Figure 1: Aeolian Filter Nord Modular patch.
One aspect that is vital to understand w/r/t Network Instrument theory is that the network is made up of interconnected instruments. Confusion can be had as electronic instruments are fundamentally made up of interconnected components and something like a modular synthesizer could appear to be a Network Instrument laboratory. Now a sufficiently large modular synthesizer could be patched in such at way that it contained discrete instruments that could then be interconnected in various ways, some of which could actually be interesting beyond simply summing signals in a mixer (for instance a four quadrant multiplier would allow you to intermodulate your signals). A large collection of individual modules, or a software based solution that works on similar principles thus can be a a Network Instrument laboratory though what particularly makes the instrument is the variety of interfaces which, especially in a soft-synth, can be severely limited. However it is the networking of the individual instruments that is of interest and is something that can be exploited. This, plus the fact that “instrument” is a rather loaded word that in an idealized network instrument might not meet the basic criteria that people hold, is why in Network Instrument theory they are referred to as subnetworks.
Synthesis v. Complex Waveform Modulation
Typically what people are doing with synthesizers is subtractive , additive, FM synthesis or a combination of these (of course there are plenty of other forms of synthesis: Wavetable, Karplus-Strong etc). A Network Instrument can utilize any or all of these techniques in its subnetworks but it is explicitly not a form of synthesis. Synthesis is taking very fundamental parts, typically quite simple waveforms (triangle, sine, saw, etc) and altering those waveforms to create a more complicated waveform. The Wikipedia links above are actually quite informative for a basic overview of these forms of synthesis and of course there is much on the web describing these techniques in greater details. Synthesis is almost always trying to create other sounds:
Subtractive synthesis is a method of creating a sound by removing harmonics, characterised by the application of an audio filter to an audio signal. For example, taking the output of a sawtooth generator and using a low-pass filter to dampen its higher partials generates a more natural approximation of a bowed string instrument than using a sawtooth generator alone. Typically, the complexity of the source signal and the cut-off frequency and resonance of the filter are controlled in order to simulate the natural timbre of a given instrument. – from the Wikipedia article on Subtractive Synthesis
Additive Synthesis which is the process of sound generation via the combining of simplified waveforms to increase the complexity of the waveform is a bit more akin to the notions of Network Instrument theory. However it is still at a much more basic level than what is done with an Network Instrument. That is the process is still trying to create a singular sound, usually tied to a pitch. The process though is much closer and if you applied the basic concepts not toward creating single sound but on an instrumental level you’d have a simple network, however the methods of modulation between subnetworks via interfaces doesn’t limit the process in which waveforms interact to just the additive. Additive synthesis simply as a process does create enough complexity that it is rarely used in commercial synthesizers but of course this doesn’t limit its use for those that have enough of the constituent parts to explore the technique.
A Network Instrument in contrast to the various forms of synthesis has no interest in replicating the sounds of other instruments or even in creating “new” sounds. It’s concerns are for utilizing the sounds in and of themselves and to increase the complexity and unpredictably of the sounds generated through the interactions of the complex waveforms generated by complete instruments. In this regard a Network Instrument is much more akin to a musical ensemble than a synthesizer. When you see an ensemble, say a string quartet, play live there are four separate sound sources that are bounced around in a space, altered by running into other objects, reflecting from walls and finally being mixed by your brain from two separated inputs. This is what a Network Instrument is trying to do
An Example Subnetwork
Pictured at the top of this post is is the Nord Modular patch utilized in the recording of the aeolian electrics (part of the Eleven Clouds [Hollow Earth Recordings 2010] series). This is a patch that is run on Clavia’s Nord Micromodular which is a DSP based synthesizer that basically runs a softsynth in hardware. With two inputs and two outputs (plus a headphone jack that can be used for additional output) it is a useful and quite flexable addition to a Network Instrument. I often use it as part of another instrument, as several instruments or for processing of other instruments. With it’s multiple Ins and Outs it can be setup to create feedback which it can then process itself. My first use of this synth was what I called the Feedback Synthesizer where I’d only use feedback as a soundsource (no oscillators). It also can internally create feedback from interconnecting of modules which allow for a pretty high degree of instability from a digital instrument. However I find I get the best results when I feed in an outside source into it and incorporate it as part of a network. The typical setup with the Nord is a feedback loop on one channel and the other input either bring in another instrument (in a network sense) or a signal source to be part of the subnetwork in the patch. These two parts can be interconnected and thus form a Network Instrument. In the above example I’m using a single input source, in this case another instrument (my Chimera BC-16) and then I’m splitting its input and feed it into two subnetworks. These two networks are interconnected to increase the interactions between the complex waveforms.
Figure 3: Annotated aeolian filter Nord Modular patch.
In figure 3 we can see isolated the two subnetworks of the aeolian filter. The object used to access the inputs is in the middle of the image (the Network Input) and the two subnetworks are in the top and bottom of the image. The input coming in is split into multiple signals which are then delayed to shift their waveforms. These signals are altered in various ways (filtering, quantizing etc) but primarily they are inter-modulated with each both internally and across the two networks. This of course is the essence of Network Instrument theory: the interactions of complex waveforms leading to increasingly complex and unexpected behaviors. The use of delays here (all very short; the Nord doesn’t have the memory for long delays and frankly I’m not very interested in the use of delay in and of itself) is always to create a new signal which increases the complexities as the are mixed with the other signals. This network (along with the instrument that was used in conjunction with this patch) was being used to invoke the chaotic behavior of the wind; effectively I should add as Brian Olewnick on his Just Outside blog described it thusly:
“One disc involves ringing metallics (a Tibetan bowl buffeted in some manner by wind?), the other more “whistling wind” (through some aperture?).”
The ringing metallics, is one disc (which was actually two contact microphones on windchimes) of this two disc set, and it is the other disc of electronics described in this post that Brian described as “whistling wind” (through some aperture?). The goal of course was not to directly replicate natural processes but to to capture that natural randomness that doesn’t seem random since being in the world it is always around us. Use of the cascading complexities of a Network Instrument worked admirably for this.
Above is an ~18′ excerpt from the ~80′ piece which gives a fair example of the sound. It can be downloaded via SoundCloud as well as played by click on the above.