Speech Recognition

Introduction

Speech recognition is the process of converting speech into text [Kimberlee A. Kemble,Program Manager, Voice Systems Middleware Education, IBM Corporation ] which enable computers (or other machines) to identify spoken words. The idea is to transform voice signal steams (utterances) into text words in a way that both voice signals and generated text would demonstrate the same meaning. It is an alternative to typing on a keyboard: which means it enables inputting data in to computers just by speaking to a microphone.

The contrary technology, i.e. transforming text to speech, “Voice synthesis” has been around for a long time since Bell Labs demonstrated a computer-based speech synthesis system running on an IBM704 in 1961.Today, after forty-six years from that time, voice synthesis technology can be found in diverse products such as talking dolls, car information systems and various text-to-speech conversion services. A satisfactory speech synthesis of respectable quality can nowadays be achieved using these systems.

In contrast, speech recognition researchers realized that the task of converting speech to text is much harder than they predicted when work began over forty years ago. However, nowadays Automatic Speech Recognition (ASR) systems have spread over various domains and applications at a considerable status. However, systems such as “Dictation” which just transforms speech to text words, “Telephony” such as Some PBX/Voice Mail systems which allow callers to speak commands instead of pressing buttons to send specific tones, Voice input telephone menu systems that are powered with VoiceXML, Command and Control systems, Systems that support human computer interaction for people with various disabilities, Wearables, Embedded Applications etc., are some examples that reflect the state of voice recognition today.

The idea of this topic is to analyze the current status of the area of speech recognition and to determine future things that are needed to be developed. Special attention is to be focused on the extent to which the development of the technology is visible related to Indic languages currently. And, last but not least, it is hoped to study the ultimate direction that the Indic languages recognition heads in future based on present context.