In this paper we discuss how theories of vision, touch, sound, and learning behavior can form a basis for the development of a testbed through which real-life task performance can be compared with task performance in a Virtual Reality (VR) environment. By taking a multisensory approach, it will be possible to simulate the functionalities of a real training setting in a VR environment. Approaching this problem from a theoretical perspective, we will view it from a new angle and discuss whether we can enhance and nuance feedback in the virtual experience through the use of ambient media like sound, scent, heat, and wind. Sound may hold a great potential here. For visual perception, it is not only crucial that we can see relationships but that we are also able to search for patterns that we recognize. If an object is taken outside its context, its meaning can easily shift. To see is to search for patterns, but vision is also dependent on our experience of other senses. We can imagine how a given surface might feel by looking at a representation of the object, and this is because of previous tactile experiences with similar objects. From a technical perspective, integrating sound in a virtual environment is a straightforward process. Research shows that the process of learning a series of physical actions can be enhanced when it occurs in parallel with verbal or written information. In the literature, this phenomenon is described in terms of enactment or subject-performed tasks. Based on theories regarding vision, touch, sound and learning behavior, we suggest the design of a testbed that can be used in a pilot study aimed at increasing knowledge on how VR and AR can support learning in an assembly or installation context in order to produce guidelines for such an environment.