Response of the column densities derived using the combined NIR+MIR technique to the adopted NIR-to-MIR opacity-law. The curves show the ratio of column densities derived with alternative opacity-laws () to the one adopted in this paper (). The curves were calculated by keeping the NIR opacity fixed and by changing the relative MIR opacity. The solid blue, green, and red curves show the case in which the true opacity-law is 15%, 30%, and 60% higher than the adopted value (). The dashed curves show the corresponding curves for the cases in which the ratio is lower. At A_V ≲ 10 mag the technique relies dominantly on the NIR data, and therefore the changes in MIR opacity do not greatly affect the column density. The large-scale background component that is filtered out by the MIR data (and assumed to be recovered by the NIR data) has its maximum at A_V = 10−20 mag. Thus, at that range there is a transition from NIR-dominated to MIR-dominated regime. Finally at A_V ≳ 20 mag the large-scale column density component starts to be small compared to the total column density, and thus the ratio approaches the value by which the original opacity law was modified (i.e., factors 1.15, 1.3, and 1.6). At extinctions lower than the 10 mag threshold value, there is a small bias in the extinctions. It is caused by the fact that the background correction value (see Sect. 3.1) for some pixels can result from interpolation from neighboring pixels instead of from the difference of the NIR and MIR extinctions. The pixels for which this interpolation is performed are not necessarily the same in the cases where n = 1 and n ≠ 1.