Description

Key Facts The measurement method referred to as the perceptual evaluation of speech quality (PESQ) was published in 2001 as Recommendation ITU-T P.862 by the International Telecommunication Union. It is used to measure the quality of speech signals that are transmitted at a low bit rate using high-compression psycho-acoustic coding methods. Psycho-acoustic coding methods use the characteristics of the human ear and remove signal sections that have no effect on auditory perception before transmission. To develop PESQ, a large of number of recorded voice samples from various speakers in various languages were used. These recordings were then compressed using a variety of speech coders – and thus different quality levels – and degraded by introducing typical network transmission interferences. A large number of test listeners rated these samples in a series of acoustic tests based on the standard scale for speech quality ranging from 1 (poor) to 5 (excellent). The development of PESQ provides a method that compares the original undegraded speech signal (reference signal) with the degraded signal (measurement signal) and then outputs an objective measured value that correlates very well with the average of the audio test results. The R&S®UPV offers this measurement in line with the method licensed by the Opticom GmbH company in Erlangen, Germany.
''