Electronic – Low voltage 2 channel stereo analog switch

Edit: Actually, I found out the answer I posted is already given in the question you rejected as "too complicated". I ask you to reconsider that judgement, as anything involving operational amplifiers (suggestions also found in that question) will be even more complicated.

While the commentors are right that your circuit is a really bad idea in practice, your thoughts that lead you to the that circuit are actually quite good. You notice correctly that the voltage you obtain at typical volumes is too low to turn on the transistor (you need around 0.6V to get a significant effect). Using a darlington transistor doesn't help, as you experienced, because you need even more voltage to turn on a darlington transistor, but you can get away with a lot lower current. You problem is that your voltage is too low, not that your supply can't deliver enough current, so forget about darlingtons.

Increasing the "baseline voltage" as you call it, is exactly the way to go. Electronics people often call this process "adjusting the DC bias" of a transistor amplifier. It can be done quite easily on your circuit without adding a dangerous low-impedance power source like a battery. Let's work on your circuit. First, add a current limiting resistor into the LED circuit and put the LED on the other side of the transistor. It makes no difference in your circuit as it is now, but it will make soon:

^{simulate this circuit – Schematic created using CircuitLab}

You can fix your problem, as you correctly identified, by providing DC bias (a baseline voltage) to the transistor base. But you should not force a DC voltage into the headphone jack, so you need a capacitor for DC decoupling. This capacitor should pass the AC music signal, but block the DC required to bias the transistor. With the circuit as drawn, you can get the required DC bias (around 0.45V) from the battery using a voltage divider like this:

^{simulate this circuit}

Most DCDC converters run well above 20KHz at your power levels .You are getting Audible noise and I think that you should put the oscilliscope.You will probably see ripple artifacts on the 5V input and 1.5V output that are in the low KHZ range.When you scope the mosfet drain waveform you will probably see shimmering on the rising or falling edges depending on how you have got your scope triggering.Even when the DC input and load are constant the shimmering will persist.This low frequency crud cuts through the orthodox SMPS filter components like a knife through butter because those components are designed to attenuate the switching frequency.Experimentaly loading the converter may cure this because these low frequency instabilities tend to be worse at light load .If the load experiment nails this then you can increase the buck inducter and you will probably have silence or use a smaller buck converter which will in effect be more heavily loaded.If things get too bad consider linear post reg using a 1.5V LDO and setting the buck up at say 2.5V .