# Multiple Options for Passive Audio Mixer

I have two mono audio inputs that I need to passively combine into a single output. Both inputs have built-in volume controls so I don't need any volume control in the mixer. I've been looking at ways to do this, and have seen a couple common options. About 50% of suggestions seem to be for something like this:

While the other 50% are for something like this:

So, my questions are:

1) What's the deal with the 20k resistor between the signal and ground wires in the second schematic? What purpose does it serve and why does the first schematic not have it?

2) Realistically, how important is the resistance value of the other resistors (1k in the first schematic, 475 in the second)? I understand it has to do with loading and volume reduction, but not sure how to choose a good value. In schematics for this kind of circuit, I've seen values ranging from 10 ohms to 10k ohms for those, with no decent explanation for the 1000x difference.

For context, this is summing the unbalanced mono input from two voice communications radios into a single unbalanced output, so audio quality is not super important.

Assuming your sources have the same nominal impedance, Zs, the resistors are the same value, ZR, and your load impedance is ZL (and if I'm doing my algebra correctly), your level of voltage attenuation (V/V, not dB) would be: $$A_v=\frac{(Z_S+Z_R+Z_L)}{(2\times Z_L+Z_S+Z_R)}$$ The total output impedance of the sources and mixer as seen by the load will be: $$Z_{out}=\frac{Z_S+Z_R}{2}$$ The input impedance seen by one source will be: $$Z_{in}=Z_R+(Z_L)||(Z_R+Z_S)$$