Skip to content

Sample code for recording system audio on macOS 14.4+

License

Notifications You must be signed in to change notification settings

insidegui/AudioCap

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

AudioCap

With macOS 14.4, Apple introduced new API in CoreAudio that allows any app to capture audio from other apps or the entire system, as long as the user has given the app permission to do so.

Unfortunately this new API is poorly documented and the nature of CoreAudio makes it really hard to figure out exactly how to set things up so that your app can use this new functionality.

This project is provided as documentation for this new API to help developers of audio apps.

AudioCap.mp4

API Description

Here’s a brief summary of the new API added in macOS 14.4 and how to put everything together.

Permission

As you’d expect, recording audio from other apps or the entire system requires a permission prompt.

The message for this prompt is defined by adding the NSAudioCaptureUsageDescription key to the app’s Info.plist. This key is not listed in the Xcode dropdown, you have to enter it manually.

There’s no public API to request audio recording permission or to check if the app has that permission. This project implements permission check/request using private API from the TCC framework, but there is a build-time flag to disable private API usage, in which case the permission will be requested the first time audio recording is started in the app.

Process Tap Setup

Assuming the app has audio recording permission, setting up and recording audio from other apps can be done by performing the following steps:

  • Get the PID of the process you wish to capture
  • Use kAudioHardwarePropertyTranslatePIDToProcessObject to translate the PID into an AudioObjectID
  • Create a CATapDescription for the object ID above, and set (or just get) its uuid property, which will be needed later
  • Call AudioHardwareCreateProcessTap with the tap description to create the tap, which gets its own AudioObjectID
  • Create a dictionary for your aggregate device that includes [kAudioSubTapUIDKey: <your tap description uuid string>] in its kAudioAggregateDeviceTapListKey (you probably want to configure other things, such as setting kAudioAggregateDeviceIsPrivateKey to true so that it doesn’t show up globally)
  • Call AudioHardwareCreateAggregateDevice with the dictionary above
  • Read kAudioTapPropertyFormat from the process tap to get its AudioStreamBasicDescription, then create an AVAudioFormat matching the description, this will be needed later
  • Create an AVAudioFile for writing with your desired settings
  • Call AudioDeviceCreateIOProcIDWithBlock to set up a callback for your aggregate device
  • Inside the callback, create an AVAudioPCMBuffer passing in your format; you can use bufferListNoCopy with nil deallocator then just call write(from:) on your audio file, passing in the buffer
  • Call AudioDeviceStart with the aggregate device and IO proc ID
  • Remember to call all your Audio...Stop and Audio...Destroy cleanup functions
  • Let the AVAudioFile deinit to close it
  • Now you have an audio file with a recording from the system or app

Thanks to @WFT for helping me with this project.

About

Sample code for recording system audio on macOS 14.4+

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages