OS X: How to pipe audio output through a processing algorithm?

What does one have to do to capture the output audio stream from an application (iTunes, QuckTime, VLC, whatever...) do some processing, maybe modify it and then send it back to the output? Any kind of help would开发者_如何转开发 be greatly appreciated because I don't even know where to look in the docs for this one...

Thanks

One approach you can try would be to use the SoundflowerBed source code as a basis for creating a virtual audio device. This can then be set as the default output device, in system preferences, so that all applications (other than those for which the output device has been set otherwise) will send audio to it.

Your processing program can capture the audio from a such a virtual audio device, process it, and pass it to an actual output device for playback in real-time.

It is possible to set/reset the default device programmatically, using the CoreAudio API, in order to provide a more integrated experience to your customers.