Can a computer-generated voice be sincere? A case study combining music and synthetic speech

Abstract

This article explores enhancing sincerity, honesty, or truthfulness in computer-generated synthetic speech by accompanying it with music. Sincerity is important if we are to respond positively to any voice, whether human or artificial. What is sincerity in the artificial disembodied voice? Studies in musical expression and performance may illuminate aspects of the ‘musically spoken’ or sung voice in rendering deeper levels of expression that may include sincerity. We consider one response to this notion in an especially composed melodrama (music accompanying a (synthetic) spoken voice) designed to convey sincerity.

Key words::

• Disembodied voice
• melodrama
• sincerity
• synthetic speech
• synthetic voice

Note

Declaration of interest: The authors report no declarations of interests.

Notes

1. In 2006 Barker took part in an exercise led by Clemens Wöllner with other musicians. Invited to compare two video recordings of the same piano music, they all discussed why they preferred a certain one over the other. After the discussion Clemens disclosed that the second video contained the same sound-track as the first, but was cleverly edited around a pianist who effectively mimed the performance. Wöllner's work was a development of research begun some years before by Behne.