Thank you indeed for your gracious contribution! I am lookong forward to try some of these scripts.

The audio-sidechain script observes the volume meters on either a bus (as the aggregate of inputs) or a single input and reduces (ducks) the volume of another bus if the volume on the observed side goes a threshold. We for instance place all inputs which are shown on the stage TVs to a bus B and place all microphones on bus A. The script then ducks all microphones if anything (for us it's usually a playing media or a guest in vMix Call) is sent to the speakers of the stage TVs. The default configuration of the script reflects exactly this setup.

Ah, now I better understand what type of audio ducking you want to achieve. Well, in theory you could do: create bus A/B/C, place MIC1 on B/C, MIC2 on A/C and MIC3 on A/B, and then instantiate the script three times -- once for monitoring MIC1 (inputMonitor="MIC1") and ducking MIC2,MIC3 (busAdjust="A", busAdjustInputs=true), once for monitoring MIC2 (inputMonitor="MIC2") and ducking MIC1,MIC3 (busAdjust="B", busAdjustInputs=true), and once for monitoring MIC3 (inputMonitor="MIC3") and ducking MIC1,MIC2 (busAdjust="C", busAdjustInputs=true). But I personally would not do this for your particular scenario, because the audio-sidechain is really just a "rough" audio ducking and intended for global/rough/simple/unprecise/all-at-once ducking inside vMix.

For your particular scenarios, balancing three mics the WTAutomixer really IS(!) the best solutions. We use it ourself IN ADDITION to audio-sidechain. The WTAutomixer balances the incoming microphones in Cantabile BEFORE ingest into vMix and audio-sidechain ducks all microphones inside vMix for emitted stage sound. The reason is that WTAutomixer is actually balancing and not just ducking: it knows that close mics always have sound bleed from one talent into the mic of the other and it tries to very precisely figure out who is the speaker and duck only the others. But WTAutomixer also knows that if there is no clear speaking, that two of three microphones have to stay open, etc. So, I would never use audio-sidechain script inside vMix for precise microphone balancing/ducking. Here WTAutomixer really is the best solution on the market IMHO. The audio-sidechain is for rough ducking of audio inside vMix. Yes, we also used in the past (on a different vMix instance) audio-sidechain also for translator voice-over. So, audio-sidechain is flexible. But keep in mind that it never can be as precise as WTAutomixer or a compressor VST plugin with a real side-chain functionality (but which cannot be used inside vMix).

Thank You 👍👍