Start a new topic

Our arms as an AR target

Hi everyone,

I would like to create a 3D model that can move with our arms. For example, a bracelet (with a name write on it) that follows the arm movement in wich it is without moving the name form the top of the arm. I think that wikitude can perform this but I'm very new in AR and I would like to know if someone would be nice to help me in this project ? From the project creation to the final "app". The objective is not to do this with a simple bracelet but learning with such simple object will help me for the real 3D model.

If this topic has no place here, please tell me where i can put it and i will replace it.

Bye, Alexandre.

1 Comment

Hello Alexandre,

If you are using Unity you will be able to hide the part of the watch that is not supposed to be shown (occlusion), so it fits as it should in the wrist. I will also post here some tips and tricks when you are working with 3D models that could help you.

Tips and Tricks for 3D models
The Wikitude3dEncoder converts 3D models into an internal representation, so-called wt3 files, which can then be rendered by the Wikitude SDK. The converter supports mesh-based 3D models which can be animated, textured and lighted.

File format: Autodesk FBX (up to version 2015). Thought the FBX format is not a well defined standard it is supported by most 3D modelling tools, like Autdesk Maya or 3D Studio Max, and thereby represents a common 3D format. Since different modelling tools use different representations of the 3D models we recommend to verify the models by the Autodesk "FBX Reviewer”.

Mesh types:
  • Triangle-based meshes: The 3D models usually consist on a hierarchy of scenes nodes where the leaf nodes contain parts of a 3D mesh.
  • Skin-based meshes: mesh vertices a placed relative to a list of joint nodes, so-called bones.

  • Key-frame-based transformations/rotations/scalings applied to mesh vertices or scene nodes.
  • Bone animations for skin-based models.

Textures: All textures are converted to the PNG file format during the import. The converter accepts textures of the usual picture file formats, like JPEG, GIF, TIFF, ...

  • Diffuse (Lambert) lighting by specifying a directional light source.
  • Specular (Phong) lighting by specifying specular light source (direction and specular exponent).

Several light sources can be used within a scene. However, only one light source (including diffuse+specular light) can be used for each mesh part, i.e., scene node. Thereby the light source which is closest to the scene node within the scene hierarchy (by traversing the tree upwards) is used. That is, a light source specified as a root node will be applied to all mesh parts in a scene. Other light sources will be applied to all mesh parts which are lower in the scene hierarchy.

Semi-transparent surfaces, like windows of a car, can visualised.

Model size: There is no limit on the size of the 3D models in terms of numbers of mesh vertices or textures. However, for rendering the 3D models fluently on mobile devices we recommend to keep the total file size below 5-10 MB and the number of mesh vertices below a few 100k. 

The Wikitude3dEncoder is steadily developed and extended. The following features will probably be available in the next releases:
  • Pivots (helper nodes) and animations on pivots
  • Several transparent layers
  • Bumb maps
  • Height maps
  • Multi-texturing
  • Animated textures
  • Shadows

For users with very advanced rendering requirements we recommend to use Unit plugin which allows to use 3D models designed and rendered within Unity to be used as augmentations in the Wikitude SDK.



Login or Signup to post a comment