In the last decade the use of the neural networks redounded to the improvement of speech recognition. With the steady growth of resources it has become possible that these solutions got more efficient than before. The improvement of speech recognition softwares has not stopped.
First the main application areas of automatic speech recognition are shown. In chapter 1, the process and history of speech-to-text transciption are summarized. In the 2. chapter I write about the applied methods. The creation of an acoustic model is explained starting from the audio processing. The specificities of the language model and the automatas are also shown. In the 3. chapter I talk about the used tools, and the work environment. In the 4. chapter I write about the two models I have created. I evaluate the results, and compare them to other systems made from the same data. In the last part the opportunities for further developement are described, and the results of the latest speech recognition systems are shown.