There was a lot going on there for certain.

The gist of it is that owing to the rise in the impedance of the headphone driver at 85 Hz and nearby, the effect of the 330 Ohm series resistor will be that the headphone's share of Vout (the amplifier output voltage) will be greater at 85 Hz than elsewhere, to the tune of 2.2 dB (in power terms). This is enough to be audible but is small in comparison to the existing response variations of the headphone. The rise in headphone impedance at 85 Hz would not produce this effect were it not for the 330 Ohm resistor in series with the headphone jack (in each channel). The same kind of thing will happen if you put a power resistor in series with a speaker, if the resistor's value is in the ballpark of the speaker's impedance. It happens because the speaker driver's share of the output voltage depends on the ratio of the driver's impedance to the sum of both impedances, which ratio will be different at different frequency if the speaker driver's impedance is not constant.

It is a small concern that most people might not notice. The solution that I would prefer, if it were me, would be to use a series notch filter in parallel with the headphones (more or less a short across the headphone driver), one for each channel. The reason I like this approach is that it will flatten the effective impedance of the headphones, such that the headphone's share of the amplifier output voltage will be essentially constant throughout the frequency range.

When a notch filter is placed in parallel with the headphone driver, it is called a series notch filter because the three components that make up the filter are wired in series with one another. The resistor establishes a lower bound for the impedance of the filter at the notch frequency, at 85 Hz in this case. The capacitor is chosen to increase the impedance for frequencies below the notch frequency, and the inductor is chosen to increase the impedance for frequencies above the notch frequency. The impedance of the series notch filter is low only at the notch frequency, thereby lowering the effective impedance of the headphones at this frequency. At other frequencies the series notch filter has high impedance and the effective impedance of the headphones is little affected.

I didn't finish the solution, so I cannot rule out the possibility that there isn't going to be something problematic about the solution. I cannot rule out the possibility that there will not be any combination of component values that will simultaneously solve these three equations. Given that there are three unknowns, you need three independent equations that must be satisfied simultaneously, in order to obtain a unique solution. One of these equations is written such that the effective impedance of the headphone driver at 85 Hz will be 300 Ohms. It was trivial to calculate that the impedance of the notch filter itself would need to be 710 Ohms at 85 Hz, in order that the combined, parallel impedance of the notch filter and the headphone driver will be 300 Ohms. The other two equations were obtained similarly, using 35 Hz and 1.1 kOhm, such that at 35 Hz the parallel impedance of the notch filter and the headphones will be 300 Ohms, and similarly for 200 Hz, such that at 200 Hz the parallel impedance of the notch filter and the headphones will be 300 Ohms. These two frequencies, 35 Hz and 200 Hz, where chosen because these are the approximate frequencies where the increase in impedance is half what it is at 85 Hz. But I will say again that not having completed the solution, I am not very confident that it is possible to solve these three equations simultaneously. I have the sense that I'm overlooking something.