Mixed Reality aims at combining virtual reality with the user's surrounding real environment in a way that they form one, coherent reality. A coherent visual quality is of utmost importance, expressed in measures of e.g. resolution, framerate, and latency for both the real and the virtual domains. For years, researchers have focused on maximizing the quality of the virtual visualization mimicking the real world to get closer to visual coherence. This however, makes Mixed Reality systems overly complex and requires high computational power. In this paper, we propose a different approach by decreasing the realism of one or both visual realms, real and virtual, to achieve visual coherence. Our system coarsely voxelizes the real and virtual environments, objects, and people to provide a believable, coherent mixed voxel reality. In this paper we present the general idea, the current implementation and demonstrate the effectiveness of our approach by technical and empirical evaluations. Our mixed voxel reality system serves as a platform for low-cost presence research and studies on human perception and cognition, a host of diagnostic and therapeutic applications, and for a variety of Mixed Reality applications where users' embodiment is important. Our findings challenge some commonplace assumptions on more is better approaches in mixed reality research and practice-sometimes less can be more.