TESPAR coded speech quality evaluation (TCSQE)

Speech transmission quality measurement in cellular networks is a major indicator of performance for end-to-end quality of service standards. Many approaches have been proposed in the previous studies, but the results correlation with subjective experiments still need further optimization, especially for quality determination using languages with unique phonetic structures i.e. clicking sounds. Moreover, the evaluation test data is always a key element in order to obtain representative and consistent results. In this paper, we introduce TESPAR coding technique for the design of a new algorithm for speech (voice) quality measurement in cellular/wireless telecommunication networks. Our experiments show that the results from this algorithm correlates well with subjective experiments using a variety of speech samples. The proposed algorithm is also computationally efficient than the existing methods and is suitable for quality measurement of longer speech signals than the current 8 seconds speech test data mostly in use today.