Addition of Simultaneous Masking

The masking from different maskers has to be added to form the final masking function. One could argue that the maskers should be added as powers or amplitudes. According to [1], though, addition of two equal maskers can give a resulting masker up 12 dB higher than two maskers alone. This would mean, that the addition of maskers would be defined as follows:

$$SUM(M_1, M_2) = 40log(10^{M_1/40}+10^{M_2/40}),$$ (17)

i.e as the addition of square roots of amplitudes. This model does not work for big differences in $M_1$ and $M_2$, so in the coder the addition was defined as addition of amplitudes:

$$SUM(M_1, M_2) = 20log(10^{M_1/20}+10^{M_2/20})$$ (18)

Using this model for summation, all spectral components of the audio are considered maskers, and their individual masking thresholds (calculated in section 3.2.5) added.

An example of the produced masking threshold, including both simultaneous and temporal masking (described in section 3.3), can be seen in Fig. .