Configure the number of speakers in Windows first and make sure you can play a test tone to each in Windows Sound Control Panel.
Otherwise most consumer-level sound card drivers still do mixing to the number of channels you've selected in Windows, even with WASAPI Exclusive.
Really? With a HDMI port? Because this most certainly does NOT happen with my AMD card...
This seems a bit fishy. What
consumer level sound cards are even out there that do HDMI out? I know there were a few very expensive ones, but certainly most of them are now provided by the GPU vendors, right? This one certainly is. Maybe you meant a more generalized "sound device drivers", certainly, but I don't know that your overall statement is correct.
I really thought WASAPI Exclusive meant exclusive. That's the way it seems to work here... I think you might be thinking about analog outs. Even then I'd be surprised, but the sound card drivers from Creative and their ilk never cease to amaze me with their cruddiness, so who knows there.
For AMD cards anyway (I've never used an nVidia to bitstream but I bet they're the same) it is
often convenient to keep Windows set to 2-channel stereo while running a HDMI output. That's because Windows sends the all 6 channels of audio for all system output then,
even if the channels are blank. For stereo sound, like for most applications (Flash in a web browser, for example), this means your receiver gets a "multi-in" PCM 6 channel stream, with 4 blank streams, and can't expand the stereo to surround. Only applications that specifically output surround give you surround with Windows set that way (like games and MC). That would be fine if you never fired up Chrome or Firefox to watch an episode of The Daily Show or a stupid video on YouTube, but... Come on.
This setup
is annoying in its own way, because you need to remember to switch it back to 5.1 if you do actually want to run a game that has a real multichannel out, or else you'll be stuck there with stereo expansion. But, there is a way to make a hotkey to quickly switch between the two setups, so ... Tolerable if you use the Home Theater PC for web video (or other "dumb" sound applications) more often than gaming.
But, it seems to work with MC beautifully.
I have:
An AMD Radeon 6870 connected to a current gen Denon AVR
Options -> Audio -> Audio Output mode: WASAPI (with Exclusive ticked in the dialog, and buffering set to the default)
DSP -> Output Format -> Channels set to 5.1, mixing to JRSS (also 24-bit and resampling only when necessary set to 48kHz)
This gives me very clear (and nice) surround mixing for music and audio playback through MC. I have another zone defined with no upmixing for when I really just want the original stereo, but I don't use it much. When the baby/wife isn't sleeping, I can run a test if you want me to, but everything seems to indicate that it is working correctly, and it certainly sounds surround, with seemingly different things on the rear channels and cool stuff like that.
For video, I have:
Options -> Video -> Bitstreaming set to Yes (HDMI)
This enables passthrough of Dolby Digital, DTS, TrueHD, DTS-MA and their ilk. I do this mostly because I used the Room Correction features on my AVR, which were easy, so I don't need Room Correction from MC, and don't want to have to configure the decoding (and all the decoding filter copying to Sys32 hoopla). Basically, I'm lazy and that Just Works. If I ever feel I really need the VideoClock feature, I might change my tune and turn bitstreaming off. So far, I've been happy.
By default, with it this way, JRSS upmixes stereo audio and sends it out in a nice high-quality format, and all the videos that use "bitstreamable audio codecs" bitstream through. Note that JRSS will also upmix any stereo audio video you play, but that's fine. It does a good job. I've been using it down in my basement with the new stereo exclusively, and it sounds good. If you
really want your receiver to upmix audio for video, but let MC do it for music, you can do this too. Switch
Options -> Video -> Audio Playback Device and point it directly at your sound device instead of "same as audio". This means that MC essentially sends your video sound through the normal Windows output, so it is stereo only (the 2-channel Windows control panel setting mentioned above), and then your receiver will do the DTS or ProLogic thing on it. (You might be able to accomplish the same thing with "same as audio" still enabled and different zones in the DSP, but I've never tried, and you'd have to manually switch zones then.)
I'm dubious about which is better, but I'm starting to lean towards MC's default way (same as audio) and just using JRSS. I used to do it the way just described, but I've been trying it out for a while now with JRSS, and I think I'm keeping it.