Caliope-SpeechProcessingLab/SpeechTester

Speech Tester is a set of Python scripts conceived as an extension to HTK Automatic Speech Recognition system (Young et al., 2002). Speech Tester aims to help language researchers and engineers to measure the intelligibility of transformed audio signals. Here the termed “transform” may describe any process that alters a speech signal or its reception by a potential listener. For instance, any type of audio compression, noise background and/or reverberation, vocoding (to simulate cochlear implants), partial auditory loss, or any combination of these and other imaginable transformations. Intelligibility refers to the capacity of a listener to recognize a transformed signal. In the rest of this introduction we describe the motivation to develop Speech Tester to measure speech intelligibility.