Toward a comprehensive subjective evaluation of VoIP users’ quality of experience (QoE): a case study on Persian language

• Published in: Multimedia tools and applications Vol. 80; no. 21-23; pp. 31783 - 31802
• Main Authors: Hesam Mohseni, A., Jahangir, A. H., Hosseini, S. M.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01-09-2021; Springer Nature B.V
• Subjects: Artificial neural networks; Case studies; Communication; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Evaluation; Gender; Methods; Multimedia communications; Multimedia Information Systems; Neural networks; Noise; Persian language; Quality of service; Special Purpose and Application-Based Systems; Support vector machines; User experience; User satisfaction; Voice over IP (VoIP); Support vector regression (SVR); Artificial neural network (ANN); Quality of service (QoS); Subjective test; Quality of experience (QoE)
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-021-11190-7

Quality of Experience (QoE) measures the overall quality of a service from users’ point of view by considering several system, human, and contextual factors. There exist various objective and subjective methods for QoE prediction. Although the subjective approach is more expensive and challenging than the objective approach, QoE’s level can be more accurately determined by a subjective test. This paper investigates various features affecting QoE by proposing a comprehensive subjective evaluation. First, we show that many unconsidered factors can significantly affect QoE. We have generated voice samples featuring different values for novel factors related to the speaker, signal, and network. Regarding the speaker, we take into account the accent and gender of Persian-speaking people. We conduct an extensive survey by employing a large number of users. Our comprehensive analysis reveals that the users’ identity has a significant influence on QoE. Our experiments show that many previously studied parameters do not affect QoE in the same way for various users with different genders and accents. Finally, we show that QoE can be accurately predicted using Artificial Neural Network (ANN) and Support Vector Regression (SVR) techniques if the new identity features are taken into account.