This chapter introduces the motivations, requirements, and architectures for studying human perception in online real-time interactive multimedia applications (RIMAs). These operate under numerous application-level measures and requirements, including quantitative metrics such as signal quality and jitter, as well as qualitative measures like perceptible delays and interactivity. Additionally, there are application-specific requirements, including the limit of human reaction time and end-to-end data transmission speeds. We present models of two-party and multiparty interactions, using online audio conferencing and fast-paced action games as case studies to illustrate the numerous issues and considerations encountered in their design. We present a general client architecture supporting RIMAs and discuss its control and network components. Finally, we outline the chapters covered in this monograph.