Here is the problem almost nobody warns you about: macOS lets you pick exactly one input device per app. One. If you have a good desk mic, a lav clipped to your shirt, and a webcam mic, your Mac does not blend them or pick the best one. Zoom just uses whichever single device is selected in a dropdown, and everything else in the room is ignored.
That is fine if you never move. Most people move constantly — turning to a second monitor, leaning back, standing up to grab something. The moment you turn away from the one selected mic, your voice gets quieter and roomier while everything else stays exactly as loud as it was.
There are four genuine ways to deal with this on a Mac. They are not equally good, and which one fits depends on how much fiddling you are willing to do.
1. Aggregate Devices (free, built into macOS)
macOS has a built-in tool most people never open: Audio MIDI Setup (in Applications → Utilities). It can create an "Aggregate Device" that bundles several physical inputs into one device your apps can select.
To build one: open Audio MIDI Setup, click the + in the bottom-left, choose Create Aggregate Device, then tick each microphone you want included.
Here is the catch, and it is a big one. An Aggregate Device does not mix your microphones — it stacks them as separate channels. Three mics become a 3-channel device. Most conferencing apps read only the first channel or two, so in practice you have not solved anything; you have just made a device with extra channels your app quietly ignores.
Aggregate Devices are genuinely useful for multitrack recording in a DAW, where separate channels are exactly what you want. For "make my voice sound good on Zoom," they are the wrong tool.
2. A hardware mixer or audio interface
The traditional answer. Plug every microphone into a physical mixer, blend them there, and send one combined signal to your Mac over USB.
This works, and it is what broadcast studios do. The downsides are money and attention. A multi-input interface is a real purchase, it takes desk space, and — the part people underestimate — you have to actually ride the faders. If all your mics stay open at once, you are not capturing one clear voice; you are capturing your voice plus every room reflection picked up by every other mic. That is what makes amateur multi-mic setups sound hollow and far away.
Open microphones also fight each other. The same sound arriving at two mics a few feet apart arrives at slightly different times, and when you sum those signals some frequencies cancel out. Engineers call this comb filtering. You hear it as a thin, phasey, slightly underwater voice.
3. Software routing utilities
Tools in the virtual-audio-device category (BlackHole, Loopback, and similar) let you create software devices and wire inputs into them. This is more flexible than a hardware mixer and takes no desk space.
It is also a genuine routing project. You are building signal paths by hand, and you still face the same underlying issue as the mixer: if multiple mics are live simultaneously, you get the same phase and room-noise problems. Routing solves "how do I combine these" — it does not solve "which one should actually be listening right now."
4. Automatic microphone selection
The fourth approach reframes the problem. Instead of blending every microphone all the time, continuously evaluate them and let the one hearing you most clearly carry your voice — handing off smoothly as you move.
This sidesteps both failure modes above. Only one mic is meaningfully live at a time, so there is no comb filtering and no stacking of room noise from four different positions. And because the selection is automatic, there is no fader to ride.
Two details matter a lot in practice:
- Per-microphone noise floors. A cheap webcam mic and a good condenser have very different baseline hiss. Comparing their raw levels is meaningless. What matters is how far your speech rises above that particular microphone's own noise floor.
- Crossfades, not hard cuts. Switching instantly mid-word is audible and jarring. The handoff has to be gradual enough that a listener never notices it happened.
This is the approach SoundMesh takes. It watches up to 8 inputs, measures speech against each mic's own noise floor, and hands your voice to whichever one is cleanest — then presents the result to every app as a single device called SoundMesh Virtual Mic.
Which should you actually use?
- Recording separate tracks to edit later? Aggregate Device into a DAW. It is free and it is the right tool.
- Multiple people in one room, all talking? A hardware mixer, ideally with per-channel gating.
- Complex routing between apps (piping system audio, multi-app captures)? A dedicated routing utility.
- One person, several mics, wants to move freely and not think about it? Automatic switching.
The honest summary: the first three all assume you want to combine microphones. Most solo desk setups do not. They want the right microphone, chosen for them, moment to moment — which is a different problem with a different solution.
A note on placement
Whichever route you take, physical placement still does most of the work. Mics roughly a fist to a foot from your mouth, angled toward you, off the desk surface if possible (desks transmit typing and bumps straight into the capsule). No amount of software fixes a microphone pointed at a wall.