The Oramics Machine

Almost every electronic music studio of the period worked by recording sounds onto tape and then cutting and rearranging the tape, so the work lived in the editing. Oram wanted to write sound down directly, the way a composer writes a score, and have an instrument play what she had written. She spent roughly fifteen years and two grants getting there.

This page covers who she was, how the machine came to exist, what it could and could not do, and what was genuinely new about it. Four companion guides go into the mechanisms in detail, each with the live diagram from the workshop deck.

Before the machine

Daphne Oram was born on 31 December 1925 in Devizes, Wiltshire, and studied piano, organ and composition at Sherborne School for Girls. In 1942, aged sixteen, she turned down a place at the Royal College of Music and joined the BBC as a junior studio engineer instead. During the war she worked at the Royal Albert Hall on broadcasts of the Proms, looking after the microphones. Recordings of the same music were kept running in case bombing forced the orchestra to stop, and her job was to switch from the live orchestra to the recording without listeners noticing. She spent years getting live and recorded sound to match closely, which is worth knowing about someone whose later work combined the two.

At twenty-three she wrote Still Point, a piece of around half an hour for double orchestra, five microphones and three 78rpm records whose sound was altered live during the performance. It is often described as the first work to combine an orchestra with live electronic treatment of recorded sound, a claim that is reasonable without being settled. When it was put forward for the Prix Italia, BBC management called it a directionless mess, and it was not performed in her lifetime. James Bulley and Shiva Feshareki reconstructed it from roughly nine hundred pages of unordered manuscript, and it was first performed on 24 June 2016 at St John's Smith Square in London. In 2018 it was played at the Proms, in the same hall where she had worked during the war.

The Radiophonic Workshop, and why she left it

By the mid-1950s broadcasters in Paris, Cologne and Milan had set up studios where composers could make electronic music, and Paris and Cologne had taken opposite positions about how. Pierre Schaeffer's studio in Paris built music from recordings of real sounds and called it musique concrète; the Cologne studio under Herbert Eimert built it from electronically generated tones starting with sine waves, and called it elektronische Musik. The two camps argued about it openly.

The BBC founded something else. The Radiophonic Workshop opened on 1 April 1958 in rooms 13 and 14 of the Maida Vale studios in London, a building that had once been an ice rink, and its purpose was to make sound effects and background sound for radio drama. The word radiophonic was itself a way of not calling the work music. That difference in purpose explains most of what went wrong for Oram inside the Corporation.

She is often described as the Workshop's founder. She and Desmond Briscoe, both BBC studio managers, had lobbied for it for years and became its first staff, with Dick Mills as technical assistant. But the committee that planned it in 1956 was led by two drama producers, Douglas Cleverdon and Donald McWhinnie, and no studio manager was invited onto it, although her work was discussed at its meetings. The founding budget was £2,000, two professional tape recorders, a basement echo chamber with one speaker and one microphone in it, and the key to the store where the BBC kept its redundant plant. More equipment came out of bins. The main sound sources were test oscillators and a wobbulator, a device for sweeping through a range of frequencies, both built for testing radio receivers rather than for music.

Five months after the Workshop opened she was told she would be moved off it for eight weeks. A doctor had advised a BBC board member that three months was the longest anyone should work there, and the policy was real: Briscoe was moved too, and a BBC document describes it as avoiding battle fatigue. Oram did not accept the reason. She wrote to her mother Ida on 31 August 1958:

I'm being taken off Radiophonics from Sept 8th - Nov 1st for a 'rest'. I'm furious. Desmond who works with me has also been told that he will be taken away on Nov 1st. He's furious too. I think they have a policy not to let either of us get important, have any publicity, or get the running of things in our hands.

Oram to her mother Ida, 31 August 1958

Tom Richards, whose 2018 Goldsmiths thesis supplies most of the dates and documents on this page, found nothing in the BBC archive that confirms her reading of it. The evidence does not settle the question either way.

There was also the matter of credit. In 1957 she wrote the score for a BBC production of the play Amphitryon 38 using sine-wave oscillators, tape and filters she designed herself, the first score in BBC history made entirely from electronic sound. On 7 October that year the Third Programme broadcast Private Dreams and Public Nightmares, whose listing said that every sound in it was new and specially made, and credited the recording engineers of the BBC without naming her. Val Gielgud, the head of drama, asked for her to be credited and was overruled by senior management. Alec Nisbett, a colleague who supported her, said management saw her as having ideas above her station. Her proposals for drawing sound were rejected.

She left in January 1959, after fifteen years. Richards gives three reasons, in order: the equipment was second hand and second rate and her proposals were ignored; she had to fight to be credited; and she wanted to compose, which at the BBC her work could not be called. He thinks the last was probably the most important.

Tower Folly

She set up Oramics Studios for Electronic Composition at Tower Folly, a converted oast house at Fairseat near Wrotham in Kent. An oast house is a building for drying hops. In January 1962 the Calouste Gulbenkian Foundation gave her £3,500 over three years to fund the research, about half what she had asked for and unusual in that the foundation rarely funded individuals at all. It added a further £1,000 in October 1965.

The rest she paid for with commercial work, most of it uncredited: music for Jack Clayton's film The Innocents in 1961, electronic sounds for Dr. No in 1962, and a great deal of advertising. None of that was made on the Oramics machine, which did not work until 1966.

Her brother John Oram built the mechanical transport. Graham Wrench, an engineer, built much of the electronics including the wave scanner between 1964 and the summer of 1966, and described the work years later in an interview with Sound on Sound. Fred Wood, a Post Office engineer, worked on the electronics at weekends and copied Wrench's circuits after Wrench left the project.

Following the signal through the studio

The diagram below is the studio as the signal moves through it, from drawing on the film to sound leaving the building. Step through the six stages with the arrow keys or the controls, and switch between the 1968 machine and the 1972 one to see the line-following circuits replaced by plain lamps.

Two floors is the only documented fact about the layout: the wave scanners were upstairs and the film programmer downstairs. Everything else in the plan, including the stairs and the echo room, is arranged for clarity rather than from a record. The echo room was real, though: the reverberation was a room with a loudspeaker at one end and a microphone at the other.

Why draw by hand

In January 1961, before any of the machine existed, Oram set out the argument in a notebook titled "Written Sound Waves":

Neither the additive sine wave methods nor the subtractive white noise methods give satisfactory 'musical' timbre because the results are, I think, too simple. There is not enough change within the note itself. Also they are too mathematically pure & calculated. A freehand drawing method might produce very much more artistic results because of the inaccuracies inherent within it.

Additive and subtractive were the two standard methods in the studios of the time: building a sound up from sine waves, or filtering noise down. Her objection to both was that they were too clean, and the inaccuracy of a drawn line was the point rather than a defect to be engineered out.

The machine was then built to tolerate that inaccuracy. Each note's pitch was written as marks across four film strips, and because they were placed by hand they never lined up exactly. Wrench added a circuit that waits 10, 20 or 30 milliseconds after seeing the first mark before reading the rest. With the film moving at about three inches a second, thirty milliseconds allows the marks to be roughly 2.3mm out of line and still be read as one note. That circuit is still in the machine. How the machine read a pitch covers the coding in detail, with a diagram you can write a tune on.

What it could not do

The machine first worked in the summer of 1966, and the list of its limits is long enough to be worth stating plainly. It played one note at a time. It drifted out of tune, because the mains voltage at Tower Folly varied; in March 1969 the tuning shifted when her electric radiator switched on. It held about forty seconds of music at the outside, and to get that much she extended the film loop around a broom handle in January 1972, which is now in the Science Museum's collection alongside the machine. Switching it on took twenty-one steps across six power supplies on two floors, and getting step nine wrong would burn the screen of a cathode ray tube. It was effectively impossible to move, and it was never reliable enough to demonstrate to an investor.

To make anything longer than a single line she recorded onto tape and overdubbed, like everyone else.

The failure that ended it came from the mains. The loudness graphs were read by small cathode ray tubes whose spot of light followed a drawn line, and those circuits needed a steady supply. On 1 February 1972 she wrote in her logbook:

Mains down to −13. Volume tracks hopeless. Am giving up CRT spot following volume tracks in despair - absolutely no use when mains fluctuates like this.

Oram's logbook, 1 February 1972

Two days later she took the circuits out and fitted ordinary lamps and photocells in their place. From then on every graph had to be filled in solid with ink or tape rather than drawn as a line, which took longer and was harder to correct. Her logbooks stop soon afterwards, and the last known recording from the machine dates from 1976. The line-following circuits were Richards' discovery; no earlier account of the machine mentions them.

What was actually new

Oram applied for patents in the United Kingdom, the United States and Japan. The American and Japanese applications were rejected on the basis of earlier inventions, one of them an oscilloscope-based function generator that David Sunstein had described in the magazine Electronics in February 1949, working the same way as her wave scanner. A photocopy of that article was in her own files.

She cut the American application down to the one part that was new, controlling pitch digitally by reading marks on film, and that was granted in November 1969. The British patent, granted in April 1970, covered the whole system. The Japanese application was never granted. She had followed the avant-garde studios closely, but the technology closest to hers was being developed somewhere else entirely, by organ manufacturers and in laboratories that recorded earthquakes.

So almost every component of the Oramics machine had been invented before. What had not existed was the combination: one instrument you composed on by drawing. Richards puts it as "not an under-appreciated musical invention, but an alternative conceptual innovation", a system design worth more than the sum of its parts.

Afterwards

Richards looked for later instruments that descend from the Oramics machine and reports that direct evolutional influences are difficult to discern, and that none have been found. The machine is often compared to the automation lanes in modern recording software, and people who used the working Mini Oramics he built in 2016 made that comparison unprompted. But that software was developed independently, and the resemblance is not a line of descent. Oramics belongs with an older family of instruments that play from a written record, such as the barrel organ and the player piano. The Oram Trust's own website calls the machine a forerunner of MIDI sequencing and the digital audio workstation, which describes what it resembles rather than what came from it.

Her influence ran through teaching instead. Large numbers of people first learned about electronic music from her lectures and seminars, her work at the BBC, and her 1972 book An Individual Note of Music, Sound and Electronics. She wrote an experimental syllabus for teaching electronic music in schools, and planned a smaller instrument, Mini Oramics, for school and university studios, whose mechanical parts were to be built by the engineering workshop at Sherborne School for Girls, where she had been a pupil. She worked on it in the early 1970s with the engineers Norman Gaythorpe and John Emmett and never finished it. Richards built a working one from her notes and drawings in 2016.

The machine itself is in the collection of the Science Museum in London.

Sources

Most of the dates, quotations and technical detail come from Tom Richards, Oramics: Precedents, Technology and Influence, PhD thesis, Goldsmiths, University of London, 2018, which worked through Oram's archive and the BBC's records and examined the machine itself. Graham Wrench described building the electronics in an interview with Sound on Sound. Peter Manning's article on the machine appeared in Organised Sound 17(2) in 2012. The quotations from Oram come from her logbooks, notebooks and letters in the Daphne Oram Archive at Goldsmiths.

This series was written for Drawing Sound, a workshop on Oram and the machine, and the diagrams are the ones used in it.