalt can be compared on the same piece of furniture. The materials and their mapping onto the model were first set up and checked using test renders.
One workflow. Two output sizes.
Helm Armchair by Studio Seitz: Three available material combinations, each shown from two angles, are presented in two sizes. The video shows a condensed version of an actual test run; it is not a real-time measurement.
In a separate export test, we also exported six existing renderings as twelve WebP files: each at 1,000 and 2,400 pixels. The material combination, view, and output size determine the file name; the image dimensions are verified. This allows for the repeatable processing of a series.
These tasks should be kept separate. Rendering calculates the image from the scene. Image processing handles existing images. Automating the export process does not automatically mean that the entire production process has been automated.
For technical classification, the documentation for 3ds Max Batch, for example, describes how scripts can process scenes and execute rendering tasks. Which steps are appropriate for a specific project depends on the project's data and requirements.
Cutouts, Interior Scenes, and Configurators
Product cutouts: comparing variants
In an online store, an entire series of variants is often displayed side by side. If the scale or viewing angle changes unintentionally, it becomes more difficult to compare them. For a CGI series, views and material assignments can therefore be defined in advance. A script can control the approved combinations and assign the output to the respective product. The image sequence shown here for the Archipen armchair by NOOM, designed by Kateryna Sokolova, illustrates this comparability.
Bathroom Furniture: Room View and Product Details as a Series
For a vanity, the online store might need a front view, the product page might need a shot of the basin and faucet, and the marketing campaign might need a view of the vanity in a room. The following series of images illustrates these different purposes using the same subject. For recurring series, preconfigured camera settings, output sizes, and product IDs can be linked together. The crop areas are defined in advance: A script calls them up but does not decide which detail is important for the product.
Interior: Compare Materials in the Same Room
Even a staged scene can serve as the basis for multiple versions. In the following sequence of images, the sofa, perspective, and room remain the same, while the fabric changes. This demonstrates why a change in material is more than just a new shade: The texture of the light-colored upholstery has a different effect than the calmer surface of the darker versions. For a script-based production, these materials must first be set up and tested in the scene’s lighting.
Configurator: Linking Variants and Views
In our 3D configurator demo, you can select material options and view a product from multiple angles. The display combines pre-rendered images. When you change the material, the selected rotation position is retained. To achieve this, existing views are processed using scripts to meet the required resolutions and assigned to the correct variant.
Image output: one source for multiple channels
An approved image file can be used in multiple sizes: for the product overview, the detail page, or another digital application. Repeated exports and filenames can be generated according to fixed specifications. Choosing a different image crop, on the other hand, is also a design decision. Whether an image fits in portrait orientation should be determined before automated output.
One image, three output formats: An example of rule-based image processing. The product remains fully visible. This step processes an existing rendering; it does not calculate a new 3D view.
How a series of images is evaluated
A file can be technically correct and still show the wrong image. That is why checking does not stop at resolution and filename. Is the upholstery assigned to the right variant? Are seams and edges still visible? Is the wood grain correct? Are framing and shadows consistent across the series?
At Danthree, this process begins with an understanding of furniture and materials. Carpentry, interior design, and creative direction determine what is emphasized in the image. The technical execution brings these decisions together; it does not make the decisions itself.
If a collection changes, the process also needs to be reviewed. A new fabric may fit into the existing material logic. A modified upholstery shape or construction, on the other hand, may require a redesign of the model. Simply treating both as additional variants would be short-sighted.
What You Need to Get Started
For an initial assessment, it is helpful to have an overview of the products and available variants, existing 3D or CAD data, material references, and the planned uses of the images. In addition, there are questions regarding maintenance: How often do products change? Who approves new combinations? What file names and formats will be needed later?
Even during testing, each file should be uniquely assigned to a product, an approved variant, and a view. If changes are made later, it will then be possible to determine which images are needed. Such an assignment does not yet constitute an integration with a store or a product data system; interfaces and ongoing maintenance must be agreed upon separately.
A limited test using representative variants is more informative than a blanket quantity specification. It shows which steps are repeatable and where exceptions remain. Only from this can a reasonable production scope be determined.
The article “From Product to Visual World” explains our creative process. If you’d like to streamline recurring tasks in your visual production, check out the solutions on our CGI automation page.
