This past week I’ve had the good fortune of getting to work with some of the most brilliant minds in vocal pedagogy, including Dr. Johan Sundberg, one of the world’s leading vocal research scientists. One of my favorite moments of the week was when Dr. Sundberg told us the story about how he created a computerized voice that got into a doctoral vocal program.
Dr. Sundberg demonstrated how he could use VoceVista, a program that does real-time spectral analysis on a singer’s voice, to determine where the formant frequencies in a sustained note lay. He’d then use another program to plug in the pitch the singer was using, along with the measured formant frequencies, in order to reproduce the sound of her voice mechanically!
One of the voice teachers in attendence asked Dr. Sundberg whether he’s ever reproduced a whole song that way. He replied that, yes, he has, and in fact created a mechanical voice that was so well done and appeared to have such good technique that the voice was accepted into a doctoral program!
This type of spectral analysis, of course, has huge implications for the music industry. Imagine a recording of us in which our voices no longer need to be even close to in tune, because the difficult notes can literally be created on a computer using formants that imitate the sound of our own voices.
We’ll finally all be able to get that high G out in “Skyscraper” with a voice tuned to sound just like ours!