Core Audio is Apple's audio framework, built into macOS itself, and it is the layer that actually handles sound: every app's playback, every microphone input, and the system's single shared idea of "the current output device." Nothing on a Mac plays or records audio without going through it, whether that is Music, a video call, or the Sound settings pane.
You never open Core Audio directly, and that is by design. It sits underneath the interfaces you do use: the Sound settings, the volume icon in the menu bar, Audio MIDI Setup, and every app that makes noise. Understanding what it is explains a handful of things about Mac audio that otherwise look like arbitrary limitations.
What Core Audio actually does
Every Mac has some number of audio devices attached to it: the built-in speakers and microphone, plus whatever is plugged in or paired, headphones, a USB interface, AirPods, a monitor with speakers built in. Core Audio is the system layer that keeps track of all of them, talks to each one's hardware driver, and gives every app the same consistent way to send sound out or pull sound in, regardless of which physical device is actually connected.
That is the part that matters most day to day: Core Audio maintains one system-wide idea of the current output device and one for the current input device. When you change the speaker in Sound settings, you are not telling each app individually. You are changing a setting Core Audio holds centrally, and every app reads from that same place. That is why switching from your Mac's speakers to headphones changes what everything plays through at once, instead of needing to be set app by app.
Why apps do not each need their own driver
Before a framework like this existed, every piece of software that wanted to play or record sound would, in principle, need to know how to talk to every possible speaker and microphone directly. Core Audio removes that problem. A hardware maker writes one driver that speaks Core Audio's language, and from that point on, every app on the system can use that device without writing a single line of code specific to it. Apps talk to Core Audio. Core Audio talks to the hardware. Neither side needs to know the details of the other.
This is also why a new audio interface or a pair of Bluetooth headphones generally works the moment you plug it in or pair it, with no separate app to install: as long as the hardware and its driver speak Core Audio, the entire system already knows what to do with it.
Where Audio MIDI Setup fits in
Audio MIDI Setup, the small utility buried in Applications > Utilities, is not a separate audio system running alongside Core Audio. It is a window onto it, exposing a handful of things Core Audio can do that Apple never built a friendlier interface for. The two device types it can create, Multi-Output Devices and Aggregate Devices, are both just new devices as far as Core Audio is concerned. A Multi-Output Device fans one signal out to several real speakers. An Aggregate Device combines several microphones into one with more input channels. Once either exists, Core Audio treats it exactly like any single piece of hardware, because from its point of view, that is what it is.
This also explains why both of those features share the same clock drift problem. Every physical device runs on its own internal clock, and Core Audio has to reconcile those small differences whenever it combines devices together, whether that is fanning audio out or pulling several inputs in.
Every feature free for 30 days. No account, no card details.
Download for MacWhere Chorus fits in
Chorus works entirely through Core Audio's own public device-grouping mechanism, the same one Audio MIDI Setup uses to build a Multi-Output Device. It is not a custom driver, and it does not install a kernel extension. That is a real, practical difference: there is no special setup step after installing Chorus, and no restart needed before it can group your speakers. It asks Core Audio to do the same kind of thing Audio MIDI Setup asks it to do, just with an interface built for picking speakers day to day rather than for the general-purpose device building Audio MIDI Setup is meant for.
The features that sit on top, one real volume for a group of speakers, per-app routing, keyboard shortcuts, auto-switching when a device connects, all live in Chorus's own interface. But the actual grouping underneath is Core Audio doing what it was built to do, the same layer that runs every other sound on your Mac.
For the full technical detail on the framework, see Apple's own Core Audio developer documentation.
Frequently asked questions
What is Core Audio on a Mac?
Core Audio is the framework built into macOS that handles every sound on the system, from playback and recording to how apps talk to speakers and microphones. It is not an app you open. It runs underneath every app that makes or hears sound, including the Sound settings pane itself.
Why does every app share the same output and input device?
Because Core Audio maintains one system-wide idea of the current output device and the current input device, and every app reads from that same setting rather than keeping its own. That is why switching the Mac's speaker in one place changes it everywhere at once.
Is Audio MIDI Setup part of Core Audio?
Audio MIDI Setup is a small utility that exposes some of what Core Audio can do, mainly creating Multi-Output Devices and Aggregate Devices. It is a window onto Core Audio, not a separate audio system, and everything it builds is a device Core Audio then treats like any other.
Does Chorus replace Core Audio or install its own driver?
No. Chorus works entirely through Core Audio's own public device-grouping mechanism, the same one Audio MIDI Setup uses, rather than a custom driver or kernel extension. That is why installing Chorus needs no special setup step and no restart.
Why does macOS not have a built-in per-app volume mixer?
Core Audio was designed around one shared output device rather than separate volume sessions per app, which is a design choice rather than an oversight. That is a deeper topic on its own, covered in the linked post on why the Mac has no native per-app volume.
