Query free-field transmission functions for ear in FluctuationStrength_Osses2016

[mlotinga]

I'm wondering if the transmission filter used to represent the ear is correct - the functions used to define the filter coefficients do not seem recognisable from the forms given and the cited sources.

One internal function given is il_calculate_a0, which cites Fastl, 2007 Fig. 8.18, page 226, yet the values given in a0tab do not resemble (the inversion of) that figure. Another comment given states % lower slope from middle ear (fig_a0.c, see Figure_Psychoacoustics_tex) but it isn't clear what data that refers to, although the lower and upper critical band parts of the function are more consistent with middle ear roll-off.

However, il_calculate_a0 is not actually used in the function, as it is commented in line 443 in preference for il_calculate_a0_idle. Furthermore, the alternative values given in a0tab within il_calculate_a0_idle do not appear to resemble any plausible free-field transmission function for the outer, middle, or outer+middle ear.

To me, the values in il_calculate_a0 appear more generally plausible as representing an ear transmission, and more closely match the cited source. However, this is not the function in use within FluctuationStrength_Osses2016.

Comparison with other transmission functions also suggests that the function shown in Fig 8.18, and represented in il_calculate_a0 (but not in the functional il_calculate_a0_idle) may be more closely representative of a diffuse field response than a free-field frontal incidence plane wave response. These plots are of the outer-middle ear filters implemented in ECMA-418-2:2022 for both field types:

[mlotinga]

Switching the function to use il_calculate_a0 instead of il_calculate_a0_idle I get validation outputs as below - some better, some worse. Perhaps better agreement might be achieved with a filter with a more representative free-field response, which better represents original test conditions (free-field-equalised headphones).