I know .4 is pretty standard, and I keep getting mixed reports of running .6... Some say it's great, other say it's a waste because sometimes there's no content. *shrugs
I likely wouldn't try to put .8 in anything but a larger dedicated HT. And even then, depending on the processor, I would probably give discrete Subwoofer channels the priority for the AVRs and AVPs that have those limited .
In terms of the soundfield and our ability to locate items within it (given an appropriate high quality 3D signal obviously) - I'm pretty sure that a well set up smallish 4.1.4 is all that is needed in terms of perception...
As the room and the screen grows, and the number of seats, as well as the distance between speakers is increased, you need to increase the number of speakers to maintain the illusion... but you also have to accurately process the soundfield for the number of speakers present and their location.
I'm not sure of the variables involved and their interaction - but there is definitely some sort of near field vs far field difference... which may therefore relate to reflections from around the room. ie when close to speakers - the direct sound completely dominates... when more distant, in many rooms the effect can be "muddied" by the reflections... so as we get more distant from the speakers, the reflections start to become a larger proportion of what is heard.... and moving to more speakers seems to help control this by providing more correct signals from the various directions (as opposed to the reflections which come from the wrong place AND at the wrong time...).
Now we have Dirac ART coming out and its associated paradigm shift... why do I say paradigm shift?
Well first of all the speakers all work cooperatively, so in the bass region where the sounds are either no localised or only semi localised, all the speakers in the room can contribute. Even if a speaker is 20db down in frequency response terms at that frequency, it can still provide a couple of db of input into the signal. So although most of the speakers have a -3db point at say 80Hz... they may well have useable signal much lower down... 40Hz or lower potentially.
Full range speakers, typically -3db at 30Hz or 40Hz - will easily go down below 20Hz in many cases.
So now you effectively have the performance of at least an additional subwoofer, distributed around the room... for many people and many systems, this may eliminate the need for a sub.
Second thing is, Dirac-ART is all about using noise cancelling style reverse phase signals to cancel some of the reflections that cause peaks and troughs within the listening room....
Active cancellation (or more accurately reduction in amplitude) of reflections.
Some of the same reflections that muddy the surround field (as mentioned above)
Yes DL-ART will allow for improved and far more even bass response through the room, and in its initial launch version, limited to 150Hz, that is its primary function.
But Dirac was also talking about raising the 150Hz limit, and moving to a 500Hz limit fairly soon... that really moves into the critical midrange - where many/most of the 3D surround effects live.
If you have active control of the reflections in this way, does this reduce the need for additional speakers? (particularly in more typical smaller spaces eg: single seat, or single shared couch)
Given cooperative audio courtesy of DL-ART, it seems likely that more than two subs would be overkill, and many systems would do just as well with a single sub (and full range L/R).
It also seems that we could be looking at optimal setups being (or starting from) 5.2.4, and anything beyond that might end up being overkill - additional expense and complexity for little or no discernible improvement.
With a proper "location" algorithm, it really isn't about how many channels are present, as the signal isn't in "channels" but in positions - and it is up to the decoder to then translate it into the number of channels present.
Even with recordings where the signal is defined in "channels" the positioning of the channels is based on standards, and where there are more or less speakers than the recorded channels have, the decoder then "positions" the signal where it needs to be...
(that's the theory anyway!)
So yeah - I am skeptical of the 16 channel + configurations.... unless you have a 3 row or larger theatre, there's just no point!