Journal of the Korean Magnetic Resonance Society, v. 6, no. 1, page. 12-19
Abstract
A shift averaged Harr wavelet transform is introduced as a new and excellent tool
to distinguish real peaks from the noise contaminated NMR signal. An application of
discrete wavelet transform (DWT) to the noise suppression in time domain NMR data
has been successfully accomplished. A simple DWT with a set of Daubechies wavelet
coefficients (1/2, -1/2), Harr wavelet in combination with shift averaging of
NMR data was applied to the noise elimination. New algorithms of shift averaged
Harr wavelet were quantitatively evaluated in terms of threshold and signal to
noise ratio (SNR), and were utilized to suppress several noisy NMR spectra for
application.