Our app needs to support multiple targets which have different sizes. We have one cube that is 8cm big, and one cube that is 50cm big.
We also have several visuals (models) that we want to apply to these targets. Which visual is actually assigned to which target is determined at run-time, we do this by instantiating the model beneath the respective target container object. The target container's scale is set to the target's scale (8cm or 50cm), as is the target size in Vuforia's dataset. This works pretty well regarding sizing the models correctly.
However, there is one challenge: When the scale is set, this also influences the lights and physics. This means: A model that has been optimized for the 8cm target looks weird on the 50cm target and vice versa.
Hence, we're looking for a best practice to deal with this. Choosing a normalized size for all visuals and targets (e.g. 1m) would not work according to Vuforia's documentation, because this influences the camera handling. But how could we adjust the lighting and physics setup then, so it works correctly on both sizes?
Any hints are appreciated. Thanks.