Orthogonal Signal Correction

Orthogonal signal correction (OSC) is a data preprocessing step that can be applied to spectral data to achieve the same aim (or should this be objective?) as methods such as MSC, SNV or differentiation: removing scattering and instrumental effects whilst preserving the signal in the spectrum. It is described in two recent papers.1,2 OSC could be used to pretreat spectra for subsequent input to any calibration method, but it is most easily understood by thinking in PLS mode, perhaps because it comes from Umeå. PLS constructs factors from the spectral data, X, decomposing it3 as