Not only would it work to give the sound more character but it also allows for more effective memory use.
The only problem that could arise would be whether or not the effects have to work at 44.1.
The goal would be top sample at a rate say 32k, process and 32k then output at 32k.
You wouldn't want it to be sampled at a 32k interpetation process at 44.1 output at a 32k interpetation.
If you take something like one of the Roland SP's (202/303) notice how the available memory changes based on the sampling mode, that's what you are looking more along the lines of.
The question is can JJ pull it off?
I guess we won't know if we don't ask.
