Dev Stack Source

Answer : Better make R2 switchable or variable with 1k being the upper limit. But the basic approach is probably OK. Speaker signals are relatively high voltage and relatively low impedance (meaning they can deliver a lot of current). What that means depends on the speaker; anything from a couple of volts and a fraction of an amp (total power 0.25 watts or so) for a little multimedia speaker up to tens of thousands of watts for an AC/DC gig... Microphones are delicate devices delivering millivolts into a high impedance input (low current, very low power). So you need to attenuate the speaker output (reduce the voltage) to avoid overloading the microphone input. The circuit you provided will protect it from damage, but it might still overload enough to distort the input signal. HOWEVER if you are referring to the coloured connectors shown, the green one is "Line Out" - 0.1 to 0.5V rms, not enough power to drive a speaker directly; a lot of PC speakers have amplifiers built...

Answer : OP here – I'd just like to add to the answer below, that I've been in contact with Apple Support. Explanation Apple claims that the poor Mono 8kHz quality which affects recording and indeed simultaneously playback on Mac when the AirPod microphones are activated, is because the SCO codec then gets employed over the entire Mac audio system. This is supposedly "expected behaviour" when trying to use the AirPods and other Bluetooth headsets together with a computer, according to Apple. The AAC codec is normally used when just listening to playback on the AirPods. It's just very unfortunate that SCO – low-quality as it may be – upon AirPod microphone activation is not only limited to doing recording, but also displaces AAC and audio playback. Apple Support claims that Apple is looking at this issue, and that improvements might be coming in future firmware updates, but I did not interpret that as a promise to be honest. But for the time being, I'd ...