Recovering Sound from Old Recordings

Adrian Tuddenham, Member, on 22 February 2002

Adrian Tuddenham collected old recordings and developed improved ways of playing them. As a result, he became a sub-contractor to the National Sound Archives and now specialises in replaying artefacts which they could not otherwise play. He has built up a range of specialised equipment in order that he can control most of the parameters required in the mechanical aspects of playing discs and some additional electronic filters and adaptations of amplification to correct some of the electrical and acoustic distortions inherent in the recordings.

Before Edison, sound was recorded by scratching a path on a piece of blackened glass. Although this was useful to provide insights into the nature of sound, it could not be played back. Edison was the first to invent a way of playing his recordings back. These were indentations in a soft tinfoil cylinder. A hard sharp tip was used to make an indented track and a softer rounded tip followed it to produce the sound again. Even in the early days, electric motors were used in some of the recording machines to drive the rotation of the recording medium even though the recording process was mechanical.

Edison later moved to using a sapphire stylus to cut a wax cylinder. This type of stylus made a U-bottomed groove, and Edison used a `hill and dale' modulation. A major disadvantage with this technique is that the playback stylus had to be under considerable downwards pressure to keep it in contact with the groove and so the recordings were subject to heavy wear.

Emile Berliner realised that a disc-shaped recording would be easier to mass produce and would occupy less space than a cylinder. By using lateral modulation he was able to achieve greater modulation of the groove to give louder playback. This also requires less force on the stylus when playing back, although it led to problems of replay geometry that had to be overcome.

When Edison started using discs, he invented the phenolic resin now known as `Bakelite', independently of Baekeland. The two inventors made an agreement that Edison would use it only for sound recordings, and Baekeland would use it only for household items, thus preventing litigation for precedence and enabling both to pursue their businesses advantageously.

Berliner found that a replay stylus with a rounded tip located itself poorly in the groove and so hit on the idea of adding abrasive to the record material so as to wear a steel needle to the right shape to fit the poorly-controlled groove cross-section of his recordings. The needles were intentionally worn down and were made easily replaceable but this also increased wear on the record itself. The overall result was considered acceptable and, of course, ensured that records had to be replaced more frequently with a better commercial return. It could be said that Berliner put the scratch into records. In later years, when lighter weight pickups allowed the use of wooden or thorn styli, these would also be worn-down to fit the groove during playing but caused less damage to the record in the process.

In 1917 Paxman invented a cutting stylus that created a V-bottomed groove and again allowed accurate replay by means of a simple ball tip. The ball would then never reach the bottom of the groove and would stay in contact with both groove walls at all times. This replay tip could have been made non-sacrificial, but the noisy Berliner abrasive system was still championed by all except one enlightened manufacturer.

Electric recording was introduced in the mid 1920s, there was then gradual improvement, but no significant step change, in commercial recordings until the Second World War. During the war, the need arose for a better high frequency response, in order to make training discs for sonar operators. The Decca `Full Frequency Range Recording' system was a commercial spin-off from this work and was achieved by means of a massive increase in the electromagnetic field of the recording cutter head stator.

After the war, electrical amplification of the playback became commonplace and that opened up the market for long play discs, and finally stereo where two channels are recorded on the opposite sides of a Paxman-type groove, cut with each side at 45 degrees to the horizontal.

The next development for playing 78s with lightweight pickups was commercialised by W. Hodgson, whose styli had truncated elliptical tips. These would remain in contact with both sides of the older groove cross-sections and leave clearance at the bottom, so preventing dirt and groove damage from affecting the tracking of the tip.

Mr Tuddenham then demonstrated his equipment that separates the motion of the tip of a playing stylus along two orthogonal directions and shows the actual movement on the screen of an X-Y oscilloscope. By combining these orthogonal signals appropriately, the noise can be derived independently of the signal and then used to cut the individual noises out of the noisy signal by means of a quick-acting attenuator.

By watching the oscilloscope display, the effect of using different sizes of playing tips could also easily be seen.

As a sound is recorded, the cutting tip does not rotate as it is modulated, but keeps a constant angle to the direction of travel along the groove. This makes the actual recorded groove appear to vary in width on loud passages, like writing with a broad nibbed pen. When tracked with a rounded playback tip, this has the effect of causing it to rise and fall twice per cycle of recorded modulation. This unwanted vertical movement would be erroneously detected by the noise-reduction circuits and trigger inappropriate attenuation of the signal, which would then be heard as distortion. Extra circuitry has to be incorporated to prevent this happening.

Some recording systems made accurate alignment of the cutting tips in the recording head very difficult to achieve. These could give badly cut records that require adjustment of the angle of the playback tip to achieve the best sound quality. All of these adjustments were demonstrated by Mr Tuddenham's hand-built equipment.

Other noise-reducing tricks he showed us included wetting the surface of the disc, especially those made of cellulose nitrate. He also had a Pye magnetically recorded disc which he eventually played with a tape playback head filed down to a point to allow it to follow the groove in the record.

When he had been asked to try to play an old Dictabelt recording, which had been badly stored and so was brittle and pressed flat, he had to find a way of softening the belt without breaking it. He solved this problem by putting the Dictabelt machine on a hot plate and running it at just below the melting point of the plastic belt, manually controlling the temperature to ensure no damage occurred.

After the presentation, some of the audience had brought their own records to play and were very appreciative of the evening's entertainment.

Adrian Tuddenham