Here is a message forwarded from Fabio Bertolotti

----------------------------------------

----- Begin Included Message -----


     A sound that is dominated by a large initial ``spike''
     in amplitude, followed by a sustain portion with an
     identifyable pitch, should be hard to reproduce with
     adsyn + hetrodyne analysis, in my opinion.

     This is so because the spike (i.e. the attack portion
     created by a hammer striking a piano string) has a Fourier
     transform containing energy at many frequencies, without
     particular peaks at specific frequencies that could be 
     identified as ``fundamental'' and ``harmonic'' components
     of the sound.  Hetro + Adsyn work well when such particular peaks 
     are present. Furthermore, hetrodyne + adsyn neglect phase
     information that is important in the reproduction of
     spikes and rapid changes.  In the sustain part of the sound,
     phase is not important to the ear.

     In my opinion, it would be better to reproduce the piano
     sound using the addition of two signals:

         1) a wavetable synthesis for the brief, but strong, attack
            portion,

         2) adsyn + hetrodyne data of the remaining sustain + release.

     I have developed a hetrodyne analysis tool-kit that should facilitate
     with the attack-sustain-release partitioning of the hetrodyne data,
     plus plotting, editing & mixing + other steps.  I have sent the
     source + documentation + binaries fo Richard Boulanger, for the
     upcomming Csound book.  I hope this toolkit will help those using
     adsyn.

     Regards

     Fabio P. Bertolotti


----- End Included Message -----