QoE Models for Adaptive Streaming: A Comprehensive Evaluation

Future Internet

Nam Pham Ngoc¹  Truong Cong Thang²  Duc Nguyen³ 

[1] College of Engineering and Computer Science, VinUniversity, Gia Lam District, Hanoi 100000, Vietnam;Department of Computer Science and Engineering, The University of Aizu, Aizuwakamatsu 965-8580, Japan;Department of Information and Communication Engineering, Tohoku Institute of Technology, Sendai 982-8577, Japan;

关键词: quality of experience;    quality model;    adaptive streaming;    multimedia services;   

DOI  :  10.3390/fi14050151

来源: DOAJ

【 摘 要 】

Adaptive streaming has become a key technology for various multimedia services, such as online learning, mobile streaming, Internet TV, etc. However, because of throughput fluctuations, video quality may be dramatically varying during a streaming session. In addition, stalling events may occur when segments do not reach the user device before their playback deadlines. It is well-known that quality variations and stalling events cause negative impacts on Quality of Experience (QoE). Therefore, a main challenge in adaptive streaming is how to evaluate the QoE of streaming sessions taking into account the influences of these factors. Thus far, many models have been proposed to tackle this issue. In addition, a lot of QoE databases have been publicly available. However, there have been no extensive evaluations of existing models using various databases. To fill this gap, in this study, we conduct an extensive evaluation of thirteen models on twelve databases with different characteristics of viewing devices, codecs, and session durations. Through experiment results, important findings are provided with regard to QoE prediction of streaming sessions. In addition, some suggestions on the effective employment of QoE models are presented. The findings and suggestions are expected to be useful for researchers and service providers to make QoE assessments and improvements of streaming solutions in adaptive streaming.

【 授权许可】

Unknown