That makes sense if you have something like a Motorola piezo tweeter .
i suggest you to move in the opposite direction ,which is having maximum efficiency ,and not throwing useful energy in a resistor that would transform it into heat . Also ,putting a resistor in parallel with a driver ,would mimic the effect of an Helmholtz resonator ...
which is not the case of a tweeter , even if some have a back chamber ...
Check the datasheet. If you accept 10 % THD, ±18 V give you 22 W into 8 Ohm e.g. the tweeter or 40 W into 4 Ohm, so 20 W into the tweeter (9,1 % less) and 20 W into the resistor (waste). And that does not yet take supply voltage sag into account.
adding a resistor to a Capacitor tweeter changes the response curve.
The resistor can be used to correct a response error/characteristic that the designer does not want/need.
But equally the resistor may introduce a response error that sounds wrong.