Sorry.. missed those posts from the 14th..
I was looking for a software solution also and I haven't found anything. I believe it is possible, but will (at the least) introduce some kind of delay.
It would most likely be a pre-processor, maybe even at a driver level. It would first have to capture the audio, filter out certian frequencies, then output to the correct channel (assuming different channels get different frequencies). So you could use your 'front' outputs for highs, 'rears' for lows, subwoofer for subwoofer, etc.
Basically, it doesn't seem feasible. A hardware soultion seems the only way right now unfortunately Maybe there is a way to design a computer controlled hardware solution like in a regular HU, who knows :P
Good work on the bezel. If you're 'not sure' if you should perfect it more, then I definitely would! :P I'd rather over perfect it then miss a few spots like I did (on the underside of the screen, but it still bugs me).