Objective Modeling of Speech Quality with a Psychoacoustically Validated Auditory Model

摘要: A new objective measure for the transmission quality of low-bit-rate speech coding algorithms is described and tested. quantitative psychoacoustical signal processing model employed to perceptually relevant deviations between transmitted, degraded corresponding reference objectively. The applied transform coded (distorted) original an internal representation which thought as information that accessible higher neural stages perception. From a comparison these representations can be derived, shows high correlation with subjective mean opinion score data various test databases if frequency-dependent weighting exhibits increasing weights center frequencies filter channels representation. inherent parameters auditory optimized from independent present study yielded also best performance in predicting compared alterations model. was investigated by comparing it alternative models regard dynamic compression PSQM/P.861 measure. proposed on average comparable PSQM