We focus on packet video delivery, with an emphasis on the quality of service perceived by the end-user. A video signal passes through several subsystems, such as the source coder, the network and the decoder. Each of these can impair the information, either by data loss or by introducing delay. We describe how each of the subsystems can be tuned to optimize the quality of the delivered signal, for a given available bit rate in the network. The assessment of end-user quality is not trivial. We present recent research results, which rely on a model of the human visual system.