Topophony, no. 1
Topophony, no. 1 is a work of live electronics. Its name is an adaptation of terms like "monophony" and "polyphony". These are musical terms whose etymology goes back to the Greek language. "Topo" is from "topos", which is Greek for "place". "Phony" is from "phonos", which is Greek for sound. Together they are "topophony", which essentially means "place sounds". Or, soundscape. I chose the term "topophony" because to me it suggests an experience positioned with the confluence of multiple environments, rather than the observation of just one.
Musically speaking Topophony, no. 1 unfolds in four phases, or four environments if you will. Elements from each appear in or persist through the others. Phase 1 is called Sequence and is reminiscent of classic sequenced synthesizer materials. Phase 2 is called Wind and Starlight, named so because of the quality of the noise synthesis and its manipulations. Phase 3 is called Groove and Glitch, named for its rhythmic and glitching activities. Phase 4 is called Twilight, and it develops on some of the sounds appearing in Phase 2.
The patcher used to create this piece is an engine for the production of a great variety of sounds from noise generating objects. It's a further development upon work I've been doing with my Noise Sculptor patches. You can visit its page here on my website to learn more about it.
​
The engine is built around a set of Noise Sculptors that produce synthy tones and a drum machine from a single variable source. The Noise Sculptors are informed and controlled by note dictionaries (manually programmed with unique mathematical values) and rhythm generators (driven by transport-synced phasors that make rhythm selections probabilistically). The sounds produced by the synth voices and drum machine are in turn further manipulated by other algorithms.
​
The noise sources available include white and pink noise generators, as well as a playback object that can load any audio file. The last source is the one used for Topophony, no. 1, which is a noise generation object called jeynoise.
Jeynoise is an object at the center of a max package developed by Johan van Kreij, Sohrab Motabar, and Jeyong Jung called ACToolkit. You can see it in the image included here (joynoise's help patch), highlighted in green.
​
Jeynoise is a sinusoid and saw wave generator that performs period-by-period synthesis. What this means is that it generates both sinusoid and saw waves at variable frequencies that change after a prescribed number of periods or vibrations. It can change the frequency after 100 periods, or even after 0.01 periods. Faster rates of change produce noisier signals. So, varying the calibrations of its settings can produce a vast array of differing sounds. It really is an incredible object.
​
Another thing I want to point out is that the variation of the frequency selected is determined by probability distribution functions. The multislider object on the left of the Jeynoise help patch determines the probability for the pseudo-random numbers (nodes within a specified frequency range) to be selected.
​
The object also has a plethora of other settings that I won't get into here. They are very much worth understanding though!
I have jeynoise objects in a subpatcher, shown in this image here. They are driven by special pseudo-random number generator objects called jeyrand. These produce values through one or a combination of two processes: 1, they can be controlled by an external signal; or 2, they can produce numbers according to an internal distribution function (gaussian in this case). Which is to say, I could have the external signal drive 70% of the randomization, and the internal function the other 30%. Or it could be 50/50, or 10/90, and so on. This is probably one of the coolest and most powerful aspects of the ACToolkit package.
​
I can also control the range of frequencies from which the pseudo-random values are chosen. I've done some work at the top of this subpatch to control that range, so I can move the upper and lower boundaries easily. In addition to the settings above, this can also have a major impact on the nature of the sounds the jeynoise objects produce.
​
All of Topophony, no. 1's sounds originate from this trio of jeynoise objects. In Phases One and Two, the sounds produced by this trio of objects bypass my Noise Sculptors. So the musical journey for the first half of the piece is determined by my manipulation and control of those objects. If you go back and listen to the piece, I hope this will give you an idea of the sonic potential the ACToolkit offers.
​
My profound thanks to Kreij, Motabar, and Jung for their work. I'm learning so much from what you've made, and I feel like I kid in a candy store!
Pictured here is the section of my noise engine patch dedicated to the synth voices and drum machine. Sounds produced by the selected noise source pass through six individual noise sculptors that produce complex tones in 20-parts (the synth voices), as well as four more to produce percussive tones (bass drum, trash snare, and two clap sounds).
The synth voices have unique configurations for the arrangement of their 20 partials, work I did specifically for this engine. This includes harmonic spectrums that are stretched and compressed, as well as spectrums produced by various mathematical procedures. Here's a breakdown of some of those processes.
​
Compressed/Stretched: Recalculations of the multiplication factors for the standard overtone series. A compressed series has values with a smaller range, and a stretched series has values in a larger range.
​
Metallic Spectrum: The metallic spectrums apply a few mathematical operations to the members of the standards series. Each spectrum raises the members by the power of a particular number, then takes the square root of each of those values as the new multiplication factor for each partial. I've dubbed it "metallic" for its timbral quality!
​
Log2: Selected multiplication factors from the standard series are input as X in the expression 2^n = X. The output is n. If X = 2, then n = 1, because 2 to the power of 1 is 2. If X = 3, then n = 1.584963, because 2 to the power of 1.584963 is 3. And so on. The output from these expressions, n, are taken as the new multiplication factors. The spectrums produced by this process are joined with the metallic timbres for their timbral quality.
​
Blended Spectrums: First, fractional values from the undertone series (a mathematical inversion of the overtone series) are multiplied to a new value higher than 1. For instance, the 3rd undertone (original frequency times 0.33333) is multiplied by 2 a few times until it reaches 2.666667. Next, this number is treated as a new "1", and a new "series" is produced corresponding to that number. Finally, selected multiples from new series is taken together with selected multiples of the standard series, to become the final series by which a fundamental frequency is multiplied. So for the example above, members from a series on "1" and a series on "2.666667" are taken together to form the new series: 1.0, 2.0, 2.666667, 3.0, 4.0, 5.0, 5.333334, 6.0, 7.0, 8.0, 9.0, 10.0, 10.666668, 13.333335, 16.0, 18.666667, 21.333336, 24.0, 26.66667, and 29.333338. I've dubbed the spectrums produced by this process "blended spectrums", because they collect together frequency values based on two different fundamental tones (an oiginal, and a second that is mathematically related by way of the undertone series).