X-Fi CMSS-3D, Headphone surround, F.E.A.R. and Quak

Question

Hi there.
I'm trying to use the 'right' audio settings to play F.E.A.R. and Quake 4 using my new X-Fi Xtreme Music.
I'm a bit confused about when / when not to use CMSS-3D. Does it upmix a stereo mix to 'virtual' surround? Or does it enable OpenAL / DirectSound to produce further positional realism?
Should I turn it on when I'm playing with two speakers?
Should I turn it on when I'm playing with four speakers? (depends on how late/loud I can be!)
Most importantly, should I turn it on when I'm playing with headphones?
Or does it depend on the game itself?
I'm a bit confused!

Answer 1

The easy question first... The surround/stereo settings are only relevant for Doom3/Quake4 when using the original non-OpenAL sound engine. In this mode, it outputs DIRECTLY to 2 channels or 5 channels on the stereo and surround settings respecti'vely. As far as the soundcard is concerned, it looks like you are playing a 2 channel or 5 channel sound file - ie. there is no 3D audio API involved. Now depending on your speaker and CMSS3D settings, you COULD get your X-Fi to "process" this output. For example, if you have 4 or more speakers selected, and you enable CMSS3D-Surround, AND you set Doom3 output to stereo (non-OpenAL mode), the X-Fi will use its upmixing algorithms to deri've 4/5/6/7 channels. Similarly, if you have 4 speakers, 2 speakers, or headphones selected, and you enable CMSS3D-Virtual, AND you set Doom3 output to surround (non-OpenAL mode), the X-Fi will use its HRTF virtualization algorithms to create 5 virtual speakers.
Regarding what exactly CMSS3D is, the best explanation is in the Soundblaster Audio Technologies Glossary (under Audio Playback Technologies) in the Creative Knowledgebase. Unfortunately, they use the same acronym (CMSS) to apply to both upmixing, HRTF based virtualization of multi-channel speakers, AND HRTF based spatialzation of 3D audio (arbitrary 3D coordinates are provided with the sound). The knowledgebase article describes CMSS as it is known in the Li've! and Audigy series. One inaccuracy with the table shown in the article - the one that describes what CMSS does based on the number of source channels and the number of speaker channels - is that 4 speakers mode also uses HRTF to spatialize 3D audio (while 5 or more channels does not).
With the release of the X-Fi, the terminolgy is cleared up a bit. You'll notice that depending on the speakers you have selected, you'll see either CMSS3D-Surround, CMSS3D-Virtual, or CMSS3D-Headphone. CMSS3D-Surround is the upmixing stuff, CMSS3D-Headphone is HRTF virtualization/spatialization, and CMSS3D-Virtual is virtualization/spatialization WITH cross-talk cancellation use for 2 or 4 speaker setu
ps.
Notice that you get both CMSS3D-Virtual AND CMSS3D-Surround in 4 speaker mode, dependin on what the source material is (stereo or multi-channel). Finally, where the Audigy series had CMSS or CMSS2, where CMSS2 included simulated room acoustics, the X-Fi does simulated room acoustics when in Entertainment mode, but not Game mode. The room acoustics is good for movies and the effect is very similar to Dolby Headphone. In fact, I think the virtual acoustics sounds pretty decent for non-OpenAL/DirectSound3D games that output directly to multiple channels rather than going through 3D audio.
Any more questions?