I have the same problem ,did you sovle it finily?

A virtual device refers to the use of AudioServerPlugIn to create a nullAudio, similar to the open-source BlackHole. As you mentioned: for synchronized audio, the CoreAudio API AudioHardwareCreateAggregateDevice is used to create an aggregate device. I did this, and it was successful. However, when I switch the microphone mode to the voice enhancement mode, the number of input channels I retrieve from the aggregate device is incorrect.

The specific issue can be seen in the following two help posts: https://stackoverflow.com/questions/79319480/mac-audio-software-aggregate-device-issue https://developer.apple.com/forums/thread/771690 Therefore, I had to abandon the aggregate device and use separate devices to capture and forward the audio. But this led to synchronization issues

"I don't know how to find XX_NECESSARY_INFO_1 and XX_NECESSARY_INFO_2." After testing, the number of input channels from the aggregate device is normal when the microphone mode is set to Standard. However, in Voice Enhancement mode, I can only retrieve the number of input channels from the main device of the aggregate device, not from the secondary devices. Is my code for creating the aggregate device incorrect? Did I not set it to be compatible with Voice Enhancement mode?

Make corrections：After testing, I found that in voiceIsolation mode, the output channel of the aggregate device is obtained normally. The number of input channels retrieved matches the input channel count of the first device in the kAudioAggregateDeviceSubDeviceListKey. This means that in speech enhancement mode, only the channels of the first device in the list can be obtained.