Speech Technologies

Speech technology is a research area and is used in biometrics to identify individuals. To understand it totally, we need to look at how the process of speaker recognition and speaker verification is carried out. Feature Extraction from the speech is used to train models, which are further used for verification of the voice. In modelling and matching a number of models such as NLP, the Hidden Markov Model, Neural Networks and Deep learning are used. Text-dependent and Text-independent are two techniques of speaker verification. Speech parameters can be found by Linear Predictive Coding (LPC) Discrete Fourier Transforms and Inverse Discrete Fourier Transforms. Mel Frequency Cepstral Coefficients (MFCC) are used for calculations. In addition, we aim to see how key concepts of text-based comparisons and interactive voice response systems are incorporated. This field also involves how the speech is synthesized and analyzed. Speech technology is used in diverse applications such as forensics, customer care, health care, household jobs, GPS navigational systems, AI chatbots, and law courts.