Speech recognition in real-time

The paper describes the basic algorithms for speech recognition. It was explained that these are features of the speech signal, and as they occur. Speech recognition systems use algorithms selection of language units. The algorithm with dynamic bending of time is used in which the voice signal is compared with other voice signal based on the characteristics of the signal. The artificial neural networks and hidden Markov models are used, which represent one of the most common algorithms for speech recognition. Implementation is made to develop systems Altium Designer NanoBoard 3000 with linear predictive analysis which are calculated LPC coefficients for voice recognition of the Croatian language. The testing of the designed systems is made for voice recognition of the Croatian language.