Literature Review - Speech enhancement



Speech enhancement is a technique that is concerned with improving the perceptual and tangibility of a speech signal by the human auditory system. In a lot of case a speech signal is degraded by some environmental noise or other additional noise which effect the human perceive and understand the information in the speech itself. It is strongly believed that human could perceive the information from a speech signal by understanding the connection between each phoneme and words that forms a sentence which then comprehended by the brain to give some meaningful understanding.

The need to enhance speech signals arises in many situations in which the speech signal originates from a noisy location. There is a wide of variety of scenarios in which speech enhancement is desired. Voice communication over noisy line, two ways communication in a noisy environment such as a factory or airport is a perfect example of such cases. In these situation, speech enhancement algorithm could be used to improve the quality of the speech at the receiving end.


There are many ways to classify speech enhancement methods. It is usually difficult for a typical algorithm to be able to perform equally across all noise types. Therefore, usually a speech enhancement system is based on certain assumptions and constraints that are typically dependent on the application and the environment  which is the key factor in determining which method is the best to use(Loizou, 2007). The simplest approach to divide the speech enhancement category would be to divide it into single channel and multiple channel method.

Non-Linear spectral subtractions is one of the single channel speech enhancement method that is explained in detail by Yoon in his research paper that is published in 2007. He pointed out that there are certain type of noise may affect the low frequency region of spectrum more than high frequency region.

Later, Young made a significant contribution in the speech enhancement field by proposing an adaptive noise cancellation as a powerful multi channel speech enhancement technique. This technique is based on  the availability of an auxiliary channel, known as reference path, where a correlated sample or reference of the contaminating noise is present. This reference input will be filtered following an adaptive algorithm, in order to subtract the output of this filtering process from the main path, where noisy speech is present. The adaptive noise cancellation cancels the primary unwanted noise by introducing a canceling anti-noise of equal amplitude but opposite phase using a reference signal


Comments

Popular posts from this blog

AWL of the day - Part 2

Book Review : Backstage Pass - a compelling adventure of a rockstar fans

Making an Argument in Academic Writing