Voice over Internet Protocol (VoIP) systems have been spreading massively during the recent years. However, many challenges are still facing this technology among which is the lossy behavior and the uncontrolled network impairments of the Internet. In this chapter, the authors design and implement a VoIP test-bed utilizing the Adobe Real-Time Media Flow Protocol (RTMFP) that can be used for many voice interactive applications. The test-bed was used to study the effect of changing some voice parameters, mainly the encoding rate and the number of frames per packet as function of the network packet loss. Several experiments were conducted on several voice files over different packet losses, concluding in the best combination of parameters in low, moderate, and high packet loss conditions to improve the performance of voice packets measured by the Perceptual Evaluation of Speech Quality (PESQ) values.