But, per Hendrik's comment in the other thread, creating a silent channel in DSD is not straightforward and would have to be added to the DSP.
To be clear, its not a real problem technically to do it, but quite simply that DSD has no true "zero" (DSDs 1-bit format is actually -1/+1, to indicate the changes in the signal). So as a workaround, typically a inaudible ultra-high-frequency noise signal is being used - the one we have in MC is the most commonly used "0x69" signal, which translates to very low energy 352.9kHz noise line (AFAIK). This seems to be mostly an industry standard, so its probably fine, and DSD DACs should filter it out entirely.
Unfortunately all that channel handling is all in the PCM DSP code, but maybe Matt could be convinced to add a special case somewhere for 5.0 -> 5.1 DSD Bitstreaming.
PS:
As a final note, as mentioned in the other thread, "Bitstreaming" for DSD is not anything magic. It has no special meaning. We refer to outputting anything thats not PCM as "Bitstreaming", and DoP falls into that category. There is no special output mode involved in Bitstreaming DSD over PCM (in contrast to native DSD over ASIO), the only difference is that MC is told to not decode DSD to PCM - if it detects DSD in the first place, if you already packed it as DoP into a PCM container, then you are by-passing all of that anyway.
"Bitstreaming" in general does not pass the source file 1:1 to the output without any changes. There will always be changes that make it possible to transport the data over a PCM link in the first place. If you have a typical DSD file, this will include packing it in a DoP container - which means its not quite the same data that it used to be, but this "wrapping" is only on the outside, and the actual audio data on the inside remains untouched. DSD is in a special place as such that you basically bitstream every channel individually, so you can re-arrange the channel, for example, and still pass the audio data untouched to the source - only in a different order. Equally, you could drop channel, or add a new one, since DSD is not "compressed" but only a different raw representation, it adds more freedom.
(Since this seems to be brought up repeatedly - if you encode a DSD file as DoP and throw it into a PCM container like FLAC, then it is effectively just PCM to Media Center, it does not know that its DSD, it'll treat it as plain old PCM, so you need to turn off all DSP. As such, you also really cannot compare this behavior to playing native DSD files)