Parametric SpectrumizationSkip other details (including permanent urls, DOI, citation information)
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License. Please contact firstname.lastname@example.org to use this work in a way not covered by the license. :
For more information, read Michigan Publishing's access and usage policy.
Page 127 ï~~PARAMETRIC SPECTRUMIZATION NIL PARENT, ELECTRONIC MUSIC STUDIO, LAVAL UNIVERSITY An analysis/synthesis systemI with 1024 digital oscillators has been developed. Each oscillator is programmable in terms of waveform, amplitude and frequency (integer, beating, ambitus) and gives access to an amplitude envelope and a frequency envelope. All the synthesis features are accessible in real time thanks to hardware implementation taking the form of 2 custom-designed2 chips per group of 32 oscillators. The analysis facility whose task is to "oscillatorize" a given audio signal calculates all the multi-stage amplitude and frequency envelopes needed for the resynthesis of the signal. As a result, four different spectra are generated, the Amplitude spectrum, the Integer spectrum, the Beating spectrum and the Ambitus spectrum. For each spectrum a second "different twin" spectrum can be entered. Since all the spectra can be modified in real time, multi-spectral interpolations can be controlled with any MIDI source. This approach is called "Parametric Spectrumization". 1 The ACXEL resynthesizer. 2. By Pierre Guillemette, inventor of the "Harmonic Processor". Starting from the hypothesis that the tool modulates the work, the development of a new instrumentation is a prerequisite to that of a musical language. It is indeed through a daily instrumental intimacy that the most vivid and durable roots of the musical language are grown. In the same manner, one could hardly expect the development of new and coherent musical articulations if the instrument itself does not maintain with the sound material it is articulating a very accomplice relation. For more than 40 years, most research in electronic music devoted to the study of sound articulation has considered sound as a whole entity, as something seen from the outside. The basic idea has been that a sound is a waveform to which various instrumental transformation and articulation techniques can eventually be appplied. "With sampling, on the other hand, you've got the sound as a whole inside a more or less sealed box, and it's very difficult to get inside the box, and manipulate specific aspects of the timbre, such as the amplitude and pitch envelopes of individual overtones. The dimensions of control that you have over a sampled sound - lowpass filtering, looping, and so on - are relatively coarse and crude compared to inherent in synthesis techniques"1 the subtlety One escape from the classical techniques is proposed here. From the notion of synthesis as "the composition or combination of parts or elements so as to form a whole" (Webster's New Collegiate Dictonary), where "each element is a mirror of the entire universe" (Leibniz, in 1714) one can affirm that a sound, that any sound, is, in conformity with the additive synthesis principle, the result of a combination of simple elements. As a result, if the values of a large number of simple elements can be changed in real time on a parameter per parameter basis, or eventually on a multiple-parameter basis, new sonic articulation techniques can be experimented which allow at last to penetrate into the interior of the sound, as if the composition process was getting to the level of the sound itself. The coordinated control of a given number of oscillators gives birth to a series of new articulation techniques, namely "Ampizing", "Intizing", "Beatizing" and "Ambizing". Each technique has, depending on the reference settings, its own musical impact or shall we say, 1 Wendy Carlos in Keyboard, 1985 ICMC 127
Page 128 ï~~as a salute to Pierre Schaeffer, its own "musical dominance". AMPIZING To "ampize" is to control in real time the dynamic interpolation between two amplitude spectra. The musical dominance here has mainly to do with what is usually called the "color" of the sound, but not necessarily in a "filtering" perspective. In fact, with respect to filtering, ampization represents an unprecedented suppleness. SOUND EXAMPLES - PART I INTIZING To "intize" is to control in real time the dynamic interpolation between two "integer" spectra, the reference frequency of each oscillator being programmable in terms of "integer" position, variable anywhere between harmonic functions 1 and 255. The musical dominance here is mainly "melodic" since it implies a continuous variation of the frequencies of the harmonic components of a given sound. The melodic pattern can be set in a up-going fashion or in a down-going fashion independently for each oscillator. SOUND EXAMPLES - PART 2 BEATZING To "beatize" is to control in real time the dynamic interpolation between two "beating" spectra, the reference frequency of each oscillator being programmable in terms of "beating" position, the synthesis system providing a four digit decimal resolution. The musical dominance here mainly has to do with the inharmonicity of the sound material, like some kind of internal detunability. SOUND EXAMPLES - PART 3 AMBIZING To "ambize" is to control in real time the dynamic interpolation between two "ambitus" spectra, the ambitus of each frequency envelope of each oscillator being independently programmable anywhere between.01 of an harmonic and 64 harmonics. It should be pointed out here that the impact of the ambitus is directly related to the harmonic function that a given oscillator is fulfilling. Let us suppose for example that a given ambitus with a size of + or - 1.0 harmonic is applied to all the harmonic components of a complex layering of oscillators. If the integer value of a given oscillator is set to 2.0000, then an ambitus of 1.0 will vary the frequency of that given oscillator any where between 1.0000 (an octave lower) and 3.0000 (a fifth higher); if, at the same time, the integer value of another given oscillator is set to 27.0000, then the frequency of that given oscillator will vary anywhere between 28.0000 (a microtone higher) and 26.0000 (a microtone lower). The musical dominance here is difficult to identify and label because the effects vary enormously depending mainly on the size of the ambitus setting, the "region" of the harmonic structure onto which it is applied and also on the internal calculated structure of the sound, especially for resynthesized sounds. SOUND EXAMPLES - PART 4 NUANCES With respect to the problematics of musical articulation, the parametric spectrumization technique represents a step forward. The atomization of the sound into its most simple elements opens the way to the exploration of the sonic objects from the interior. As a result, because the synthesis system give access to all facets of the sound's interior in real time, it becomes possible to develop a new approach in the field of instrumental articulation, an approach whose foundation is situated within the sound it self. In many areas where technology and music are related, musical nuances shall take an increasing importance, provided that higher level (compound spectrumization) articulation facilities are proposed in order to revalue and enhance "live" music. ICMC 128-129-130