Extended Data Fig. 8: Simulation of pulsed-SBS spectral accuracy under different scan ranges.
(a) The Brillouin shift and linewidth accuracy of single-peak fitting as a function of scan range. The signal to be fitted is the simulated sum of a Lorentzian (with an amplitude of 0.625, peak shift 5.5 GHz, FWHM 0.63 GHz) and a random noise with a standard deviation amplitude of 0.0312 (that is SNR of 20). (b) The Brillouin shift and linewidth accuracy of a double-peak fitting as a function of scan range. The signal to be fitted is the sum of two Lorentzian waveforms (L1 with an amplitude of 0.625, peak shift 5 GHz, FWHM 0.63 GHz, L2 with an amplitude of 0.5, peak shift 5.4 GHz, FWHM 0.63 GHz) and a random noise with a standard deviation of 0.0312 (that is SNR of 20). There are 100 frequency points in the scans, to mimic the actual experimentally acquired spectra. The accuracy and error bar are the mean and standard deviation of 100 independent simulations, respectively. Note that above 1.5 GHz, the double peak is fitted with L1 and L2, but between 0.8 and 1.5 GHz, only L2 gets a meaningful fit and L1 is minimized because it doesn’t recognize the second peak.
