I found that the aggregated device correctly obtains input channels in the standard microphone mode. However, in voice isolation mode, it only retrieves channels from the first sub-device in the aggregated device's list. If I want to properly obtain channel information in voice isolation mode, how should I do it?