A multimedia system includes a source device for providing a media stream and a sink device for playing the media stream. The source device encapsulates the media stream into data packets with corresponding timestamps associated with a first wall time, and transmits the data packets to the sink device based on the timestamps and the first wall time. The sink device provides a second wall time based on the second global clock, synchronizes the second wall time with the first wall time through a network protocol, generates a local media clock, and locks the local media clock to the second global clock. The sink device decapsulates the data packets, and then plays the media stream with the locked local media clock. A play-out time of each data packet is determined by a corresponding timestamp and the second wall time.