The paper presents different methods of implementation for Ambisoncs recordings in a wave field synthesis system. Using the Ambisonics B-format, virtual microphone signals are extracted from the recording. The signals thus generated can be placed and reproduced at the corresponding positions in the wave field synthesis system. To further improve the performance of Ambisonics recordings in the system the possibility of sound source localisation using a sound field microphone is explored. With the presented methods it is possible to determine the position of individual recorded sound sources and to create a virtual, aligned microphone signal pointing at the source. Reproducing a single sound source gives good results in the system. When playing multiple sound sources, the problems of ”sweet-spot” based systems arise.