AEOLIAN VOICES (2013)
for eight digital soundtracks

---

In many cultures, wind is regarded as the “breath of nature”, and frequently described in terms of human vocal expression such as sighing, moaning, howling and so on. As an invisible force of moving air, wind is largely experienced through its interaction with ourselves and the objects it encounters, becoming audible by moving through and activating these materials. Also, wind can be stirred up by moving objects and can scatter the sounds that it carries, hence its elusive character, at times playful, at other times relentless.

This soundscape composition takes the listener on an imaginary journey, starting in a windy wheatfield with a windmill, where a passing car kicks up some stones and wind gusts, and the wind scatters the sound of distant bells and geese. We then move into a shed where wind whistles through the cracks and crevices, banging anything loose enough to move, creating its own resonances and percussion. Finally we encounter ourselves in a heavy windstorm where our imagination may be tricked into thinking we hear more frightful sounds that are partially masked by the wind. All of these effects are the voices of the elusive wind.

Scenes:	The Wheatfield	The Shed	The Windstorm
Production Score:	Prod1	Prod2	Prod3
Time:	0:00	4:30	6:45
Audio:	Part 1	Part 2	Part 3
Spectrogram:	Part 1a, Part 1b	Part 2a, Part 2b	Part 3a, Part 3b

Materials:

All sounds in the piece were recordings of wind in various environments (e.g. forest, wheatfield), or sounds activated by the wind (e.g. windmill pump, wind whistling through a shed) or distant sounds that are being diffused by the wind (e.g. bells, geese). Fifteen files, consisting of various types of wind, were convolved with themselves  using SoundHack to produce various hybrids, each normalized with high frequency boost, and later EQ'd with fades added; granular time stretching and digital resonators using Chris Rolfe's MacPod were also used.

I. Wind Convolved with itself or Resonated

Original Source Materials	*self	Resonated	Resonated*self
Wind 1 Wind 2 Alberta Wheatfield		Wind2A Wind2B	Wind2A*self Wind2B*self
Wind 3 Alberta Poplar Grove	Wind3*self		CombMix
Wind 4 Wind 4A Quebec Field Wind 4B Alberta Lodgepole Pines
Wind 4C Nova Scotia Outhouse	Wind4C*self
Wind 5  Wind5A Saskatchewan Bushes	Wind5A*self Wind5B*self
Wind 6A Saskatchewan Window	Wind6A*self
Wind7, Wind7A B.C. Pine Forest	Wind7*self Wind7A*self
Wind 8 Alberta Aspen Grove	Wind8*self
Lake 1, Lake 2 Alberta Mountain Lake

II. Wind Activated Percussion Convolved with itself, Stretched or Resonated; Car Passing; Trees Creaking

Original Source Materials	*self	*self*self	*self*self*self	Resonated	Resonated**+Stretch	Mix
Shed 1 Shed 2, Shed 3 (Alberta)
WindCry	WindCry*self	WindCry*self*self
WindRattle	WindRattle*self			WindRattleRes	WindRattleResStretch
WindSlam	WindSlam*self	WindSlam*self*self	WindSlam*self*self*self			WindSlamMix
WindWhistle	WindWhistle*self
WindWhistle2	WindWhistle2*self	WindWhistle2*self*self			WindWhistle*selfStretch WindWhistle*self*selfStretch
Door Slam (Yorkshire)	DoorSlam*self	DoorSlam*self*self				DoorSlamMix
Door Rattle				DoorRattleRes	DoorRattleResStretch
Door Rattle short	DRattle*self
Door Rattle 2	DRattle2*self
Car Passing	Car*self	Car*self*self	Car*self*self*self			CarPassingMix
Car Passing 2	Car2*self	Car2*self*self	Car2*self*self*self			Car2PassingMix.aif
Creak 1	Creak1*self
Creak 2	Creak2*self	Creak2*self*self
Creak 3 (short)	Creak3*self	Creak3*self*self			Creak Sequence

** Delay line resonators set to 5.6 and 8.4 ms for left and right channels, producing a pitch of approx. B-flat/F, with feedback varying from 70-90%.

Production Score

III. Spatialization

All of the material in the work is spatialized with the TiMax2 SoundHub, developed by Harmonic Functions. Our version uses it as a 16 to 8 matrix mixer where at every cross-point in the mixer, an "image" describes how the signal is routed to the 8 output channels in terms of level (dB) and time delay (ms). In this particular work, the images are all members of a family called DelRamp (delay ramp) where one output channel is strongest (0 dB), and the neighbouring output channels drop off sequentially by about 0.9 -1.4 dB and each adds a 10 ms delay. Two subsets of the family are a Clockwise one (e.g. DelRamp1) and a Counter-clockwise (e.g. DelRamp1R). The aural effect of the use of these delay ramps is that the signal is smeared over all of the other speakers, a little like the trail behind a jet. The visual and numerical representations of some sample images are shown in the table below (with the relevant numerical values marked in red, the amplitude levels shown as faders, the delays shown with an up arrow fader, and the output channels listed on the bottom row).

DelRamp1	DelRamp5	DelRamp8
DelRamp1R	DelRamp5R	DelRamp8R

In all cases a Trajectory is created by placing 8 such images in a row, indicated in the Production Score, for instance as DelRamp1 CW->8 which means that the sequence of images goes clockwise around the ring of speakers in the sequence DelRamp1, DelRamp2, .... DelRamp8. Likewise, the notation DelRamp8R CCW->1 indicates that the sequence is DelRamp8R, DelRamp7R, ... DelRamp1. In all cases, the right hand channel is one speaker to the right (e.g. when the left channel goes to DelRamp1, the right goes to DelRamp2, etc.). Note that the speed of the trajectory depends on the length of the file, and in this work, one or more complete rotations around the 8 speakers occur during the entire file as noted, meaning that each file with its own duration will be spatialized at different speeds. At certain key moments, as noted, the trajectory will speed up and slow down.

Other spatial patterns are also used in this piece such as DelV1, Echo1 and Comb1, as shown below.

DelV1	DelV5	DelV8
Echo1	Echo5	Echo8
Comb1	Comb5	Comb8

---

home