Harmonizing Curve Subtraction Parameters across ASTM and GB Textile Methods
Harmonizing ASTM and GB curve subtraction parameters requires unifying baseline anchor coordinates, invariant scalar peak ratios, and grid registration.
Savitzky Golay Smoothing
Digital Filter
Digital signal processing algorithms located within spectrometer software pipelines eliminate high-frequency instrument noise from raw spectral data without blunting narrow absorption peaks. When evaluating infrared spectra of synthetic yarn treatments, Savitzky Golay smoothing fits localized low-degree polynomials across adjacent data points using simplified convolution matrices. This mathematical smoothing preserves spectral line shapes and peak heights far better than standard moving average filters.
Polynomial Convolution
The mathematical filter operates by moving a localized window across consecutive wavenumber data points in the raw infrared spectrum. Within each local window, Savitzky Golay smoothing performs a unweighted linear least-squares fit using a predetermined polynomial degree. This process reduces high-frequency noise generated by detector electronics while maintaining the underlying physical shape of textile absorption bands.
Selecting an overly wide window width flattens narrow absorption peaks, whereas an overly small window fails to remove electronic baseline noise. Proper selection of polynomial degree and window frame size ensures accurate quantitative analysis of subtle chemical finishes.
Noise Reduction
Reducing random spectral noise improves signal to noise ratios in low-reflectance fabric samples and dark-dyed technical textiles. Smoothed baselines enable accurate automated peak picking in computerized dye library matching.
Spectrometric Verification
Analytical laboratories verify smoothed spectral derivative data to pinpoint precise peak locations for functional group identification in unknown synthetic fibres. Consistent algorithm settings ensure reproducible batch testing across separate laboratory locations.