The Digital Voice Interface is a live analysis of voice which extracts individual features that control any parameter of electronic sound. These are achieved through a combination of signal analysis, signal processing and spectral analysis.
The voice is both the instrument and the interface for live-electronic sound which is, in turn, infused with “vocality”.This raises my question of vocality: is vocality more than the sum of its features, and if so, does the process of voice lend a greater vocality than the sound of the voice itself?
Between 2012 and 2014 I had the opportunity to study and collaborate with Pete Johnston and develop vocal feature extraction in Kyma as well as the indicated digital voice instruments.
The following digital voice instruments can be heard throughout my work, on their own or in combination with another, for brief moments or for entire compositions. These descriptions are intended for future collaborations with composers writing for my voice and computer or musicians improvising with myself and ensemble. The following signal flow diagram is a source for creating these instruments–combinations and juxtapositions of different signal processes serve as an index and navigational tool.
In all cases, the following instruments can be used to generate fixed media, live or automated compositional plans, or voice-controlled relationships to any parameter using vocal feature extraction.
This overview of voice-controlled electronics may be applied to any live variable in the subsequent instruments. Live features of the voice which have been implemented successfully in Kyma are: frequency, amplitude, brightness, vibrato rate, vibrato depth, breathiness, growliness, scream and formants. They have all been implemented so that they are differentiated from each other and can be controlled independently. The following diagram depicts the range of compositional relationships these features can have on a sound over time: synchronous voice-controlled sound (a 1:1 relationship), asynchronous or offset durations and mappings (offsetting the relationship of the voice to the sound), to time-warped relationships that include impulse response compositions or simply using the voice as influential means of controlling sound (the voice input is far removed from the sonic output) .
This instrument investigates the musical potential between amplified and resynthesized voice. The final variable bypass allows me to crossfade between the two, allowing the composer to exploit the varying degrees of these differences.
Choosing the signal path of the resynthesized voice, the variables that have been implemented are to freeze the signal with a variable number of oscillators, to add an envelope to that frozen state, to pitch shift the original input to either extreme of the audio range before resynthesis, to derange the harmonic spacing of the partials, and to isolate any single or group of the 64 partials.
Essentially, this voice instrument is capable of moving seamlessly from the amplified voice and its breath dependancy to resynthesized voice that, with a programmed envelope, can seemingly sing without breath. On the other hand, this instrument can gradually change from a remarkably recognizable voice to a continuous atonal instrument.
One example of using this instrument in an unusual way is to pitch shift the input to a frequency impossible for the voice to sing and freeze the signal at a moment of breathiness in the voice. The bank of oscillators will seek frequency data within the breathy sound and result in pitch clusters. An entire composition could be made from breath, pitch shifting and the varying number of oscillators that are interpreting the breath input.