For more than 50 years, simulation models for the prediction of acoustics in rooms have been implemented in software. Today, room acoustic simulations are established tools for architects, consultants and researchers. However, as most of the underlying models of these simulations are based on the simplifying concept of geometrical acoustics, the uncertainty and the validity of these simulations should always be considered when simulated results are evaluated. This thesis presents a comprehensive database intended for the validation of room acoustic simulation software. It includes eleven acoustic scenes, covering both simple environments and complex rooms. In addition to a detailed description and the definition of input data, acoustic measurements were carried out for all scenes. Furthermore, two studies are presented in which simulated and measured data for selected scenes of the developed database are compared. The first study, a round robin experiment with up to six participants, reveals various shortcomings of simulation models based on geometrical acoustics. The second study demonstrates how deviations of simulated results decrease when measured data is known beforehand and input data of the simulation is adjusted manually or systematically. The database along with the findings of the two presented studies provide a basis for the future revision of room acoustic simulation software.