| In practical communication application, such as speech telecommunication, speech synthesis, speaker recognition, the degradation of speech by all kinds of noise is inevitable and it affects the performance of the speech equipments very badly, So, it is necessary to denoise for the degraded speech, this is called SPEECH SIGNAL ENHANCEMENT. The primary purpose of speech enhancement is to extract more pure speech from speech signal with noise.The approaches of speech enhancement are numerous. Several enhancement approaches are discussed in this paper, and some knowledge related to speech enhancement are given, for example, the models of speech and noise. A algorithm of speech enhancement using LMS is mainly discussed based on self-adaptive algorithm. On the basis of presenting an overview of a number of speech enhancement algorithm in common use, the self-adaptive system for speech enhancement is described in detail based LMS algorithm. Because the LMS algorithm is simple, small in calculation and easy to realize in real time signal processing, LMS algorithm is widely used in the technology of speech enhancement. The standard Least Mean Square algorithm(for short LMS algorithm) possesses evident advantages that it is a simple algorithm easy to realize and the misalignment rate of the output is almost zero when the relativity of input signals is zero.But when the order of the filter is bigger and the relativity of the input signals is strong, there appear its disadvantages, such as slow speed of convergence and large amount of calculation and misalignment. The variable step size and parallel LMS algorithm has a higher convergence speed and less misalignment than those of standard LMS algorithm. It is very applicable to realization of DSP as well as processing of voice signals. In this paper a simulation of the self-adaptive speech enhancement was made with Matlab language. For realizing the application of the speech enhancement with LMS algorithm, Simulation result shows It is better than classical LMS. the algorithm can effectively suppress noise in speech, improve SNR and reduce distortion. The simulation results show the improved algorithm can remove a lot of noise.Background noise can destroy the intrinsic acoustical characteristic and model parameter of sound signals, so that deteriorates the communication quality of sound. The adaptive anti-noise speakers can improve effectively communication quality in the strong background noise and is widely applied to industries with high environmental noise. With the rapid progress of techniques on Digital Signal Processor, the price of chip falls, which makes it very useful to research on new fashioned anti-noise speakers based on new hardware and new algorithms. The variable step size parallel LMS algorithm combines with advanced DSPã€A/D and D/A chips in the design of adaptive anti-noise which guarantees the advance of technology of the anti-noise in the aspects of the algorithms and represents the mainstreams of anti-noise speaker at present. Believe that it will be widely applications in the future. |
