3.4.1 Experiential categories of processed voice sound


As discussed in Chapter 2, the choices made in the improvised interplay are (mainly) based on an “inner ear” experience of the sound prior to the musical action. For me, a grouping of sounds and techniques in categories is one way of structuring the possible choices.  Being a musician and not a music technologist, I am more interested in the sounding result of different types of sound processing than the technology involved. I find myself working with four main categories for manipulating voice-sound. These are experiential categories ‒ based on how I experience the sound quality:

–       (a) Broadening: adding something to the voice

–       (b) Narrowing: filtering the frequencies of the voice

–       (c) Placing: putting the voice in different rooms/spaces and distances

–       (d) Reconstructing: changing the voice sound more substantially

(a) Broadening: adding something to the voice

Example III, 5: “Numb Street Cabaret” with Thomas Strønen – from the CD Voxpheria (2012).



From the start to 2.52: I am using a “cluster”; a pitch shifter effect that I have programmed on the Lexicon (see Chapter 2), by adding two close notes to the original sung note. (This is also the effect I use in the example “Spring is like a perhaps hand” in Section 3.2.3) This makes the voice broader, and more unnatural. It also triggers, here, a way of singing that is theatrical, and it sounds a bit absurd. This inflects the musical idea for the whole expression, I think.

One variant of broadening includes a tonal/harmonic focus. This happens when the interval between the voice and the added pitch is increased above the second, especially with consonant intervals – and if there is a melodic or tonal focus.

Example III, 6: Excerpt from a version of “Mercy Street” (Peter Gabriel), studio recording with Krister Jonsson and Per Oddvar Johansen, 2009.



Here I am using another pitch shifter effect programmed on the Lexicon, adding a 5th under, for tonal reasons as well as for the broadening effect.

I would like to compare this with a non- tonal focus:

Example III, 7: Excerpt from ‟Western Wind” from the CD Numb, number (2012), BOL + Hans Magnus Ryan and Stian Westerhus.



Here I use the same effect as above, but with a speaking voice.

Example III, 8 : Solo part from “Singing again” from the CD Numb, number (2012), BOL + Hans Magnus Ryan and Stian Westerhus.



I use a “slicer” effect, “cutting up” the sung note (from the Roland SP 555). I balance the clean signal and the processed signal so that the effect is in focus. This is an overdub done in studio, after first doing a take with a more acoustic solo that did not work too well with the music. The use of a slicer effect created a kind of mechanical vibrato that allowed a more instrumental, sound- focused approach. This connected better with the guitars.

In all these examples, I experience the voice-sound as becoming richer and broader, and at the same time less direct due to the unnaturalness. This is something that I experience with different pitch-shifter modes, and also with effects like flanger, chorus and reverse delay, etc.


(b) Narrowing: filtering the frequencies of the voice

The term ‘narrowing’, points towards my experience of reduction of the voice sound working with frequency filters.

Example III, 9: “Numb Street Cabaret” with Thomas Strønen – from the CD “Voxpheria” (2012):



At 2.52 I use an EQ/filter-effect (Roland SP 555) and a compressor to create a radio-like voice. The filter removes  lower frequencies and adds more of the upper middle frequencies. I find it interesting to see how removing low frequencies can make the voice seem “closer”, more insisting and more important. Could one of the reasons for this be that we are used to receiving important information (airports, trains, fire alarms, rehearsals, etc.) on very bad sound systems which lack low frequencies? Or is it because of the reference to radio or telephone sound; someone far away, but still close to you ear? The compressor makes the consonants very clear, small sounds are amplified and the voice seems to be very close – it resembles whispering, which is often meant to be important and is just for you to hear (personal).

Example III, 10: Excerpt from “Waiting time”, from the CD Numb, number (2012), BOL + Hans Magnus Ryan and Stian Westerhus.



From 0.39 I use the same type of filtering and compressing as above, on my singing. I experience, again, that this technique makes the voice sound “closer”,even though it is unnatural and not real world. This ‘sound’ is also inspired by what I think of as the “Stina Nordenstam-sound” (see 1.5.5). I also register that when I use filtering on my voice in real time, I am often enhancing the upper middle-tone area, while I tend to bring out the low frequencies when I filter sampled sounds. I probably use the upper middle tone filtering to make my real-time voice more present, while I often use the filter to vary sampled sounds and therefore use a wider frequency range, also including lower frequencies.


(c) Placing: putting the voice in different rooms/spaces and distances

By using different kinds of reverbs, delays and compressors, the voice sound can be put in different rooms, and be experienced as though it is coming from different distances in the musical scenery.  This inflects the experience of room and space in the musical expression as a whole.

Example III, 11: ‟When what” improvised studio session with percussionist Marilyn Mazur, Copenhagen 2011:



Throughout the session the percussion is “placed” in a medium “distant” room, while the voice has various “spatial placements”. The pre-samples I use at the start are experienced as being a little bit closer than the percussion, and the natural voice as being much closer, by using – among other things – less reverb. At 3.37 I  change from relatively “close” natural voice sound, to – by using a longer reverb – making it more distant, and sampling phrases that stay in this more distant place due to the use of the reverb. The whistling sound towards the end is also placed in this more distant area. I experience the placing of the voice as being a musical parameter here.

Example III, 12: Excerpt from “Udu”, studio improvisation with Marilyn Mazur 2011.



Here, I work with a talking voice with a short reverb (“dry”), a whispering voice with a long reverb (“wet”) and thereafter a “wet” singing voice. Since the percussion also has a long reverb, the “dry” talking establishes a spatial reference giving perspective to the experience of the total space.


(d) Reconstructing: changing the voice sound more substantially

This is the widest category in the sense that the various techniques can sound very different from each other. Examples of these techniques are:

–       Granular synthesis in Max MSP

–       The use of different states in the Hadron

–       The more extreme effects on the Roland SP555

(like the DJFX-looper and the Ring Modulator-effect)

(All of these techniques and machines are described in Chapter 2.)


Granular synthesis and filtering in MaxMSP

I use a custom-made Granular/Filter-patch (G/F patch) programmed in MaxMSP (see Chapter 2). Vocals with granular synthesis sounds very different to what I can create with my other effect-machines. One major difference is that I can create random variations, both regarding the granulating process and the filtering. This makes it possible to work with loops that are constantly changing “on their own”. I experience the processed loops as being organic rather than static (this can solve some of the challenges with looping, which I will mention later). Sometimes the use of the G/F patch gives a sensation of playing with someone else; choices are being made outside my realm. The patch has a lot of possibilities, and in the following I will provide examples of some of them.

I use the G/F patch to create accompanying layers or an underlying “fundament” in my music. Very often the voice is not, or hardly, recognisable as the source of the sound (at least to people not familiar with this type of processing). I experience that granular synthesis can give some very organic-sounding variations. I will return  to example no. III, 12, the “Udu” session with Marilyn Mazur:



An important musical fundament for this session is a sample of short sounds made with the lips, processed through the G/F patch, with short grains and varying density. I am attracted to the way it sounds, and how it ‒ in a minimalistic way ‒ changes all the time, by the stream of short grains with varying density and deviation. The sound immediately led Marilyn to choose the udu as a matching instrument.

I also use the G/F patch to create tonal fundaments:

Example III, 13:  “Grains” with Thomas Strønen, from the CD Voxpheria (2012):



The processed loops define the tonal fundament, and they are also an important timbral element, giving the piece a certain colour. At 1.49 I sample a sung note for a loop in the G/F patch, where the pitch is set to an octave+4th below the original note. Between here and 2.54, I sample this down-pitched loop on the Repeater loop-machine, and then at 2.54 I transpose the processed loop up a 4th in order to have a broader tonal fundament that I can vary. Without the granular synthesis, this would sound like “a singer accompanying herself with pitched voice-loops”, but in this way the loop acquires an instrumental character. Again, it is also the sound in itself that attracts me. The way the sound is transformed by the synthesis and varied through the filters in the G/F patch makes it very different to the sound of the pure voice. The connection between the first sung note and the repeated sample is not very obvious, and this takes away the focus from the looping technique as an ‟effect”. The tonal layers “grow out of nothing”. I also use the G/F patch for loops without a specific tonal focus:

Example III, 14: Excerpt from solo performance “ Eugenie – short stories of sound”, live at the Ultima Oslo Contemporary Music Festival 2011:



At 0.07 I use a sampled loop with no tonal focus, in a low pitch, with longer grains and a lower density. This creates a deep and “disturbing” sound-layer. The irregular random fluctuation and density of sound impulses make it “alive”.

Example III, 15: “Numb”, from the CD Numb, number (2012), BOL + Snah & Westerhus :



Here, at 2.19 and 3.07, I use granular synthesis of spoken words (not heard as words, but sometimes as word-like), as a textural “sound-shower”, which I experience as being connected to the spoken samples that are used from the start.


Hadron plug-in synth in Ableton Live

The Hadron is my newest device and is still something that I am exploring. (see Chapter 2). What has attracted me so far, but which also is a challenge, is how it creates a very different output compared to what I produce with other devices. To me it sounds very “digital”. What I have used in the Hadron is the possibility of making a pulsating, highly processed sound-layer from vocal samples. It sounds different to what I can do with the DJFX-looper on the Roland SP555 (which has some similarities; it can create a pulsating, processed sound from a sample and change it by pitch modulation and tempo). The Hadron-pulse, which I often combine with various plug-in filters in Ableton Live, can be varied in many more ways than the Roland effect, but it is also more unpredictable and complex, as discussed in Chapter 2.  Still, trying to implement the very different sound quality in the Hadron is interesting; I want to examine further what it can bring into my vocabulary. Up until now, I have been using the pulsating effect in some improvisations:

Example III, 16:  “Thhh” (again), with Michael Duch:



I use the Hadron pulse as a returning element during the first part (until 4.55), starting with sampling breath sounds in 0.17 and at 0.45 entering the pulse for the first time. For me it has two main functions: contrast and energy. It sounds very “machine-like”, in contrast to the other sounds, especially the acoustic bass. The pulse gives the music a movement, not necessarily “forward”, but adding something more rapid and continuous into a kind of “punctuated” interplay.

Example III, 17: “Raised, Rave “ (again), with Thomas Strønen:



I use the Hadron pulse (1.34) here very much in the same way as in the last example, as a returning element, contrasting the rest of the sound picture (although not contrasting as much as in the example with the acoustic bass,) and bringing in a continuum moving in and out of the, in general, abrupt expression of the piece.


Roland SP555 – “extreme effects”

The Roland SP555 sampler (link) has a lot of built-in effects, and some of them have become part of my musical vocabulary. When I think of effects as reconstructing the voice, I mean that they make the voice sound like “not voice” or “something else than voice” to a certain degree. The sound relies of course very much on how you balance the effect with the clean input signal: distortion/fuzz can be subtle, adding spice to the voice, or it can sound very alien if you turn down the original vocal input in the live mix. I will exemplify two of the effects I often use, effects that give me an impression of a “reconstructed voice”: the DJFX-looper and the Ring Modulator.

The DJFX-looper

Example III, 18: Excerpt from the solo performance “Eugenie – short stories of sound”, live at the Ultima Oslo Contemporary Music Festival 2011:

From 1.07, I use the Roland DJFX-looper, like I often do when working with noisy parts. It has several functions and controllers: looping a fraction of the input and repeating it, changing the pitch of the voice or the repeated fraction, and changing the tempo of the repetition. It functions, for me, as a means of creating a powerful and sometimes aggressive expression. It is voice and “not voice” at the same time, often giving a feeling of the voice trying to “break through” or being “broken down” by something powerful.

Here is another example, where I use it in a much more moderate way:

Example III, 19: “Raised, Rave” with Thomas Strønen, again:



Here, I use the effect in a less “dramatic” way, picking up some phrase endings and samples at the beginning (from 0.07) and then repeating fractions of some of the phrases (from 2.02). The “machine-like” effect are important, providing a surprising, contrasting and abstract element, especially when compared to the natural voice.


The Ring Modulator-effect

Example III, 20: BOL: Excerpt from “Nature is not Beautiful”, live performance at Teaterhuset Avant Garden, Trondheim, 2011:



I use the Ring Modulator-effect on the Roland SP555 with “word-like” speech .The acoustic voice sound is heard, but rather distant, sometimes only through the reverb. I experience that the sound of the Ring Modulator “reconstructs” my voice into sounding like something else. The effect creates a kind of fragile, transparent and distant expression using this type of vocals as an input signal.


The mix of categories

 The different experiential categories can, and will, of course, be mixed in different ways. The processed sound, broadened, narrowed or reconstructed, always has a defined position in the musical space, by the use of reverb (or lack of it), perhaps a delay and/or compression. The broadened or reconstructed voice sound can be narrowed with a filter and the broadened sound can be reconstructed, and so on. Working with a mixer, I have the choice of balancing and blending the amount of processed and natural sound in different ways. The balance between unprocessed and processed voice sound, the shifts and the cross-fading, are also musical parameters, related to the play with zones, to the whole sound-scenery and the structural elements of the music.

Next page ➡


  1. […] 3.4.1 Experiential categories of processed voice sound […]

  2. […] 3.4.1 Experiential categories of processed voice sound […]

  3. […] 3.4.1 Experiential categories of processed voice sound […]

  4. […] 3.4.1 Experiential categories of processed voice sound […]

  5. […] I am categorising roughly four different types of processing (see 3.4.1): […]

  6. […] I am categorising roughly four different types of processing (see 3.4.1): […]

Comments are closed.

Scroll to top