A queueing system model for the design of feedback laws in rate-controlled MPEG video encoders

The MPEG standard is the most popular technique in video and audio encoding. However, when MPEG encoding is used to transmit video information, one intrinsic drawback is its high burstiness. A large number of proposed solutions to this problem implement encoding control schemes based mainly on on-line adjusting of the encoder quantizer level through the quantization parameter. However, the choice of the feedback law is still a challenge to make MPEG encoding suitable for transmission over the forthcoming guaranteed service scenarios, where a negotiated bandwidth profile is available for video transmission. Up to now, this problem has only been approached by using simulation; no analytical framework exists yet which allows designers to study the most effective feedback law for transmitting video in the network environment to be used. This paper introduces an analytical paradigm modeling a rate-controlled MPEG encoding system in which a feedback makes the quantizer scale parameter dependent on both the state of the output counter/buffer in the controller and the activity level of the frame to be encoded. To this end, an SBBP/D/N/K queuing system is used, where the SBBP (switched batch Bernoulli process) arrival process is varied according to the quantizer scale parameter chosen by the addressed feedback law at the frame level. The analytical framework allows us to evaluate system performance in terms of both the distortion introduced by the encoding system and the resulting output process statistics; it is general enough to be applied whatever the feedback law is, provided that the parameter to be varied is the quantizer scale parameter. A case study is introduced to assess the proposed framework