المجلة الدولية للعلوم والتقنية

Voice recognition has become a primary approach as part of many advanced technologies and systems in our modern age as it has been adopted by different technology manufacturers for different purposes. Modern technologies, which use speech recognition systems, need more accurate algorithms to perform real speech recognition functions. In this paper, an effective speech recognition system has been proposed to achieve better results compared to the classical methods. Advanced signal processing technologies are adopted to analyze the recorded speech, which measures the parameters required before performing a voiceprint recognition task. The speech was recorded using MATLAB software by setting a 32-bit two-channel recorder and then this speech is used as a sample to complete this study. The system includes a noise-removable section to remove syllables without sound so that it takes less time to perform speech recognition. Autocorrelation was calculated to match the recorded signal with the test signal, and then the pitch period was determined to ensure the identity of the speaker. The results are observed by performing speech recognition to see if the system can identify the contents of the speech and identify the speaker, so access is allowed if the same speaker with the same content enters his or her voice. For this purpose, MATLAB software was used to build the system and the results were presented throughout this paper.