|Code Listing by Esfandiar Zavarehei|
Calculates the average log-spectral distance between CLEAN and NOISY signals. It has the capability of calculating this distance for a specified sub-band as well. This measure is used for evaluation of processed speech quality in comparison to the original speech.
The spectral Subtraction Method for enhancement of noisy speech signals proposed by Boll 79. The method implements spectral averaging and residual noise reduction proposed in the paper. Note that the first 0.25 sec of the speech signal is assumed...
Multiband spectral subtraction as proposed by Kamath 2002. Uses adjusts the subtraction coefficient with the frequency as well as the SNR. note that the first 0.25sec of your signal is used to model the noise.
Calculates the average log-spectral distance between CLEAN and NOISY signals. Has the option of calculating this distance for speech active periods only. Also you can select a frequency range to study the distance in a sub-band.
This is an implementation of Ephraim's MMSE STSA method (Ephraim 1984). The decision-directed method is used for tracking of a priori SNR tracking.
This function is for training a codebook for vector quantization. The data set is split to two clusters, first, and the mean of each cluster is found (centroids). The disttance of each vector from these centroids is found and each vector is...
This is an implementation of Ephraim's MMSE log-STSA method 1985. The decision-directed method is used for tracking of a priori SNR tracking.
Spectral Subtraction method proposed by berouti for enhancement of noisy speech signals.