The convergence was so fast when we used the educated guess spectrum as a default spectrum in the unfolding procedure and the resulting solution was identical for both deconvolution
codes (same shape as in Fig. 5). This suggests that we are closer to the “real” solution.