The aim of interactive computer graphics scenes and Virtual Environment (VE) technologies for simulation and training is the creation of accurate, high-quality imagery and interaction interfaces that faithfully represent a real-world task situation. Reliable fidelity evaluation techniques are essential in order to assess relevant implementations. A commonly employed strategy is to compare task performance in the VE in relation to the real world scene being represented. Spatial perception tasks are often incorporated in benchmarking processes as such, since spatial awareness is crucial for human performance efficiency. Such approaches, however, are not sufficient to assess the fidelity of VE systems. They are often limited to a specific application and are not based on formal frameworks but in most cases on arbitrary selected spatial perception tasks. This thesis introduces a metric, based on human judgements of spatial memory awareness states for assessing the simulation fidelity of a VE in relation to its real scene counterpart. This framework is based on the cognitive processes participants employ in order to retrieve the memory of a space. Participants could describe how they make their spatial recollections by selecting between four choices of awareness states. These depend on the level of visual mental imagery involved during retrieval, the familiarity of the recollection and also include guesses, even if informed. In order to demonstrate the differences between using task performance based metrics and human evaluation of cognitive awareness states, a set of VEs displayed mainly on Head Mounted Displays (HMDs) were created. Resulting scenes were then compared to the real task situation they represented by employing the spatial memory awareness states methodology as well as assessments of presence, simulator sickness and responses to lighting. The experimental results are presented in this thesis, with an emphasis on probability-based formal analysis, revealing a variation of the distribution of participants' awareness states across conditions, especially when task performance failed to reveal any. Simulation of task performance does not necessarily lead to simulation of the cognitive processes employed in order to complete the task at hand for VE display technologies. The general premise of this thesis is focusing on 'how' tasks are achieved, rather than only as in earlier VE simulation research, on 'what' is achieved.