EFFICIENT SNR-BASED SUBBAND POST-PROCESSING 
FOR RESIDUAL NOISE REDUCTION IN SPEECH ENHANCEMENT ALGORITHMS
-------------------------------------------------------------
Audio demonstrations
********************

This archive contains audio demonstrations for the "medium SNR" case presented in the paper. The pre-enhancement algorithm used here is the log-spectral amplitude estimator from Ephraim and Malah [1]. There are three conditions tested, using noise files from the NOISEX database and speech from the TIMIT database (resp. [2] and [3]).

* 'Cafeteria.wav', 'Cafeteria_enhanced.wav', and 'Cafeteria_postprocessed.wav' respectively represent the noisy speech in cafeteria noise conditions, the pre-enhanced signal with Ephraim and Malah's algorithm, and the postprocessed file using our proposed method.
* Similarly, 'Factory*' and 'Military*' follow the same terminology for the Factory and the Military vehicle noise types.


REFERENCES:
**********

[1] Y. Ephraim and D. Malah, "Speech enhancement using a minimum mean-square error log-spectral amplitude estimator," in IEEE Transactions on Acoustics, Speech and
Signal Processing, Vol. 33, Issue 2, pp. 443–445, April 1985.

[2] A. Varga, H.J.M. Steeneken, M. Tomlinson, D. Jones, "The NOISEX-92 study on the effect of additive noise on automatic speech recognition," Technical report, Defence Evaluation and Research Agency (DERA), Speech Research Unit, Malvern, United Kingdom, 1992.

[3] J.S. Garofolo, L.F. Lamel, W.M. Fisher, J.G. Fiscus, D.S. Pallett and N.L. Dahlgren, "The DARPA TIMIT Acoustic-Phonetic Continuous Speech Corpus, NTIS Speech Disc," NTIS order number PB91-100354, 1990. 

