Review of Noise Reduction Techniques in Speech Processing

Abstract

Present systems advances in speech processing systems aim at providing sturdy and reliable interfaces for sensible preparation. Achieving sturdy performance of those systems in adverse and screeching environments is one in every of the most important challenges in applications like dictation, voice-controlled devices, human-computer dialog systems and navigation systems. Performance of speech recognition systems powerfully degrades within the presence of background, just like the driving noise within a automobile. In distinction to existing works, we have a tendency to reduce the boost in noise strength that present in levels of speech recognition: feature extraction, feature improvement, speech modelling, and coaching. Thereby, we offer a summary of noise modelling ideas, speech improvement techniques, coaching ways, and model design, that square measure enforced in speech orthography recognition task considering noises created by numerous conditions.
DOI: 10.17762/ijritcc2321-8169.150757