Strange DC offset with laptop audio output and some noise.

[Corben]

If I connect my oscilloscope to my laptop audio output, I see a DC offset of about 50mV. If I place a 4.7µF ceramic cap between the probe and the signal wire, the probe ground connected to the signal ground, the DC goes away as I'd expect. But if I then put a 68K resistor between the probe and ground, simulating the DC decoupling cap and volume pot of a preamp, the scope shows DC again even though the cap is still there. The scope switch in the AC position does remove the DC.

There's also some 20mV noise that appears digital even when the probe is shorted to the signal ground, so I suppose there is some SMPS noise in a ground loop, which is annoying. The laptop charger is connected to the same power strip as the scope and I can't run the laptop with the battery as it's at the end of it's life. Any ideas what's going on here?