I found
viewtopic.php?f=12&t=49016#p284253 which indicates each frame contains a sample for each channel in sequence
However I don't find any information on the actual mapping of those physical samples to the output channels. This is indicated via the hdmi avi info frame as per cea-861-d, does the sdk apply such a mapping before it delivers the packet? Or is it just passing it through as is without exposing that metadata from the infoframe?