Soft Audible Noise Masking in Single Channel Speech Enhancement for Mobile Phones

Mobiles phones are not only intended for voice communication but also for other multimedia communications. Hence any additional features will increase the space planned for the device. Therefore single channel speech enhancement with soft audible masking is more feasible for mobile phones. This paper aims at development of algorithms for speech enhancement using soft audible masking. Proposed method uses the ideal binary mask to examine and blend the input speech signals in both the domains. Masked Signals are evaluated and quality is measured in terms of SSNR. The illustrated algorithm estimate spectral power of disturbed signal in very short time by masking threshold. Binary decision are made in iterative mode until all the words are estimated, thus a suitable quantity of noise is reduced without disturbing the clean speech. A result demonstrates that the proposed enhancement algorithm rebuild the desired speech signals competently from the despoiled speech.