Speech Quality Assessment of VoIP: G.711 VS G.722 Based on Interview Tests with Thai Users

Abstract

Abstract-This paper presents the comparison between two codecs, G.711 and G.722 at 64 kbps, referring to speech quality perception using a subjective method called interview tests. These subjective tests have been conducted with 201 subjects, who are Thai native speakers that use Thai which is a tonal language, for accuracy and reliability of results. The results from testing with both codecs are almost the same; the scores are 4.17 for G.722 and 4.14 for G.711. After analyzing the results, it has been confirmed that G.722 does not provide better speech quality than G.711 to the Thai subjects significantly, which is consistent with previous information. However, these results could be used as the benchmark of G.711 and G.722 for speech quality assessment within Thai environments.Index Terms-VoIP, speech quality assessment, speech quality evaluation, speech quality measurement, subjective methods, subjective methods tests, G.711, G.722, Thai

On Some Biases Encountered in Modern Audio Quality Listening Tests (Part 2): Selected Graphical Examples and Discussion

“…They explained the null results by stating that for Thai speakers, who took part in the experiment, the spectral content above the conventional narrowband was of no benefit (most important speech spectral content bound within the traditional narrow telephone bandwidth) [47]. This explanation can be challenged.…”

“…In their experiment Daengsi et al tested five hypotheses and despite a large number of assessors none of them proved to be statistically significant [47]. They explained the null results by stating that for Thai speakers, who took part in the experiment, the spectral content above the conventional narrowband was of no benefit (most important speech spectral content bound within the traditional narrow telephone bandwidth) [47].…”

“…11 presents the results that could be explained using the contraction bias model. The MOS values plotted at the top of the figure, labeled as Interview Monadic Test, were extracted from Table 2 of the paper by Daengsi et al [47]. The error bars were calculated by this author based on the standard deviation and the number of votes provided in the aforementioned table, using the standard equation for 95% confidence intervals.…”