Oram's Wave Scanner

The glass slide

Drag to paint one cycle of the waveform. Solid ink below the line, clear above.

Timebase

The number of sweeps per second is the pitch: 440 sweeps a second gives 440 Hz.

Freq 220 Hz

Level 45

What comes out

The voltage that moves the spot up and down, which is the audio signal.

Scan rate220 Hz DC offset removed0.0% Harmonics in range100 Peak, after normalising1.00

How it works

A six-inch cathode ray tube sits in a lightproof box with a photographic slide clipped over its face. A flying spot sweeps left to right. A photomultiplier behind the slide detects light coming through it, and a feedback loop moves the spot up until it does. So the spot follows the edge of what you painted.

The voltage that moves the spot up and down is the audio signal. It is scaled down from several hundred volts to about one volt and sent to the amplifier.

filled in with solid black below the line and left transparent above it, looking rather like the silhouette of a mountain rangeGraham Wrench, who built it, describing Oram's slides

Filled, or outlined?

Switch between the two and the sound stays the same, because the edge is in the same place. What changes is what the photomultiplier looks for.

Filled: ink below, clear above. The spot moves up until it reaches the light and settles on the edge. Outline: the spot follows the dark line itself. Each needs a different feedback circuit.

The machine used both at once. Wrench describes the glass waveform slides as filled. But from 1968 the ten control films downstairs were read by their own spot-followers, so those graphs could be single traced lines. On 1 February 1972 Oram gave up the followers, and from then on every graph had to be inked in solid.

Caveat: “filled” for the glass slides rests on Wrench's memory alone. Richards never says so, and the follower circuit he describes, which he says is the same in both places, would not need it. The surviving slides would settle the question.

Sources

Tom Richards, Oramics: Precedents, Technology and Influence, PhD thesis, Goldsmiths, 2018; the mechanism is described on pp. 111–114. Graham Wrench interviewed in Sound on Sound. Peter Manning, Organised Sound 17(2), 2012.

The oscillator here is bandlimited, so what you hear has no aliasing. The 1965 original was not: the spot's slow return sweep added extra overtones, a fault recorded in the first prototype in 1963.