Perceptual Speech Enhancement Using Hilbert Transform

A new speech enhancement algorithm using a Hubert transform (HT) based time-frequency (TF) representation of speech signal with respect to human perception is proposed. TF representation is carried out by the means of analytic decomposition of speech signal in the hearing critical bands (CB) where the envelope and phase components of the analytic signals are used. For the purpose of enhancement the envelope in each CB is modified, based on the conventional spectral subtraction method, using a time varying gain function which takes into account the threshold of hearing. This threshold is calculated on the basis of masking effects of all bands using a perception model. Signal is reconstructed from the modified envelopes and the original phases of noisy signal in critical bands. Experimental results show that using the threshold of hearing in which temporal masking is included can effectively eliminate the musical noise without a significant decrease in intelligibility