FS1R Programming Human Voices


By Thor Zollinger, writer/musician/engineer at large.

I've been working on the Yamaha FS1R synthesizer, creating female voice patches on it for comparison to the PLG100-SG card which I also have.  I've created a particularly nice patch based on the lower register of Christina Aguilera's voice.  You can listen to it below where I used it in a classic Jazz piece called 'Misty'.  The singing voice is the FS1R, the other instruments are played on a Yamaha Motif-XF keyboard.

  FS1R Christina1 Voice Patch - Misty MP3 

I first learned how the patches were constructed on the Yamaha PLG100-SG card, then I created a life-like FS1R patch using what I learned.  I was having trouble before I bought the SG card, I couldn't quite get the patch to sound right.  The key seems to be in all of the other characteristics of the patch, not the timbre as much.  To make the patch sound more human it requires a large amount of key triggered portamento, correctly shaped sound envelopes, vibrato, reverb, and then the timbre has to be correct.  

Yes, I know Vocaloid and Utau are both superior tools to mimic a singer with your computer, but both of those tools require the singer sit down with you and record over 200 voice samples in order for it to work.  My chances of getting a famous singer to sit down with me and record are about nil.  Using the FS1R I can create a simple voice patch from just one clean recording, creating a nice backup singer in about an hour.  Pretty cool, eh?

The construction of the patch is described in the back of my FM Synth Programming Guide in a prior article.  You can get the Programming Guide here.

  FM Synth Programming Check out this article on my FM Synth Programming Guide