Physical modeling by computer, especially for sound synthesis of musical instruments, is based on discrete time and space numerical simulation of the targeted physical systems. Digital waveguides (DWG) have been found to be a particularly attractive technique due to their efficiency and inherent capability of modeling systems with propagation delays. Other techniques used are, for example, finite difference schemes, wave digital filters (WDF), modal synthesis techniques, etc. The relation between these approaches has been investigated to some degree, but systematic studies on how to build mixed models, including subsystems of different types, remain to be carried out. The aim of the paper is to explore some of the relationships between particular modeling paradigms and to show how to construct flexible yet efficient models for sound synthesis.