Speech recognition in noisy environment, issues and challenges: A review

The voice is most prominent & primary mode of communication among the human beings. With this speech human can communicate with machine, thus this technique is used in education, military and medical sectors. Though this is not the new area, from last few decades researchers are working on the improvement of accuracy in voice recognition system. The design of that system concerns major issues such as speech classes, speech styles, vocabulary, transducers, illness and channels; due to all this constraints the noise factor in automatic speech recognition is high. Many researchers have tried to overcome above issues. The recent work in speech recognition is briefly summarized in this paper. Some basic fundamentals of clean and noisy databases of voice samples are also discussed.