The process involves rotating the head. For example, the center is created by looking directly at the right speaker. For the most part I didn't move the speakers (except to create the height channels), rather the idea was to record impulses two speakers at a time with the speakers being at the angles the speakers would be in if they were actually there in a multi speaker room. Each of these "PRIR" pairs are then added to the virtual room until you have the full complement (in this case 16) to create a full virtual surround sound room. And, they do indeed really sound like my live speakers, except perhaps better b/c the dynamic limitations of a speaker like the LS 50 are transcended when something like an HD 800 or LCD X (which can play at 114 db) are programmed to emulate the LS50's sonic signature (transfer function). Obviously, the real LS50s could never approach that kind of volume, but the virtualized ones can--and with the distortion levels of the headphones (check out the LCD X's distortion profile in A,mir's review!) rather than the speakers.Nice, HD800s might be the classical choice as a basis. How's it compare to your sound localisation with your real speaker system? Am I right in thinking you're asking the software to create more speakers than you currently own in your real speaker system? If so, then how convincing is that.....how does it go about extrapolating the extra speakers? How many speakers you got in your real system?
EDIT: do you just measure your speaker at different angles (the angle that virtual speaker would actually be at in relation to you) as part of the setup process? So if it was rear right roof you'd just have to make sure your speaker was up there and you were down there when you were doing the setup calibration sweeps? Then the software combines it all?