Hoka Shoe by Deckers – 3D Design Workflows for the Footwear Industry

Reusing the existing CAD data

Reusing the existing CAD data

CAD data without materials

The vast majority of manufacturing companies have at least some of their products designed in one of the engineering 3D packages like Romans CAD, Modo or Rhino. This data can be further used for material iteration (testing colorways and other design-related uses) and for marketing purposes, including photorealistic visualization.

The challenge here is the fact that engineering data is based on NURBS, which usually needs to be converted into polygons before being used in other visualization software. The conversion creates polygonal topology which is hard to modify, including the creation of proper UV maps.

There are two ways to use this data:

• Use it as is and try to utilize the existing UV shells, if they are pre-made in the CAD software or;
• Use the CAD file as a reference to create a new topology.

In this case, we’ll go with the first option, since it is more efficient.

Creating UV maps

Creating UV maps

The shoe has been split into 6 groups, each with a dedicated UV map

The UV map is a flattened representation of 3D geometry and it serves to correctly map textures to the surface. It allows cross-platform usage of 3D geometry with its materials.

The first step is to organize 3D objects into logical groups (sole, exterior, interior, details, etc.). Each group will have a dedicated material assigned to it, so the object within that group needs to have their UV shells laid out in 0-1 UV quadrant.

If the CAD file included the unwrapped UV shells, those should be pre-scaled to have a correct relative ratio to the other objects from the same group.

If there are no UV shells, they should be unwrapped. Depending on the object itself and the tool used, this might be a time-consuming process, but it has to be done properly. This will ensure we can correctly apply the materials.

Creating Digital Materials

Creating Digital Materials

Materials extracted from a single color input

The best way of creating digital materials is by scanning the physical samples. However, this is a time-consuming and (usually) pricey, so it is not always an available option.

Instead, we’ve used Project Substance Alchemist. The process is almost completely automatic. The initial input Alchemist requires is a single color texture sample. From there, it is able to reproduce all other textures, including the normal and height maps. It also uses the AI algorithm for a smart light recognition, which removes the shadows baked into the color texture input.

Extracting all of these textures from the color input would traditionally take an experienced texture artist at least 20-30 minutes.

With Project Substance Alchemist, the process takes about 10x less time (a few minutes per material).

The HOKA shoe asset had a total of 6 Substance materials, so here is the time-saving estimate:

The Substance materials are exported as .sbsar, so they can be used in Substance Painter.

Applying the Materials to the Objects

Applying the Materials to the Objects

When the object is loaded into Substance Painter, the first task is to bake the utility maps, which will provide the software with various information about the nature of its meshes (such as the position in space, occluded parts, curvature lines, normal maps baked from the high polygon resolution mesh, and so on).

Now the object is ready for the material assignment. Each group will have different materials applied on its various parts, and sometimes it might be necessary to paint custom material masks (like in the case of the sole).

Furthermore, it may be required to project certain design elements, using pre-made, semi-transparent images.

Having multiple groups means the object has more resolution “real-estate” (up to 8196×8196 pixels, in this case).

Preview from Substance Painter with one of the UV maps

Textures can be exported from Substance Painter for multiple uses. We will export using two different presets: Adobe Dimension and gltF. The latter will provide us with an asset, which can be used in Project Aero.

Iterating on colorways, using Dimension and Illustrator

Iterating on colorways, using Dimension and Illustrator

Assets exported from Substance Painter using the Adobe Dimension preset can simply be drag-and-dropped into the Dimension viewport. The models and their materials should load up automatically.

We can access all the materials and their textures by selecting individual meshes.

Using UV maps as a reference, we can set up templates in Illustrator (each of the 6 groups requires a separate template since it has a unique UV map). This will allow us to quickly iterate on the colorways and potential designs and patterns. We can use as many artboards as desired and dynamically pick between them when we connect the .ai file in Dimension.

This is an example of the template with 4 different artboards:

If we replace the color texture on the appropriate material in Dimension with this .ai file, we will have the option to pick between the artboards:

The asset in Dimension with one of the materials selected

Dimension also allows you to edit the attached textures, by creating a live link with the appropriate editing app (Photoshop or, in this case, Illustrator). This allows further edits, which are immediately reflected back in Dimension, upon saving.

Marketing visualization using Dimension and Photoshop

Marketing visualization using Dimension and Photoshop

We can use Dimension to create photorealistic renderings, suitable for marketing purposes. Besides setting up the materials properly, it is very important to spend time working on the lights and to blend the 3D scene with a background photo.

The key feature we can use in order to do so is called Match Image. This function uses an artificial intelligence algorithm to perform a set of (selected) operations at the same time:

• Resize canvas to match the image size ratio or the image size
• Create Environment Image (makes an IBL out of the background image)
• Match Sunlight
• Match Camera Perspective

Further light adjustments might be required to bring the 3D scene even closer to the background image as well as adding certain finer details, like focus and ground plane reflections.

The scene is now ready for rendering. After this process has been completed, Dimension will provide us with a multi-layered PSD file, which allows further adjustments in Photoshop. The file will include Object and Material selection layers, and Depth Pass; in addition, every 3D element will be separated from the background, and placed on its own layer.

Dimension render before and after the final adjustments in Photoshop

Immersive experience using Project Aero

Immersive experience using Project Aero

We can use the exported .glb file in Aero, which is the augmented reality application specially designed for smart devices. Aero provides users with the ability to integrate 3D content with the real world.

Project Aero is still in closed beta at the moment.

https://video.tv.adobe.com/v/3419353