A next step towards measuring perceived quality of speech through physiology

For service providers, it is important to know the perceived quality of a transmitted speech signal, as this can be an indicator for choosing one operator over the other. To evaluate the perceived quality, several methods have been introduced in the area of Quality of Experience (QoE), such as subjective, instrumental and physiological methods. In the current study, we expose the test participants towards several speech files, while brain activity was recorded with electroencephalography (EEG). Additionally, a subjective quality judgment is obtained after each presentation which is summarized as the mean opinion score (MOS). To validate the selected stimulus corpus, a quality predication algorithm is used which calculates quality scores based on the speech signal. The recorded EEG data were set in relation to the subjective data and show promising results. In contrast to most previous studies in the domain of QoE using EEG, the study at hand uses a standardized test paradigm proposed by the International Telecommunication Unit (ITU).