Speaker cable for the LS50 Meta: does 20 AWG instead of 12 AWG change anything?
Over a 4.6 m run, thin cable costs 0.32 dB against the nominal 8 Ω — and 0.79 dB at the 3.2 Ω dip. The loss is not flat, which is the part cable arguments always miss.
Yes, but not the way cable arguments usually go: the loss is not flat across frequency.
Over a 4.6 m (15 ft) run, 20 AWG has 0.30 Ω of loop resistance. Against the LS50 Meta's nominal 8 Ω that is 3.7% of the voltage — 0.32 dB, inaudible. But the speaker is not 8 Ω. At the 3.2 Ω dip near 100–200 Hz the same 0.30 Ω eats 8.7%, or 0.79 dB. The cable is a voltage divider working against a curve that swings by 2.5 to 1, so what thin cable removes is bass, not volume. It behaves like a gentle tone control that you did not install.
With 12 AWG the loop falls to 0.048 Ω: 0.6% at nominal, 1.5% at the dip. The damping factor at the terminals rises from 23 to 82.
What to do. Half a decibel of frequency-dependent bass loss is not a catastrophe, and 16 AWG (0.12 Ω, damping 47) already removes most of it — that is the sane default for a run this length. The money matters far less here than the amplifier's minimum impedance rating. But the general claim «cable gauge is inaudible» is only true for speakers with a flat impedance curve, and this is not one of them.