VoiceWorks: Harmony-Style parameters explained


What does the HARMONY STYLE parameters located in the HARMONY menu mean?


The Hmny Style is the humanization style that is applied to the harmony voices.  For a good understanding of what it is, some knowledge of a few humanization concepts is required:


This is when a singer glides either up to or down to the desired note.  The VoiceWorks can scoop notes either up or down to the target note.  To do this the VoiceWorks detects the note onset based on the pitch and whether the input is voiced or un-voiced.  Styles with scoop randomly select which onsets to apply the scoop, so that the harmony voices sound more like real people are singing.

Time Randomization:

Time Randomization is where the harmony voices are randomly sped up and slowed down. Applying this to the harmony voices simulates how real people sing together, since it is impossible for people to be perfectly in sync.  The VoiceWorks has a maximum time randomization of 100ms.  When using time randomization it is important to understand that the harmony voices will be delayed overall, as compared to when no time randomization is applied.  This is because we can only delay a signal to speed it up and slow it down.  Turning the Delay Compensation parameter under the Setup Menu to ON will delay the lead voice so that the harmony voices are equally ahead and behind the lead voice.

Pitch Randomization:

The research group analysed the vocal pitch of real singers to model Pitch randomization in the VoiceWorks.  The pitch randomly goes above and below the target pitch.

Styles are a combination of the above concepts. Here is a list a brief description of the Styles:

  • ScpDwn: Scoop Down to the note.
  • ScpUp:  Scoop Up to the note.
  • ScpTim: Random Scoop up and down and Time Randomization.
  • Loose:  Random Scoop up and down, Timing and Pitch randomization.
  • Timing:  Time Randomization.
  • TimPit: Time and Pitch Randomization.
  • Pitch:  Pitch Randomization.
  • Detune: Static detune of the harmony voices. Maximum detune amounts are: Voice 1: -25 cents; Voice 2: +25 cents; Voice 3: -19 cents; voice 4: +19 cents.

Share this page