Quality of Experience (QoE) is the key criteria for evaluating the Media Services. Unlike objective Quality of Service (QoS) metrics, QoE is more accurate to reflect the user experience. The Future of Internet is definitely going to be Media oriented. Towards this, there is a profound need for an efficient measure of the Quality of Experience (QoE). QoE will become the prominent metric to consider when deploying Networked Media services. In this thesis, we provide several methods to estimate the QoE of different media services: Voice and Video over IP. We study the performance and the quality of several VoIP codecs like iLBC, Speex and Silk. Based on this study, we proposed two methods to estimate the QoE in real-time context, without any need of information of the original voice sequence. The first method is based on polynomial regression, and the second one is based on an hybrid methodology (objective and subjective) called Pseudo-Subjective Quality Assessment. PSQA is based on the artificial neural network mathematical model. As for the VoIP, we propose also a tool to estimate video quality encoded with MPEG-2 and with H.264/SVC. In addition, we study the impact of several network parameters on the quality, and the impact of some encoding parameters on the SVC video quality. We also made performance tests of several SVC encoders and proposed some SVC encoding recommendations.
