Split always gives half of the inductance and parallel, 1/4 of it. So, normally, parallel is quieter (electrically). And it shouldn't have a "hollow sound": if anything, it "symmetrizes" the response of both coils, cancelling the slight comb filtering effect always due to series path with a 4 cond. wiring.
Maybe what you did find "hollow" is the effect of the resonant peak way higher in the spectrum, as implied by the divided inductance. But what you describe doesn't match any of my experiments with parallel wiring these last four decades so I wonder if an unseen problem of wiring wasn't there in this case (I've seen plenty of situations where wiring issues were making conductors accidentally resistive and/or capacitive).
Maybe "reverse wiring" of coils would give the small amount of aliveness that you want. Try series with black and green altogether. Red to ground. White to hot. Both coils will still give normal series wiring but with opposite start and finish paths. Example of measured electrical effect illustrated below (with reverse wiring giving a brighter response in this case).
View attachment 6326439