A system includes a server to generate a real-time stream of audio packets and a client device to decode and playback the audio content of the stream. The client device includes a network interface configured to receive a stream of audio packets via a network and a buffer configured to temporarily buffer a subset of audio packets of the stream. The client device further includes an audio decoder having an input to receive audio packets from the buffer and an output to provide corresponding segments of a decoded audio data stream. The client device also includes a stream monitoring module configured to provide an audio packet of the subset in the buffer which was previously decoded by the decoder to the input of the decoder again for a repeated decoding in place of a decoding of an audio packet that is lost or late.