Techniques for Comparing Objective and Subjective Speech Quality Tests
Stephen Voran · 1994
Objective (or instrumental) tests of speech quality have been proposed as ways to reduce the need for expensive and time-consuming subjective (or auditory) tests. Both types of tests attempt to quantify the range of opinions that listeners express in response to a group of speech transmission or storage devices, but objective test results often show measurable deviation from subjective test results. This deviation may be judged to be acceptable if the objective test offers significant savings of time and expense. This cost-performance judgment cannot be made without a meaningful statistical measure of the deviation that is likely to be associated with the objective test. This paper offers a number of techniques that compare both the central tendencies and the uncertainties of two tests. The resulting statistics have direct, intuitive interpretations.