ASA S3.50 PDF download

ASA S3.50 – Method for Evaluation of the Intelligibility of Text-to-Speech Synthesis Systems

This Standard is to be used for testing the speech intelligibility of text-to-speech systems, providing ameasure of human listeners? recovery of words that correspond to the intended phonemic content ofspeech created by the system. Listeners are tasked to record the words or sentences they hear.Scoring may be either at the word or segment level. A normalized edit distance of the response fromthe intended message is the measure of the system?s speech intelligibility. This Standard specifiesmethods for selecting test material, which may depend on the purpose and constraints of the test. TheStandard also specifies methods for selecting and training the listeners; for designing, controlling, andreporting the test conditions; and for analyzing and reporting the test results. The Standard alsoprovides background material, important for designing the test. Informative software is provided toassist the user in creating stimuli and scoring the test results. Use of the software is not mandatory.