The only big win is to look at why you need to do this, and perhaps think of a better approach to the problem as a whole.
these inputs are in fact audio-streams which need to be added
together. There is no other way around that if you want
to process say 30 different streams through the same filter.
And I can't imagine it being more effective to apply a filter
on each stream instead, and for some algorithms that
is not even an option!. Ahh well, I'll keep tryin to find
a better way

thanks
/btq