Automatic Speech recognition is the translation of spoken words into text. It takes speech data as input and divides it into small time domain frames. Speech signal processing considering speech signals stationary for a small time interval. From point of view speech signals are divided into small units Morphims or Phonims. Any speech data can be sorted as word uttered followed by voice and silence intervals. Voice activity detection can be are employed to detect voiced and unvoiced part of speech. Speech processing consists of speech recognition, speech synthesis, speaker recognition, understanding of speech with reference to context, speech coding, speech enhancement, speech transmission, speech to text conversion & text to speech conversion etc. In general speech to text conversion system will convert input speech data to output text data. If the input speech data is inappropriate with some errors then there is a possibility to get incorrect output data. The proposed system contains options for correction of inappropriate input data so that the output text and speech data produce and pronounce is correct. The proposed system will be employed as learning assistance in educational field for students to learn correct pronunciation of words. The proposed system will also help tourists for conversation in local language.
