Somewhat similar to this post: https://developer.apple.com/forums/thread/110816
I am attempting to use the VoiceProcessingIO and RemoteIO audio units simultaneously to provide full-duplex VoIP communication while also simultaneously recording just the high-quality mic input audio to file, which does not appear to be possible / supported currently on iOS.
It seems that using the VoiceProcessingIO unit to record and playback audio (IO enabled for both input and output) for both its echo-cancellation and automatic gain control features, while also using the RemoteIO unit to just record audio (IO enabled for just input) results in the samples recorded by the RemoteIO unit to also be "preprocessed" by the VoiceProcessingIO unit, in that the gain is boosted (via the AGC), and the frequency response is filtered, resulting in audio that "sounds like it was recorded over a Zoom call" and is not fit for musical instrument recording.
Attempting to use two RemoteIO unit instances works as expected for the recorded audio, and results in high-quality samples being recorded. It only seems to be when the VoIP functionality instead uses the VoiceProcessingIO unit that the RemoteIO unit no longer records the "raw" audio.
Is this the expected behavior, and therefore a known limitation on iOS, where the audio hardware inputs can only effectively be "connected" to one audio unit at a time?
I am attempting to use the VoiceProcessingIO and RemoteIO audio units simultaneously to provide full-duplex VoIP communication while also simultaneously recording just the high-quality mic input audio to file, which does not appear to be possible / supported currently on iOS.
It seems that using the VoiceProcessingIO unit to record and playback audio (IO enabled for both input and output) for both its echo-cancellation and automatic gain control features, while also using the RemoteIO unit to just record audio (IO enabled for just input) results in the samples recorded by the RemoteIO unit to also be "preprocessed" by the VoiceProcessingIO unit, in that the gain is boosted (via the AGC), and the frequency response is filtered, resulting in audio that "sounds like it was recorded over a Zoom call" and is not fit for musical instrument recording.
Attempting to use two RemoteIO unit instances works as expected for the recorded audio, and results in high-quality samples being recorded. It only seems to be when the VoIP functionality instead uses the VoiceProcessingIO unit that the RemoteIO unit no longer records the "raw" audio.
Is this the expected behavior, and therefore a known limitation on iOS, where the audio hardware inputs can only effectively be "connected" to one audio unit at a time?
