well, as charles suggested, set up a few different reverbs of varying decay lengths, and send varying percentages of each grain to the different reverbs.
also, reverb usually has a 'dampening' setting, which will dull the signal as it goes through, so another thing you can try is to mix in a lowpass or bandpass filter after every voice before it goes in to the reverb. if you alter the dry/wet ratio of each voice going into the filter, as well as altering the cutoff and q values for each filter it will make it sound similar to if each voice were going through a different reverb.