ASR system evaluation of speech intelligibility
In this study, we used version 4.8 of iflynote software (Automatic Speech Recognition System, Hefei, Anhui, China) developed by the iFly company. iFly developed a speech recognition system using a feed-forward sequential memory network (FSMN), which is a deep convolutional neural network in which image recognition is introduced, and its recognition framework is developed with a core spectrum recognition technique created by the company.