Electronic noise suppression in the wavelet, frequency, and time domains
Juan I. Arvelo · The Journal of the Acoustical Society of America · 2005
Automatic machine gun recordings were severely contaminated with electronic noise. A detailed examination of the character of the electronic noise revealed that it consists of periodic pulses at an interval of about 4.2 ms and that its frequency content is rich in narrow-band lines with a colored broadband background. Given the impulsive nature of the electronic noise in the time domain, wavelet decomposition was applied in an attempt to harness the noise and to null it out. An alternative approach was also implemented where a bank of stop band Butterworth filters were applied in the frequency domain to suppress the narrow-band components of the electronic noise. Other approaches include the use of adaptive cancellation in the time and frequency domains. We will demonstrate and compare the performance of these approaches emphasizing the need to minimize their influence in distorting the signal of interest.