There is more than one approach used to determine the Signal / Noise for a Raman spectra. For example, one approach is to use the difference of peak and background signal, divided by the square root of the Background signal. In this case the peak signal is measured at the main Raman peak (E.g., Si samples have a single peak at ~520 cm-1), while the noise can be measured in a region) where no Raman signal is present (e.g., for Si one may look at ~1500 cm-1, which will not be impacted by any factors except for the noise). For an ideal system this signal will be zero, but in reality, the noise signal is low but not zero (thermal effects, the set up for the signal sensitivity level, etc.). Another commonly used method is when to divide the difference of signals defined as; Peak signal – Background signal, by the root mean square of the background signal associated to noise.
" The SNR for a particular measurement is rigorously defined as the inverse of the RSD of the measured value. The SNR for the peak intensity of a Raman band is the average peak intensity, usually above the baseline, divided by the standard deviation of the peak intensity." R. L. McCreery, "Raman spectroscopy for chemical analysis" (Chemical analysis; v. 157), 1st edition; Wiley-Interscience, 2000. ISBN 0-471-25287-5. See previous answer of: @Aurelio Sanz Arranz
Bahram Roughani Thanks for explaining, Its clear for Si, but what if i have a spectrum from single bacterial cell, phenylalanine band is my marker band. Also spectrum has many peaks from different bio-molecules. Its quite confusing for spectrum of bacterial cell
Priyanka Mishra The signal to noise measurement depends on the signal being investigated. The noise is basically the same for all signals present, but the signal to noise will be different for each signal, depending on the strength of the signal comparted to the background noise.