10th June 2003 - A Technical History of Recording, 1890-1980

This can only be a short precis of the wealth of knowledge and experience that Sean was able to share, accompanied by photographs and replayed examples stretching back to the earliest days of recorded sound.

Before the comparatively recent invention of recording there was only writing in place of speech and musical notation in place of recorded music. However, these notations of sounds are not very accurate records of how things actually sounded.

For a competent recording we need to provide an adequate frequency range and an adequate dynamic range between the system noise floor and excessive distortion.

For discs or cylinders, a constant velocity recording provides a flat response, where the velocity is measured as the stylus crosses the centre-line of the groove. This results in the groove amplitude being proportional to the reciprocal of frequency.

The next part is the soundbox. Although horns and diaphragms had been used for years as tools related to sound - think of drums and ear trumpets - this was the first application that required them to work together without contributing a character of their own. The horn matches the impedance of the air in the room with the mechanical impedance of the diaphragm. In an electrical analogy, a constant velocity recording is equivalent to a constant current source. Therefore, for recording, a high mechanical impedance is required in the stylus and diaphragm and a low impedance in the groove.

Edison liked the cylinder because the groove speed could easily be held constant. However, the vertical recording technique he used gave rise to practical problems. A recorded sine wave will be distorted asymmetrically; the shallower half cycle will be quite different to the deeper half cycle. Edison's machine used a lightweight stylus and diaphragm and a comparatively small horn. The LF performance was poor and, although the average speed of the groove was constant, the wow and flutter was typically appalling!

At the beginning, all recorded cylinders were original recordings but soon the volume of sales demanded a means to duplicate multiple copies from a single master recording, but this isn't easy with cylinders. A conventional split mould would produce a seam at the join that would be quite objectionable - imagine two hefty clicks per revolution of every recording you owned! Edison evolved a way to make a one-piece cylindrical mould of a master cylinder by electroplating. A type of wax was found which, after being cast into the mould, shrank just sufficiently on cooling that it released itself from the grooves to be withdrawn from the mould.

Spotting an opportunity, Emile Berliner patented a disc-shaped record. He reasoned that the flat shape should be easier to reproduce in quantity compared to Edison's cylinder. A suitable material, a mixture of shellac and various mineral fillers, was found at the Scranton Button Company in New Jersey.

Early masters were made using a zinc disc covered with a thin wax coating. The recording stylus cut through the wax and exposed the zinc which was then etched to form a groove using acid. The metal master was then copied to make a stamper which was then used to press discs for sale through the new Gramophone Company.

Later, discs were mastered by cutting the full groove directly into a wax blank. We heard a recording made in Milan which nicely showed an unfortunate side-effect of the chisel-shaped cutter, called the Pinch Effect. As the cutter moved sideways, it would cut a narrower groove than when it was still, like a calligrapher's italic pen. When this groove was later replayed with a spherical stylus tip, the groove-width variations would show up as a double-frequency component in the reproduced sound, particularly at high frequencies. The previous etching process did not show this effect.

The variation in level as the artist moved further from the horn was critical but did not entirely fit the inverse square law that we would expect. We heard a test recording of Nellie Melba singing scales at different distances from the recording horn, presumably to enable the technicians to establish a balance!

After the introduction of the triode, it took some time before this new technology was applied to the now mature business of mechanical sound recording. When the Service for the Unknown Soldier was recorded at Westminster Abbey in November, 1920, possibly the first electrical recording ever issued, the disc was cut using a Columbia lathe and electrical cutter head and also an array of multiple carbon microphones. The fidelity was limited but it was clear that the possibilities were enormous.

Other early mics were not impressively flat. For example, the Western Electric condenser mic had a 10dB resonance peak at 3kHz. Blumlein's first job for Columbia in 1929 was to fix a problem with response peaks in Columbia's moving-iron cutter. This used an energising-coil magnet. He introduced a special circuit that used rectified AC cutting current to modulate the energising current and hence reduce response peaks. Blumlein's own cutter design used electro-mechanical damping of a moving coil system to avoid the Western Electric patents on mechanical damping. The quest for high fidelity was not always apparent. For many years, EMI considered that an extended frequency range was dangerous and it was actively discouraged. EMI even used a choke across the microphone to limit excessive HF! EMI also persisted with a constant-velocity cutting characteristic, while Decca used a slight HF boost.

In 1931, Henry Olson introduced a ribbon microphone which provided a wide-range response to 15kHz. We heard a test piece of girls chatting plus a series of struck tone bars up to 10kHz, followed by an 'inflicted' harmonica solo, all reproduced from disc.

An alternative moving-coil cutter was developed by Paul Voight for Edison Bell. This design formed the basis for the Decca FFRR head. The development of the Decca FFRR disc cutter was in answer to a contract for the Admiralty to produce recordings for sonar training. The 6kHz limit, which was considered quite adequate prior to WWII, needed to be stretched substantially and Decca rose to the challenge. At about the same time, tests in the US were also exploring an extended frequency range.

By the mid 1930s, there were already many recipes for recording with high-frequency boost to improve subsequent reproduction quality. Standards don't always reflect the leading edge of technology, however. In the early 1950s, RCA introduced their "New Orthophonic" equalisation curve which was based on the use of pre-war RCA ribbon microphones. The equalisation curve was arguably inappropriate for condenser microphones and close mic techniques.

Through the RIAA, a 75 microsecond top lift record curve was standardised in the US. In Europe, Deutsche Grammofon initially adopted a 50 microsecond top lift curve - which was technically preferable - but, to avoid confusion in the marketplace, they later adopted the RIAA de facto international standard.

By the 1960s, engineers such as Joe Meek were experimenting with compression and artificial reverb, not to compensate for deficiencies of microphone technique but as a new type of creative tool. The days when recording quality was determined simply by microphone techniques and recording technology were numbered.

Mark Yonge