Streamed MPEG video over IP networks, like any other information sent over "best effort" networks, is susceptible to loss and unpredictable delays. However, in MPEG video, loss and unacceptable delays are propagated to other frames due to Intra/Inter frame encoding nature of compression. Different adaptation techniques are used to enhance Quality of Service and provide "controlled" degradation on the streamed video.  In this research we consider a source driven adaptation technique through a detailed simulation. Then we propose two schemes for encapsulation of MPEG video inside RTP/UDP/IP packets consistent with the standard and examine how this affects the network level QoS of the video before and after adaptation. A comparison between the two proposed encapsulation methods will also be done from bandwidth efficiency point of view. Effect of different network parameters on QoS for the both suggested encapsulations will be investigated too.