my own personal sense is that the plate choke might be good... and I'd bet even more money on a "grid choke" (ac choke) in lieu of the 1 meg resistor might even be more hip. But, I obviously like iron :=)And AC choke (with as little as say 500H all the way up to 5000H) in parallel with the output capacitor would enable you to reduce the size of the 4.7mfd output coupling capacitor.
with the 4.7mfd cap and a 500H ac choke in parallel.... the resonant frequency would be 3.28 hertz.
so...say you back off of the cap size and reduce it to 1mfd. then with a 500H choke your resonanct freq would be 7.11 hertz. More than a full octave lower than twenty hertz if you take that to be your lowest freq of interest.
a 1mfd cap with a 1000 henry choke would get your resonance down to 5.03 hertz.
a .47mfd cap with a 1000 henry choke would get your resonance down to 7.34 hertz.
increasing the choke's L while keeping the c constant lowers the resonance. the lower the amount of L the larger the C needed for the same resonance point.
but the beauty here may be that you could use caps as small as say .22mfd or .47mfd with an appropiately sized ac choke (in the output part of the circuit) and still get response well below the audio band.
this may yield sonically even greater improvements than just switching to a plate choke as opposed to using the CCS as drawn.
MSL