My dealer says M told him that what I see is intended behavior. No reason to do split decoding with an 818. And yet, MQA has published this diagram:
So this contradicts what people tell me here and in the Roon forum.
Sigh. Would be good to get some authoritative answers.
Btw, it is of course correct that with an 818 there is no reason to split the processing. But there is the room correction issue. I use Acourate to measure and calculate and Roon to run the convolution, which breaks MQA. And there is no question that REQ is more valuable than MQA. So I hope, without evidence, that Roon could implement first stage decode and DSP while allowing 818 final decode.