What about the non identical subs being used to help smooth overall room response? Would this be preferable than using them collocated. If output could be matched, there still would be 3dB gain compared to 6dB collocated. In my cause with the 20-39, The gain would only be 1.2dB. So the only potential benefit I'd have would be smoothing the response.
Also, regarding the combining of distortion, are you sure it would be an issue? If using the gain (4dB in my case) from collocation to allow the HSU and/or SVS to not have to work as hard, wouldn't that lower overall distortion? In other words each configuration should have nearly the same overall output level:
HSU by it self, relative level (0dB)
HSU -7dB, SVS -5dB = +0.07804dB
HSU -4.5dB, SVS -8dB = -0.05429dB
HSU -2.5dB, SVS -12dB = +0.00940dB
And if I wanted to play louder than just the HSU by itself, I could do the following (yes with more distortion, but it does give more output):
HSU +0dB, SVS -5db = +4dB
So I can see the argument with the +4dB gain being used for max output (causing more distortion), but I'm not sure that there is more distortion when the gain is used to lower the level of BOTH subs. I would think there would be less.
Are you saying that this latter case would just be a wash and it would be better to just use one? In the former case, you have more output capability despite the extra distortion. That may still be useful for movies with lots of deep bass.