In working with magnetic tape Gerhard was aware that he was adopting different working methods from those he normally employed when working in the instrumental realm, and was gaining fresh insights into the nature of sound itself. In his notebooks he writes:
The composer at the tape machine is like a commander in the field, he is in the very thick of events. This is a tremendously exhilarating situation. Direct action with actual sound stimulates aural alertness to an unsuspected degree. And – what is even more important – it also stimulates thought as applied to tactics and strategy in quite new ways.

And also:
After a full day’s work by the tape-recorder one suddenly discovers that one’s ears have become […] atuned to all manner of sounds, indoors and outdoor-sounds to which, one realizes, one had been completely deaf before.

The strategy that Gerhard refers to is one in which empirical rather than a priori methods came to dominate his practice when working with magnetic tape and as such marks a different methodological approach in his ‘sound compositions’ than in his notated works of the same period. In his notebooks there are long lists of sounds that form the numerous mixes or ‘compounds’ that he produced before the final montage of a work. There are, however, no sketches, notes, or diagrams referring to the sound compositions themselves. In his instrumental compositions of the late 1950s onwards Gerhard used the serial set to govern large-scale pitch and temporal structure, but within this framework he was able to work out much of the local detail of a work intuitively. In his sound compositions Gerhard took this intuitive process much further, writing that:

The basic resorts brought into play are the same as in ordinary composition on paper – only more so, as it were, which is to say that intuitive and imaginative approach rule supreme. There is no system, no computation, there are no blue-prints. Sound firing the imagination, sound for the love of sound is the prime mover.

This seeming dichotomy between the rigour of Gerhard’s own interpretation of serial time and pitch structures and the freedom offered by tape composition is reflected in an isolated statement in one of his notebooks from 1957, in which he writes, ‘pre-compositional hurdles (parameter organization) = paralysis of the reflexes’. This seems to suggest that Gerhard was aware that extreme parametric organization could only take him so far and that tape music offered him an additional means of structuring material through more textural and gestural means.

As Gerhard continued to work with magnetic tape it is clear that he began to adopt a coherent personal aesthetic towards tape composition. His notebooks become increasingly filled with ideas about the temporal nature of composition, about timbre and texture. While his radio and theatre productions continued to use Foley sound, such as ‘taps on a cardboard tube’ for some of the sounds for the incidental music for Macbeth, Gerhard’s more autonomous sound compositions utilized more abstract or processed sound materials, often instrumental sounds, which in some cases underwent considerable metamorphosis to form hybrid ‘sound families’. Gerhard’s thinking at this time is best summed up in ‘The Composer and His Audience’, in which he writes:

One of the hardest discoveries for the musician to make, it seems, is that music, contrary to a generally held belief, is not made with notes. The eye leads the ear astray; it easily persuades it that the notes are really there – and nothing but the notes, as far as one can see. The ear, therefore, misguidedly concentrates on locating notes, or disentangling their clusters, on tracing the patterns they form. Yet the basic stuff of music is sonic motion, not notes or sounds. Manoeuvre is the raison d’etre of the formations […] The true business of the composer is to release the flow and shape, and steer the stream of sonic events in time.

Gerhard’s notebooks contain numerous annotations of source materials and comments on these. For Gerhard, the first step toward creating a sound composition was to gather a repertoire of raw materials on tape. This process is described in ff. 1-10 of the sound score for the incidental music to King Lear (1955) , which contains detailed instructions for recording a catalogue of instrumental sounds using different dynamics and modes of attack, including: maracas, cymbals, xylophone, turkish cymbal, tam-tam, piano, chromatic timpani, bass drum, gong and mbira. In his studio, Gerhard had a microphone available for making recordings of piano effects – or smaller percussion instruments. But the sound materials he utilized were by no means limited to instrumental sources. Production notes reveal the regular use of daily objects for making sounds (packing paper, paper tissue, combs, ashtray), as well a wide range of incidental noises (birds, dogs, axe strokes, cracking tree, thunder, wind, rain and storm, whipping gusts, crowds, chatter, laughter, screams), which could be home-made or taken from the everyday environment. In his notebooks, Gerhard writes,

[…] we all have got to start in the same way: by building up a repertoire of sounds which are stored on tape. […] The sounds selected may either be appropriate in their original form to the sound-picture one has in mind or else require further treatment before being used. Most of my stored sounds are of instrumental origin, recorded on tape through microphone. The next step – what I called my second stage – is directed towards a certain transformation of that original sound, ideally towards a metamorphosis of the sound [in] which [its] origins are blurred, and a far-reaching change of identity might be achieved.

Gerhard’s methods for obtaining such source materials for his compositions are documented by Lindsay Anderson and Dick Mills. Anderson writes that:

I remember visiting Roberto in Cambridge, talking about the score, and even assisting him in throwing various objects down the stairs, in an effort to product the right kind of abstract sounds which he felt he needed.

Dick Mills, who worked at the BBC Radiophonic Workshop describes recording sessions in which Poldi Gerhard was fond of participating too, writing that:

Roberto had a rather difficult problem to overcome when attempting to record his basic sounds, as he lived on a busy trunk road in Cambridgeshire and the only quiet period was around 3.30 in the morning. One can imagine the scene as Roberto twanged and banged and bonked metallic objects as his wife Poldi acted as recording engineer. Both of them were in their sixties at that time.

Aside from sound sources recorded in his own studio, Gerhard also recycled fragments of recordings of his own instrumental works. Where the materials he needed could not be easily recording or created in his own studio Gerhard would resort to commercial sound catalogues or to outsourcing the recordings to a professional facility when a wider palette of instrumental sounds was needed. One such example is the music for the Royal Shakespeare Company’s performance of Pericles (1958) for which Gerhard produced the incidental music for ensemble and electronics. The box of tape 254 credits ‘Studio Black, Queens Way’ for the recording of percussion and exotic instruments. The multiplicity of sources from which Gerhard would obtain sounds included his close friend Joaquim Homs who provided the recordings of castanets which were required for the tape part of Symphony No. 3, ‘Collages’. Although Gerhard had a preference for sounds of acoustic origin, this did not rule out the occasional use of synthetic sounds, such as white noise or sine tones obtained from test and demonstration records or from sessions in the BBC Radiophonic Workshop. In the instances in which Gerhard required variable speed playback, the transformation would again be organised in an external facility, most often the BBC Radiophonic Workshop. In his notebooks, Gerhard used capital letters to identify the sound patterns that resulted from the combination of multiple sources as he developed his compositional materials. Such processes enabled Gerhard to mix sources at fixed or variable loudness to obtain more articulated sound images, and successively build up several strands up to ‘multilevel compounds’ ready for editing in the final composition.

In the second stage of the production process, Gerhard listened intently to the internal characteristics of his material, abstracting the sounds from their physical sources through various means of processing. During this stage of processing the primacy of the original sound as a means of grouping material developed from it became redundant as a means of classification. As Gerhard processed his material he regrouped it so that the timbral or gestural relationship between the sounds now assumed the most important means of classification. This processing stage allowed Gerhard to re-classifying the transformed sounds into sound-families – what Gerhard referred to as his ‘theory of change of family through sound mutation’ in which material is grouped together because of its similar sound behavior or timbre. Gerhard came to develop the idea of a genealogy of sound, stating that:

I have come more and more to believe that the overall sonic domain is perhaps not as vast and as diverse as one is a first inclined to assume. Rather does it seem a finite and bounded domain, and that in more aspects than that of frequency range alone. And I suspect that when acousticians take these matters up, it will probably be found that the number of existing ‘distinctive families’ of sound is not so inordinately large. What makes one think that this might indeed be so, is suggested by the fact that sound of a given family can be modified, by suitable operations, and made to resemble less and less the original sound from which we started. Gradually, it will adopt a novel character, but the degree of novelty that can be obtained is not unlimited. After a certain amount of change, there comes a moment when the sound simply begins to show characters of another, different, but already well-known family. In other words, it would seem that it is possible to develop new varieties, but now new families. The number of basic sound-families seems to be comparatively small. Most of them are probably already represented in the modern orchestra. If a section of tuning-forks were added – standing for spectrumless sine tone – the representation might be considered to be fairly complete.

From these sound-families Gerhard developed a series of clear compositional stages and his own terminology for each:

– small mixes Gerhard termed sound images or sound aggregates;
– these aggregates were mixed to form compounds;
– numerous compounds were mixed to form multilevel compounds;
– from these multilevel compounds the final assembly or sound montage would be mixed through editing.

The origins of this terminology can be found in his notebooks. Gerhard writes:

To compose and compound. To compose, in the sense of putting things together, in mere linear consecutiveness, or even in placing and spatlising [sic] a plurality o[f] events in more complex synchronicity is not all, it is indispensible, at the same time, to compound, i.e. to potentiate factors by settling differences and contradictions – by which is not meant that they should be ‘de-fused’ in the ballistic sense – but that their continuing struggle should be harnessed to the form-generating process, building up a manner of chain-reaction.

Here a further comparison with Maderna may be drawn. About electronic music, Maderna once said, ‘we no longer listen in linear time – our consciousness casts various projections of time that can no longer be represented with the logic of one dimension’. Working with electronic music made Maderna trust in his compositional intuition. The influence of electronic music in Maderna’s instrumental composition can be found in works such as the Serenata per un satellite. Gerhard himself wrote that ‘the way time is felt in electronic music differs entirely from the way time is experienced in traditional music.’ Gerhard was adamant that there is a fundamental difference between working with electronics and instruments. He uses the term sound-behaviour to characterize this difference. Gerhard writes,

[…] the operative word is behaviour, it will be noticed, not colour; colour is never of decisive importance. Instead of ‘behaviour’ I might have used the term sound-activity. The electronic medium, in effect, makes possible new modes of action with sound which have greater freedom of tonal movement, of configuration and of textural weaving than those which our traditional instruments permit.

Gerhard’s notion of sound-behaviour bears a close conceptual resemblance to what Denis Smalley would later term spectromorphology – literally the shaping of sound through time, an extension of Schaeffer’s typo-morphologie proposed in the Traité des objets musicaux (1966). Interestingly, these sound behaviours are never codified to the extent that they were by Schaeffer and later Smalley. Again, there is no abstract schema being formulated, merely the observations and thoughts of a practicing composer. What is clear, however, is that Gerhard considered these sound behaviours as directly contributing to the form and structuring of a work, writing that ‘wave-shape = prototype of form’. These new modes of action and of composing with sound contribute to what the composer termed the ‘temporal shaping’ of a work – one that provided the listener with an aural blueprint, which could be enhanced by repeated listenings. Gerhard writes:

I care enormously about shape, a telling shape, an apprehensible shape, a shape you could almost remember as shape, not the first time, to be sure, but after a time, after a number of times of listening to the piece, almost as you can remember a spatial sky-line, of town or hill – or mountain – range once you’ve become familiar with it; there is such a thing as a temporal sky-line, I believe, that’s what I mean when I say shape, only a temporal shape has got to be per-formed [sic].

In certain works, such as Symphony No. 3 ‘Collages’, the temporal shaping is extremely dynamic and highly profiled. In other works, such as Audiomobile 3 ‘Sculpture’, the temporal shaping is far less differentiated, but because the work is built on essentially one sound type – recordings of a sculpture made from brass rods created by John Youngman – the subtle differences are at once further metamorphoses of the sonic materials as well as a means of forwarding the musical argument.

In line with thinking in fields of sound-activity the electronic works are driven by gesture and texture led sections. Although Gerhard did not care for Schaeffer’s term for the basic perceptual unit in musique concrète, the objet sonore, it is clear that in his electronic works and increasingly in his later instrumental works, he nevertheless moved away from the ‘note’ as the essential unit, to his own notion of the sound object or sound-field as building blocks for his works.

Gerhard most prized intuition and imagination when working with magnetic tape, and, like Xenakis, he worked quickly and drew material from any source at his disposal when it suited his needs. As a result, there are sections of the Symphony No. 2 metamorphosed in Lament for the Death of a Bullfighter, and the same sonic materials shared between works: Audiomobile 3 ‘Sculpture’ and the final section of Symphony No. 3 ‘Collages’ use the same piano sounds; a keening vocal loop originally designated as ‘for the end of SCULPTURE’ was used in both Asylum Diary (1959) and in a variant form in Caligula (1961). For a composer known to destroy his sketches upon completion of the final score, this practice suggests a very different working philosophy. Part of this has to do with Gerhard’s notion of the metamorphosis of sound materials and their grouping into sound families. He wrote ‘nothing that instruments or the orchestra can do as well or better can be justified in the electronic medium. To be justified, both the sound-stuff and the way it is organized must be original growths of the medium’.

Following the initial recording of sound materials for use in a composition, Gerhard listened intently to the internal characteristics of his material, abstracting the sounds from their physical sources through various means of processing. The piano, percussion and the accordion were particularly favoured as source materials, as is evident in the number of tapes in the Gerhard Tape Collection in the Cambridge University Library that contain recordings of Gerhard and his wife Poldi making source sounds with these instruments for processing at a later date. Gerhard once said that there are more sounds in the piano than one can imagine, and utilized this instrument and sounds derived from it in his earliest sound composition, the Audiomobile ‘in the manner of Goya’ recorded for his Audiomobiles radio programme on the BBC Radio Third Programme in 1960. Gerhard was also aware that all instruments were not equally useful. In a notebook the composer observes that the processing of long wind notes, such as those of the flute and oboe, may result in awkward vibrato effects. This is one reason perhaps that he favoured the accordion (which Gerhard also considered a wind instrument) so much. Percussion and pedal glissandi on timpani also feature often in Gerhard’s sound compositions.