There is one you tube video showing this process.
Basically, you record in WAV format, a number of notes (C E G A for example) from low octave to high octave(within the range of that instrument), one WAV file per note. Then use Grand Suite software to load all these , again one WAV per note. Grand Suite will generate a voice for that instrument. I plan to do this but still do not have time yet. (Recording WAV on my Genos for a dozen voices I like to use).