Composers' Desktop Project
Sound Transformation software such as the CDP System is designed to transform sounds in weird and wonderful ways, and with extreme subtlety in the fine-tuning. This touches on today's forays into 'sonic art' and 'sound design': composing with the qualities of sound itself as well as with notes and rhythms. The realm of orchestral scoring has always dealt with tonal qualities in its own way, but music has moved much deeper into this realm since the advent of the 'classic' sound studio based on the use of tape recorders, synthesisers and samplers, FFT analysis and manipulation of digital data by computer, using software algorithms.
Working with sound in this way in an effective manner requires a fundamental extension of musical aesthetic. We are now exploring the inner regions of sound in a way never before possible. This is exciting enough as a scientific and technical endeavour. It becomes even more exciting when the tremendous richness of sounds begins to be employed in the service of musical expression. To do this, we have to use the technical, but move beyond it, making connections with the basics of shaping time: musical art. The terms described in this musical glossary are familiar enough. Here, however, a special effort is being made to connect these familiar terms with relevant aspects of the electronic manipulation of sound, to show how the latter can be used in the service of musical art.
Accelerando means getting faster. Music moves along at a given tempo, meaning the rate at which the main beats occur. During an accelerando, this tempo increases. Just as the heartbeat accelerates when running or in a fear-inducing situation, accelerando implies excitement, the nature of which is determined by the surrounding context. The opposite is a decelerando.
The word 'articulated' is used to describe flexibility in movement, such as an 'articulated truck' in which the cab and storage section are independent units connected by a ball joint. The term is used in music to describe flexibility in a performer's movements, such that the manner of performance is always changing: extra energy at the beginning of a sound (attack transient), increasing and decreasing pressure on the bow or mouthpiece (crescendo / decrescendo), staccato and saltando bouncing, dancing effects, use of mutes to alter the tone, whisperingly quiet, etc. This performance flexibility brings a piece of music to life, distinguishes one performer from another, projects the inner intention of the music, and is altogether one of the most important aspects of music-making. When working with electronically generated sound, every effort needs to be made to achieve a similar flexibility: the music should not just play back through the speakers, the listener should feel that it is being performed.
During a crescendo, the music gradually becomes louder. The opposite is a decrescendo. As music is a time-based art, every change over time has a poetic meaning. A crescendo indicates a gradual increase in excitement or tension, a decrescendo, a relaxation of effort or an easing of tension. These changes of volume are one of the most basic ways to give the music a living, organic feel. Similarly, moving along at the same volume can be part of communicating a steady state, contented, or resigned frame of mind.
Crescendo and decrescendo can also be realised in an 'orchestral' way. Sometimes this is accompanied by a change of volume, but not necessarily. An 'orchestral crescendo' is achieved by adding in more instruments, so that the overall sound and tone becomes more complex, fuller, richer.
An 'orchestral crescendo' can also be achieved in the electronic domain. The sound itself can be enriched by changes which make it timbrally denser and more complex: i.e., by compressing or adding in more frequency components. We can also layer in more instruments or effects by chaining processes and mixing operations.
A musical form is really the idea behind a piece worked out in both the inner pattern detail and the overall temporal design such that everything hangs together and serves a common purpose – whether this purpose is harmonious unity or chaotic diversity. The verse and chorus sections of a song, for example, alternate narrative detail and general statements which capture the essence of attitude or mood. The musical setting of these words usually supports their meaning with gestural shapes, harmony, rhythm, articulations and tonal qualities, unless deliberate irony is sought.
Suppose the idea for a song or piece of music were to be urban complexity. This might suggest interleaving widely divergent sonic material, fast cutting from one scene to another, tonal complexes, angular rhythms, and incisive articulations. Putting all these features in place constitutes the practical, step-by-step process of form-building.
Another idea for a song or piece of music might be isolation and loneliness, perhaps poignantly summed up in the song's chorus section. The task of realising this idea in the music might focus on this chorus section, approaching it and leaving it through moments of quietly sustained material. Gestures and contour shapes might maintain low, slowly changing profiles.
The basic idea as expressed in words, or just a simple abstract idea such as 'rising' becomes the guide to building a form, shape by shape, step by step, until the whole emerges from the detail. The overall idea guides the shaping of the detail, but it is also a fact that what happens when working on the detail can affect and reshape the original idea. The various musical terms described here show how the realm of sound, electronically manipulated, can be used to create this detail. But the purpose, the basic idea, needs to remain in focus as the guiding light in order to achieve organic unity and emotional power.
The term 'form' as used here relates to the overall organising idea of a composition, its 'commanding form' as Susanne Langer says, or its 'ur-form' to use Schoenberg's term. A Endrich is currently seeking to articulate a lower level of form that lies between the ur-form and specific techniques of composition. He calls this level 'micro-forms' and considers them to be generic time-based processes that generate the musical argument of a composition.
We are most familiar with the concept of gesture from the realm of theatre and dance. We perceive it as a distinctive and commanding shape which conveys psychological intent (or just pure form). It is also a key ingredient in music, giving it intent, purpose, 'feel'. Perhaps we could describe gesture as contour with meaning. Similarly, the underlying contour shapes in a painting play an important role in its interpretation.
Changing parameter values inevitably create a contour as they move higher and lower. Whether this contour is gradual with soft edges, abrupt, slowly changing or jagged already has feeling / psychological implications. The amount of displacement and rate of change are major factors in gesture because they create varying levels of intensity: the shape itself implies different levels of energy needed to create it. The sonic artist continues to develop the meaning of this contour in many other ways, matching attack transients, tonal color and other articulations to give it the right energy profile.
Sometimes an artform is full of 'dramatic' gestures. At other times, the gestures are soft, supple, relaxed or elegant. Getting the gestures to match the overall form, character and intent of the artistic creation is a key task. The organic unity of a piece is to some degree tied into the repetition of gestural shapes, often in different parameter fields and time levels. These are, then, important considerations to bear in mind when designing envelope contours and automating parameters.
Except for the odd humorous trombone slide or expressive slide across the strings of a violin or the keys of a piano, glissando has been a rarity in the music of the past. In our own time, this has changed enormously. A glissando is a pitch slide, and pitch slide (or pitch 'bend') has been used extensively in all forms of jazz and popular music, both in voice and instruments. Particularly with the voice, pitch slides with timbral changes (e.g., rough to smooth) are an incredibly important and skillful aspect of performance. Similarly, composers such as Iannis Xenakis have astonished audiences by making whole pieces dominated by glissandi.
Even so, Nature has been way ahead of us. What do we hear when a flock of 100+ starlings gather but a whole texture of glissandi? What about the roar of volcanoes and the whooshes of avalanches, the fluid movements of landslides and mountain streams and waterfalls? Now that we are working directly with sonic material, these shapes become relevant, become possible! Paradoxically, while working with the most advanced computerised systems, we find ourselves coming closer to Nature.
The ability to create glissandi is, therefore, an important facility in all of the CDP pitch/frequency transposition programs. Using programs such as MODIFY SPEED, REPITCH TRANSPOSE or FOCUS ACCUMULATE, finely crafted pitch bends, gigantic slides or flowing, throbbing complexes can be created. Emotional nuances, sensations of flying or falling, flowing, fluid turbulence found in Nature – all have a sliding, flowing form and can be evoked by musical glissandi. Glissandi are achieved by creating time-varying breakpoint files.
We are familiar with grains of sand, salt and sugar, but not necessarily grains of sound. Yet this is entirely possible with the computer, which can divide up and play with the stream of numbers representing the sound in any way one likes. A granulated sound usually comes out as a series of tiny pulsations, a sonic surface like sandpaper, with the degree of coarse or fine depending on the size of the grains and overlap factors.
By extension, any sound with a noticeably rough surface of (semi-)regular attacks can be said to be 'granulated' or grainy. Trevor uses the word 'grain' as it may help the user to find the GRAIN programs, specifically meant for dealing with such sounds. Also, natural grainy sounds have only semi-regular attacks – regular attacks immediately make them sound synthetic.
Harmony takes on a new, much broader meaning when creating music with sounds. One of the reasons for this is that many sounds from the natural world are very complex. Some birds, such as the wood thrush, can sing three pitches at once, warbling on two of them and sustaining the third. The sound of a train is a veritable compendium of sounds.
We are mostly used to 'pitches' in music: clear tones in which one main pitch (called the 'fundamental') is loud and clear, while the other partials in the sound are harmonious with it: locked into it by being frequency multiples: 'harmonics'). Large bells have such a big surface area that they can create more than one pitch at the same time, easily perceived as separate. I recall hearing a beautiful bell ringing out across a valley in Switzerland: two pitches a fifth apart were audible and slowly phased against each other.
As we move into the realms of nature and machines, sounds become altogether more complex, so if we are going to use these sounds at all in music, much less use them effectively, we need to broaden our understanding of harmony.
In general, this refers to repeating patterns. Many software packages automatically loop a sound (repeat with a smooth join) so that changes are applied continuously. CDP's EXTEND LOOP program offers a different set of possibilities by controlling the movement of segments of specified length: gaps or overlap.
In software such as Cakewalk's Project5, the input musical sound from the software synthesiser repeatedly loops back to the beginning to give a continuous sound output. Using the LFOs in its plugin Spectral Transformer (a CDP creation), various parameters can be set to loop round a waveform pattern, whether quickly or slowly – more noticeably when the LFO is set to a slow speed. This repeating pattern of parameter change is then phased across the repeating input sound. The result can be a wild, churning sound, or one with slow, controlled timbral evolution. There is, then, also looping that affects small internal segments of the sound.
In general, regular looping is an extremely mechanical process, not found in naturally occurring sound sources. It appears to be a very important psycho-acoustic feature of sound that completely regular repetition always sounds contrived, whereas natural sounds that appear to be repetitive (e.g., the rolled 'rrr', the sound of crickets, etc.) are full of subtle variations of timing, pitch and spectrum.
To produce a more natural effect, therefore, CDP (T Wishart) has developed EXTEND ITERATE, which allows you to vary the timing, pitch and loudness of the repeated elements. ITERATE tries to produce a 'naturalistic' rather than a mechanical 'looping' of sounds. Compare the output of EXTEND LOOP with the output of EXTEND ITERATE using very small (less than 0.1 semitone) pitch variation, small randomisations of the timing and small amplitude shifts. It is difficult to imagine that the two output sounds are from the same source.
This technique has been familiar in music since the evolution of the 'development section' of sonatas. However, more specifically, it was developed by Russian film-makers in the early part of the 20th century, such as Eisenstein and Verkapich. Both sound and image were patterned in complex ways to achieve specific effects. The Britannia Micropaedia article on the subject (Vol. 8) describes how Alfred Hitchcock repeats the word 'knife' in the thoughts of a frightened girl who believes she has committed murder.
The montage form of assembly has been an important part of the evolution of music in the 20th century and particularly of musique concrète since its earliest days. The emphasis is on the amalgamation of disparate material, while with other types of texture the emphasis is on replicating the same or similar material. Pieces such as Stockhausen's Momente and Hymnen provide many examples of audio montage.
The various approaches to morphing audio are described in the Reference Manual and in the 12 Step Tutorial.
The key words with morphing are 'change' and 'gradually'. We are familiar with the way visual images are distorted as one shape is bent out of shape in the process of changing it to another image. These striking images are staple diet in cartoons and now have a new lease of life through CGI (Computer Graphic Imaging). The movement is relatively slow, whether at 24, 25 or 30 frames per second, which is OK for the eye.
The process is not at all easy in audio, because the ear is so finely attuned to the smallest changes – the ear can detect quality differences in sample rates at 22500, 44100 or 96000 samples per second! Our generic recommendations therefore are:
The ability of computers to produce random number series provides the basis for many procedures useful to composers. They are often generated within (time-varying) upper and lower limits, to give added compositional control.
On the one hand, it can help to give a natural feel to rhythms, especially with multi-event textures. But many other parameters are suitable to randomisation, such as segment positions and lengths, pitch, volume, spatial placement etc. All of these are implemented in the TEXTURE Set, BRASSAGE, and GrainMill. The TEXTURE Set also uses randomisation to select among several defined motifs or ornaments.
Randomisation also gives flexibility in the EXTEND programs DRUNK, SCRAMBLE and ZIGZAG. Whenever there are min – max lower and upper limits, selections are made between these (time-varying) limits by a randomisation procedure. As usual, compositional control increases when there are constraints to work within.
Although it can be over-used, randomisation is an important tool for the sonic composer. The temporal placement of sounds (as opposed to pitches) can seem artificial if they are too regular in a context that doesn't justify such regularity. Randomisation enables the composer to loosen things up, as well as to introduce unpredicatability without having to hand-code every variant.
Working with sound also requires, at times, a more flexible approach to rhythm. For much of the time, altering the tonal qualities is designed to give just the right feel to the beat, whether with the type of attack transient, the brightness or warmth of the sound, or the pitch bend which shapes its pitch. A slight randomisation of where the beats fall has been shown to 'humanise' the rhythm, to sound less mechanical and more as if produced by live performers.
At other times, the sounds and textures are more complex, and we come closer to the world of Nature, in which regular rhythms are the exception rather than the rule. If we recall the texture of birdsong emitted by a seagulls or listen to the wind moving through the leaves of an aspen, we realise that there is a significant random dimension in the sounds of Nature. When using natural sounds, or abstract sound complexes for which there are no known equivalents, it is often useful to make use of some degree of randomisation in the rhythmic feel.
Effective use of these possibilities requires daring and imagination, but also a regard for the overall idea (form) of the music.
Western music divides up the octave into 12 more or less equal divisions, called semitones. Because of variants in temperament (tuning method), the actual size of semitone divisions can vary. Semitones are themselves divided into 100 cents. All specification of semitones in the CDP software may be fractional. Two decimal places equates with cents, the division of a semitone into 100 parts: 0.01 to 0.99. A whole tone comprises two semitones. Thus 1.50 means 1 semitone + ½ semitone, i.e., ¾ of a whole tone. Microtones of great precision are therefore catered for in the CDP software. This is also true of the (CDP) Spectral Transformer plug-in for Project 5.
'Smooth; is a simple word, but enormously relevant in this context. It can refer to soft tones, harmonious spectrum, gentle onsets and decays, and relaxed, lilting rhythms. It can also refer to the ways in which a sound can be blurred or dissolved, using functions such as BLUR BLUR, FOCUS ACCUMULATE, the FILTERS, HILITE TRACE and HILITE BLUR-TRACE. The computer sound manipulation tools therefore make it possible to affect the smoothness of a sound in many different ways. This affects emotional tone and the recognisability of the sound.
Tempo is the number of beats per minute, the pulse. It is usually as
expressed as a note-value (e.g., a crotchet/quarter note) = the
number of times per minute it occurs. MM stands for Metronome Mark.
e = 60. In this case, a crotchet
occurs 60 times each minute: each beat lasts 1 second. The general
formula for the duration of each beat at a given tempo is:
Number of seconds in a minute Number of beats per minute (tempo)The inverse of this is the number of beats per second:
Number of beats per minute (tempo) Number of seconds in a minuteThus 60 sec. ÷ 60 beats = each beat lasts 1 second. When the tempo is doubled, e = 120, so 60 sec. ÷ 120 beats = each beat lasts ½ sec. and 120 beats ÷ 60 sec. = 2 beats per second. For MM e = 84, 60 sec. ÷ 84 = each beat lasts for 0.714 sec. and there are 84 ÷ 60 = 1.4 beats per second. There is a more detailed discussion of tempo in Time Chart.
Tempo is directly related to human biological activity: the heart-rate increases when there is more activity in order to increase the blood flow to the muscles. Thus tempo is a component in expressing action or relaxation, excitement or peacefulness, and fear or feeling comfortable. Which tempo to use, as well as changes in tempo, immediate or gradual, are therefore musically important features.
In sound transformation software, the concept of tempo is further extended by including the rate of movement of features internal to the sound. The glissando parameter in FOCUS ACCUMULATE controls the rate of internal pitch bend.
Another aspect of tempo relevant to sound composition is the rate of information flow. More information in less time increases the sense of movement and compresses the sense of time passing.
'Texture' is a very generic term in that it can relate to so many things.
Sonically, it relates to the smoothness of a sound's flow, just like the surface of a painting, which can be almost completely flat, like a watercolour, or have increasingly rough surfaces: applied with a palette knife or a collage of disparate materials. Granulation is perhaps the primary technique for texturing a sonic surface.
Texture also relates to the layering of sonic events, such as homophony (melody + chords), counterpoint (simultaneous independent melodic lines), and heterophony (intertwining variants of the same melody). Similarly, short motifs can repeat and vary in musical forms such as fugue bridge passages, sonata development sections, and the motivic repetitions that result from (atonal) 'series' built from repeating figures.
This brings us to textures in sound composition, where the same sound or similar variants of it are repeated in various densities and overlappings in horizontal and vertical space. Specialised software such as Trevor Wishart's TEXTURE Set facilitates the creation of such textures, with additional features such as randomisation, mapping to harmonies, using specified rhythms and motifs, and building textures on defined linear subsubstructures.
As a time-based art, music is intensely concerned with how all its component ingredients change over time. These patterns of change, these time contours, create a sonorous symbol which is perceived both by the mind and the body as the experience of music. Without time-varying sonic events, nothing is happening, from a musical point of view.
However, it should also be remembered that temporal changes, as with gestures, need to be in step with the overall musical form and intention. It is easy to create helter skelter wild textures with computer tools. While this is sometimes fun and appropriate, the other end of the spectrum should be explored: where the changes over time are slow, easily perceived, and tied in with the overall form and gestural nature of the music.
In the electronic domain, therefore, it is essential that the component parameters for a sound vary over time one way or another and that these changes are consonant with the overall form and intent of the music – unless a contrast of outer form and inner pattern is deliberately intended.
The majority of parameters in the CDP software vary over time. The changing times and values are specified in a breakpoint file. Our first venture in real-time automation was with the Spectral Transformer plug-in for Calewalk's Project 5, where there are low frequency oscillators (LFOs) and ultra low frequency oscillators (ULFOs) as well as the parameters themselves.