Yes it does, but there is more, the ohmic response at Audio - run the values thru for 1kHz to come up with the ohmic result - lots' more.
The coulomb charge "capacity" at the higher REASONATE frequency is different (Greater) - you don't want that cycling thru the amp as a unit. So it's supposed to bypass it. (lowest ohmic value means the more of the power at that frequency is diverted)
Remember the envelope issue - a range of low frequency to high frequency - think of it as heterodyne only you make room in the lower frequency waves, for squeezing in the higher audio frequencies.
And also - a 2 to 4 ohm speaker coil has CONSIDERABLY more inductance which affects the bandpass using the L at mili-Henries not MICRO-Henries.
The Higher frequencies trying to enter into that coil will encounter a considerable amount of reactance - so they are trying to use that T1 as a FIXED inductor - so they need a path to remove the otherwise parasitic high-frequency oscillations and smooth out the waveform going into the transformer so the losses are less going thru it,
Below is a crude example of what I'm talking about...
The graphic is only trying to show SPECTRUM as a range of power - if you go back to the CB tricks days, I also talked about PHASE angle inclusive in the Asymmetrical Mods thread... you can think of the above as a indication of the DISTORTION the power input will make because not all the frequencies entered into T1 will make it out alive - at least not without considerable modification...
Because Inductance and the effects of DC bias flowing thru the transformer at the same time - skews the results - it's not symmetric and we have distortion products too (Grey / pink area added in) because of that inductor changing and "borrowing" from both Audio and Bias to make this envelope - and why a narrow bandwidth is better for to reduce the distortion products to increase intelligibility as the cost of fidelity - so they use the heavy inductance of the T1 coil winds and the amps ability to have an output of a higher frequency bandwidth - so they use more "treble" to offset the "bass" the T1 will take in more readily.
DC carrier will always be there, but the envelope forms AROUND it like a sphere - so the angle of power DEVIATES from the carrier as a means of frequency to power vector - so less higher treble equates to less DEVIATION or re-vectoring (if that can be used here) of the power around the DC steady state the carrier bias is.
Why am I telling you this?
Simple , the Cobra, Uniden, Radio Shack - any AM-Only CB radio using a Reactance Modulator has a DC blocking cap that when combined with all that inductance on the other T1 used in these AM-only units - they produce a bandwidth product of power, BIAS and Audio Envelope - throw it thru a Spike Diode to remove the negative component and there you have it...
So they are making the TDA's do the push pull with enough oomph to make the envelope without introducing the DC component to it - it's supposed to be separate - but they also generate a considerable amount of EMF kickback that is a lot like a spark plug coil - induce a large current change in a small coil induces a larger voltage change in the larger wind coil of the plug side...
So that is why the Bandwidth on the amps themselves - so they don't interact and generate a larger phase issue of similar and dissimilar (a heterodyne effect of sound some people feel as a pulsation and hear as a squeal - which can cause Cardiac Arrests due to this low-level motion and it's sonic effect- SEE: Rollercoaster The Movie...) Hence the "Smashing" moniker I used as a reference to this event I posted earlier .
Have fun for Labor Day!