A Gain Bounded Speech - enhancement Algorithm for Improving Intelligibility ?

A higher intelligibility subspace speech-enhancement algorithm based on a theoretical upper bound on the gain function is proposed. The majority existing speech-enhancement algorithms cannot effectively improve enhanced speech intelligibility. One important reason is that they only use Minimum Mean Square Error (MMSE) to constrain speech distortion but ignore that speech distortion region differences have a significant effect on intelligibility and the theoretical upper bound on the gain function does exist for speech intelligibility. A priori Signal Noise Ratio (SNR) is estimated in order to compute the upper bound for the highest value allowed for the gain function. Then the values of the gain function are bounded in accordance with the theoretical upper bound. Both objective evaluation and subjective audition show that the proposed algorithm which can limit the values of the gain function to fall within this upper bound does improve the enhanced speech intelligibility.