Oops, I forgot some parameters. The modulation frequencies should be in the range of voice vibrato to get the soprano-effect, but the segregation can be had with a much larger range. Steve