Mobil networkler için deneyim kalitesi tahminleme modeli

Rate adaptation is a general problem in streaming services where sending and receiving sides should compromise on a common rate, e.g. to minimize the re buffering failures, also called start-stop failures here. Currently, start-stop failure has the dominant effect on QoE (quality of experience) since the other types of streaming errors, such as blurring and color distortion of media frames, can be significantly avoided with the protection of the streaming channels against packet losses, e.g. using Forward Error Correction or TCP/IP protocol. In this work, the term Playing Rate Adaptation is introduced to further increase the rate adaptation capability of the streaming services by introducing a synthetic failure type, controlled fluctuations in playing rate, which in turns can reduce or eliminate the start-stop failures. Playing some video segments in negligible reduced rate is argued to provide better QoE than the start-stop failures that otherwise would be possible. As a case study, the streaming service is considered to be progressive downloading type over the 3GPP's MBMS (multimedia broadcast multicast service) as the underlying network and as worst case scenario the playing rate adaptation is applied to all frames of 2 minutes video. In order to prove the accuracy of the proposed model, objective and subjective study using DSIS (double stimulus impairment scale) and DSCS (double stimulus comparison scale) are provided. The results show that the playing rate adaptation even in worst case scenario increases the user satisfaction from the service.