The selection scaling mask operation which is based on peak detection

摘要

The audio signal which is received tupple of the peak in the audio vector ^S which was re-constructed is detected, the scaling mask which is based on the tupple of the peak which is detected (^S) generates. At least gain vector g generates on the basis of with the scaling mask and the index j which displays gain vector. The reconstruction audio signal scaling is done with gain vector, generates the reconstruction audio signal which scaling is done. The deformation generates on the basis of with the audio signal and the reconstruction audio signal which scaling is done. The index of the gain vector which is based on the deformation which generates is output.