Singers with nice voices have a natural vibrato. For some people, this vibrato starts the instant they start a note. For others, the note begins without vibrato, but it's added after a short delay. The frequency of vibrato also varies. Dolly Parton's, for instance, is almost irritatingly shaky; vibratos faster than 7 cycles per second sound nervous. Frank Sinatra's vibrato, on the other hand, is subtle and mellow, with fewer cycles per second. A well-developed vibrato undulates at a rate between 4 and 7 Hz.
Regardless of the style, this vibrato is a trait that the singer is born with. It cannot be acquired or learned. Many individuals have very nice voices and can hold a note quite well, but they lack that nice vibrato, which becomes glaringly apparent when they sustain a note for an extended period of time.
Is it possible to design an effects device that would add this natural vibrato to a voice that doesn't have it? It seems very straightforward from an engineering perspective, given today's DSP technology. You could easily obtain a singer's vibrato "fingerprint" using a computer, then apply it to a vocal input. To enhance the effect's "naturalness," the vibrato would have to come on after a note is sustained longer than, say, one second. Such an effects device would certainly have a huge market.Glenn Johnsonvia e-mail
Glenn - Vibrato is a slight, periodic variation in pitch. Therefore, in order to add vibrato to a static singing voice, you would need to send the vocal signal through a pitch shifter that was set to shift the pitch up and down periodically over time. As you suggest, this is easy to do with today's DSP technology; all you have to do is apply the pitch shifter's low-frequency oscillator (LFO) to the pitch, specifying the frequency (typically 4 to 7 Hz), depth (amount of pitch shift from nominal), and onset delay (typically one second), if available.
The waveform of the LFO (and thus, the vibrato) would traditionally be a sine wave; however, many software synths, samplers, and sound-design systems (such as Native Instruments Reaktor, Applied Acoustics Systems Tassman, Synoptic Virtual Waves, and Symbolic Sound's Kyma sound-design workstation) let you apply a custom waveform, such as that derived from the analysis of an actual singer's vibrato, to the LFO. Sophisticated vocal processors are available from TC Electronics, Eventide, and DigiTech, but they can't truly duplicate the "natural" sound of a well-trained singing voice.
EM author and vocal coach Joanna Cazden disagrees that vibrato is innate and not amenable to training. According to Cazden, it's difficult to teach and learn, but not impossible. Your observation that the degree of vibrato varies from one style of music to another supports the notion that it is at least partially a learned behavior, a "dialect" of singing that is no more innate or genetic than a regional speech accent. Furthermore, the ability to control vibrato - that is, to start a note "straight" and then add vibrato later - is definitely a learned skill.
Cazden does not recommend using an effects device with a fixed delay. Musically, you would do better with a device that lets you dynamically adjust the vibrato frequency, depth, and onset delay to shape longer or shorter phrases in a more or less dramatic fashion. However, this would be difficult if not impossible to coordinate smoothly in real time with footpedals, MIDI faders, or other controllers during a performance. On the other hand, if you use software to create the vibrato, you could add the effect offline and could thus tweak it to taste. Even then, it would probably be exacting work.
So although you might be able to get the job done with sound-design software and a sequencer, the best tool for this job remains the human larynx.