#+TITLE: CoreEQ #+SUBTITLE: System-wide equalizer for macOS
CoreEQ equalizes everything you hear — music, video, calls, games — whatever application is playing, with no virtual audio driver and no kernel extension. It runs on Apple's native Core Audio process-tap API, so there is nothing installed into the system and nothing left behind when CoreEQ is not running.
[[./design/screenshots/core-eq-1.8.png]]
- Features
A global process tap captures the mixed output of every application, equalizes it, and plays it back through the current output device. No =/Library/Audio= component, no reboot, no per-application setup. Quit CoreEQ and system audio returns to its normal path immediately.
** One filter chain, two ways to edit it
The Graphic editor is the professional ISO octave ladder — 32, 64, 125, 250, 500 Hz and 1, 2, 4, 8, 16, 20 kHz — at fixed centre frequencies, each adjustable ±12 dB. The Parametric editor adds up to sixteen bands of your own anywhere between 20 Hz and 20 kHz, each a bell, shelf, or pass filter with its own frequency, gain, and Q.
The two editors are tabs on the editing area's own border, and both edit the same chain and the same graph: the sliders are eleven filters that happen to be locked to the ladder, and an added band is one that is not. Nothing is converted, approximated, or reconciled between the two views, so a change made in either is exact and reversible.
** Live frequency response, directly editable
The window plots the true combined magnitude response of the chain, computed from the same coefficients the audio path renders with — not an interpolated approximation. Drag a handle to change a band, double-click one to reset it, or double-click empty space to add a parametric band exactly where you clicked.
** Real-time spectrum analyzer
A Hann-windowed FFT of the audio actually reaching your speakers is drawn behind the response curve on the same frequency axis, with fast-attack / slow-release smoothing, so you can read the programme material against your shaping.
** Presets
Twenty-two built-in profiles — the classic set familiar from Apple Music and Spotify — plus your own, managed from the sidebar. Built-ins are immutable; your presets can be renamed, duplicated, edited, and deleted. Switching is instantaneous and glitch-free.
** Automatic gain
The preamp computes itself, and does so by default. Boosting makes a preset louder, and louder always sounds better — which would make every comparison a loudness test rather than a tone test. With Auto on, the trim follows the chain and holds the preset at the loudness it started from, so what you hear when you compare two presets is the shaping.
It is measured as the average lift across the audible range, weighted equally per octave. A preset that lifts one narrow band is corrected gently; one that lifts everything is corrected hard. Nineteen of the twenty-two built-ins lift the chain on average, so most arrive with a small trim already applied.
The correction only ever attenuates. Cancelling a cut would mean raising every frequency the preset deliberately left alone, which is a way to make something clip that would not have clipped otherwise — so a preset that mostly cuts is left quieter, as it was meant to be. Switch Auto off, below the preamp slider, to set the trim by hand.
** A/B comparison
Two working states of one preset, switched from the header. Each holds a complete sound — the preset, the chain, the trim, and the tone — so A/B compares two versions rather than undoing one, and the pair belongs to the output device like everything else you have not saved. Reaching for the other slot the first time changes nothing: it starts as a copy of what you were hearing, and the first difference is one you made.
** Per-device settings
The preset in use, any unsaved edits to it, and the preamp follow the *output device*, the way volume does on macOS. Your headphone curve comes back when you plug in headphones and gives way to the speaker curve when you unplug them. The preset library itself is shared across devices.
** Menu bar control
CoreEQ runs as a menu bar application with no Dock icon. The status menu carries the master switch, output device, the preset list, a live response preview, and three Quick EQ tone controls (bass, mid, treble) layered over the active preset — all without opening a window.
** Built to stay running
The engine rebuilds itself when the default output device changes, when the sample rate changes, and after the machine wakes from sleep, with bounded automatic retry on failure. Bypass is a real-time flag rather than a teardown, so comparing processed and original sound is instantaneous. The render path never allocates or blocks: parameter updates are staged behind an unfair lock and picked up only when uncontended, and gain changes are smoothed over ~50 ms so preset switches and slider moves never click or step.
- Requirements
- macOS 14.2 or later
- Permission to record system audio, requested on first launch
- Installation
#+begin_src sh brew tap andreypudov/core-eq brew trust andreypudov/core-eq brew install --cask core-eq #+end_src
Upgrade later with =brew update && brew upgrade --cask core-eq=.
CoreEQ is currently unsigned and unnotarized, so macOS may block first launch. Allow it in System Settings → Privacy & Security, or clear the quarantine flag:
#+begin_src sh xattr -dr com.apple.quarantine /Applications/CoreEQ.app #+end_src
** Build from source
Requires Xcode 16 or later. Locally built applications are not quarantined, so CoreEQ launches with no Gatekeeper warnings:
#+begin_src sh git clone https://github.com/AndreyPudov/core-eq.git cd core-eq make install #+end_src
This builds a Release copy and installs it to =/Applications=.
** Prebuilt binary
Download the latest =CoreEQ-x.y.zip= from the Releases page, unzip, and move =CoreEQ.app= to =/Applications=, then clear the quarantine flag as above.
- First launch
- Launch CoreEQ. A sliders icon appears in the menu bar; there is no Dock icon.
- Grant System Audio Recording when macOS asks — CoreEQ needs it to process
- Choose a preset. The sound changes immediately.
- Using CoreEQ
Settings — a gear in the window's header, or Settings… in the menu. Two tabs: General, which holds Open CoreEQ at login and the status of the System Audio Recording permission, and About. Clicking the CoreEQ mark at the top of the sidebar opens About directly.
Window — presets in the sidebar, with a search field and your own presets above the built-ins; the response graph, the Graphic / Parametric editors, and the preamp on the right. The output device sits in the centre of the header and names the preset playing on it underneath, with a dot when that preset has unsaved changes; where the machine has more than one output, that control is also how you choose between them. Revert sits beside it — next to the preset it would discard — and the gear and the master switch close the row.
CoreEQ has no volume control. Its one level control is the preamp, which is the same thing a hardware equalizer calls Level: makeup gain for what the curve took away. Volume belongs to macOS, which offers it in the menu bar, on the keyboard, and in Sound settings.
Right-click a preset for New Preset, Rename, Duplicate, Save Changes, Reset to Preset, and Delete. Right-click a band slider for *Edit as Filter…*, which moves that band into the parametric editor without changing the sound.
Editing the active preset marks it Edited; Save Changes writes the values into one of your own presets and Revert discards them. Your edits, the active preset, and the on/off state are restored on every launch, per output device.
- Limitations
- Aggregate and Multi-Output Devices are not supported. Core Audio will not
- On a device CoreEQ cannot capture stream by stream it falls back to a stereo
- Presets cannot yet be imported or exported as files.
- Development
#+begin_src sh make build # build the Release configuration make test # run the unit test suite make release # build and package a distributable zip #+end_src
- License