I don't understand the solution. A 90° splitter is intended as a lossless (in matched case) device. As said, it's not feasible at audio frequencies. The best lossless impedance matched splitter at low frequencies (not DC) would be a transformer, in my opinion.
A resistive splitter in contrast will work downto DC, but has losses. If you don't require a constant transmission factor and impedance matching at low frequencies, then a combination would be feasible. 90° phase shift at low frequency would still require large LC values.