Speech noise Reduction and Automatic Recognition Based on Active Detection Algorithm

Abstract

Abstract In practice, people need to carry out intelligent speech analysis on the sound collected or recorded in some specific scenes. These sounds are recordings of a period of time in some specific places. Due to the influence of the phonetic environment and the limitation of acquisition or recording means, the actual effective and available voice content is very small, and the voice quality is relatively poor. Therefore, it is necessary to enhance the speech of the acquired recordings, eliminate the noise and the interference of non-human voice, and improve the understanding degree of the target’s voice. Voice quality is improved, and the voice can be targeted content recognition and keywords detected, the audio files into a text file, by setting the key words found early warning, the valuable data information in mass speech quick preview, retrieval, analysis, according to the established goals and objectives by using automatic speech recognition in order to improve work efficiency.