杏吧原创

ASK any movie fan about old films and they鈥檒l tell you that classic features are slowly turning to dust as the old cellulose nitrate film on which they were made crumbles away. Tell music enthusiasts, or radio or theatre buffs, that irreplaceable sound recordings are being lost in much the same way, and they will probably look blank. After all, Edison鈥檚 original cylinders and the 78s that replaced them were made of tougher stuff. Unless dropped and broken, they should last forever, shouldn鈥檛 they?

Unfortunately not. Sound archives around the world are in a desperate race to save valuable recordings before they fade away, and sink into a hissing sea of background noise.

The extent of the problem is huge. In Britain alone, where the British Library runs the National Sound Archive, vaults in London and Yorkshire hold 2.5 million recordings: more than a million CDs, vinyl LPs and shellac 78s; 3000 wax cylinders; several hundred thousand tapes of radio programmes, personal interviews and unpublished musical performances; and the largest collection of wildlife recordings in the world. The Smithsonian Institution in Washington DC has an equally important archive that includes recordings of native Americans made in the 1890s, ex-slaves interviewed in the 1930s, and a huge collection of traditional music from around the world.

The long-term aim of both the Smithsonian and the NSA is to make all their collections available online. But nothing can be done until the sounds have been converted to digital code and saved on disc. Time is running out. Just playing many of these recordings to copy them can ruin whatever sounds are left. And to cap it all, audio engineers are now discovering that even recordings laid down just a few years ago will soon be unplayable.

Conservators at the NSA and Smithsonian are being forced to classify the recordings in their archives according to how rapidly they are fading. Cylinder recordings made a century ago on Alexander Graham Bell鈥檚 Graphophone or Thomas Edison鈥檚 Phonograph come top of the risk list. The cylinders consist of a cylindrical core covered with a coating of soap-like wax that is soft enough to cut with the recording stylus but hard enough to survive when played back with another stylus.

Simply storing these things has been a headache. Over the years the wax can dry out and start to crumble, and conservators need to find temperature and humidity conditions that halt or at least slow this process. This is made all the harder because there was no standard formula for the coating: different companies used different waxes, often mixed to their own secret recipes. This means there is no set of conditions that is ideal for all of them. Another problem is that merely playing the cylinders can destroy them. Although modern pickups can minimise wear and tear by tracking the grooves very lightly, every play inevitably damages the wax a little bit more.

Most of the recordings are now in such poor shape that engineers at the NSA must do their best to clean up the sounds before they are re-recorded. Simply filtering off the hiss won鈥檛 do, as this can remove meaningful sound as well. There are computer systems that can distinguish noise from sound, but they are far from perfect. So engineers at the NSA have had to develop techniques of their own to try and unscramble the precious recordings.

One of them uses a specially designed stereo pickup to capture two separate recordings simultaneously, one from each side of the groove. A digital processor can then compare the two recordings and extract the wanted sound from random noise. It can even switch from one recording to the other, choosing the one that is of better quality. Switching too sharply can distort the recording, so the trick is to fade one channel out slowly as the other fades in. Engineers at the NSA are reluctant to say more as they are in the process of patenting the device.

They have revealed, however, that the technique will get its first big challenge later this year when Peter Copeland, who was the archive鈥檚 conservation manager until he retired in 2002, returns to his labs to try the stereo pickup on a wax cylinder reputed to hold the only known recording of Queen Victoria. In 1888, one of Bell鈥檚 European representatives demonstrated the Graphophone to Victoria in her Scottish home at Balmoral. A cylinder recording said to be of the queen鈥檚 voice turned up at the Science Museum in London in 1929. When the cylinder was eventually played, in 1991, one section was blank, another had male speech and whistling, and the third was clearly a female voice speaking in English. The sound was so faint and the surface noise so loud that only around 40 words could be distinguished, and most of those were unrecognisable.

Copeland called in a forensic expert skilled in analysing noisy surveillance tapes, but even he could make no sense of the words. Computer analysis did not help either. So now Copeland plans to try the stereo groove-switching technique. With luck, he says, the recording on the cylinder will survive long enough for its pedigree to be established.

Sticky stacks

Another group of seriously at-risk recordings are those made on the 鈥渁cetate鈥� discs that were used for personal recording in the 1930s and 1940s, before magnetic tape came along. Recordings were cut in a layer of cellulose nitrate coated onto an aluminium or glass disc. However, with time, the cellulose nitrate hardens as the oily softeners dry out, and its surface turns to white powder. In the worst cases the coating separates and starts to break up into small pieces.

Even on discs that are playable, noise can totally mask the recording. Things are slightly easier, however, in the rare cases where there are two copies of the same disc. When they are played in exact synchrony, the speech or music is the same from each source, making it far easier to pick out from the random background noise. The hard part is synchronising the two recordings. In the past the two discs had to be played on identical turntables, locked together one above the other on a structure resembling a rotating cake stand. A less cumbersome way of synchronising the two tracks is to re-record the discs digitally and then use characteristic peaks in the music, such as loud percussion, as markers to lock the two recordings together as they are played back.

One of the most technically challenging projects Copeland has faced was an attempt to recover broadcasts from Germany by William Joyce, alias Lord Haw Haw, made during the Second World War. BBC listening stations recorded Joyce鈥檚 propaganda on 30-centimetre discs made from a mix of cellulose nitrate and lacquer on an aluminium base. After the war some 50 discs were stored at the Public Records Office in London where they were stacked in piles. Unfortunately the warm surroundings made the surface of the cellulose soften so that the discs are now stuck tightly together.

Copeland鈥檚 first idea for separating them was to squirt water jets between the discs to push them apart without tearing off the outer coating. This didn鈥檛 work. So his plan is now to let water seep into the cracks by capillary action, and then repeatedly freeze and thaw it. The idea is that the expansion of the ice when it freezes should push the plates apart, allowing slightly more water in when it melts which pushes them further apart next time it freezes. Unfortunately the project is now on hold while the NSA鈥檚 audio archaeologists look for an 鈥渁nti-crystallisation agent鈥�, a chemical that they can add to the water and which will help the ice to expand and contract smoothly, hopefully minimising damage to the grooves. Meanwhile, Copeland is not banking on success. 鈥淚 estimate a chance of only about one in a hundred that this will work,鈥� he says.

Copeland has already managed to extract some of the recordings 鈥� by balancing the huge piles of discs one at a time on a strengthened turntable. This let him play the sides on the outer surfaces of the stacks. And at least the problem won鈥檛 get any worse if the Haw Haw recordings have to sit for a few more years while Copeland and his colleagues work out how to split the plates apart. Since the surfaces of the discs are stuck together, the grooves inside should be protected against atmospheric attack.

A potentially far worse problem is with some magnetic audio tape that has been in use since the 1970s. The chemicals that bind the magnetic oxide to the plastic base go soft over time, and in tapes made by a few leading manufacturers the binder begins to ooze out. Even if the tape somehow holds together, this goo sticks the turns of tape to on another on the reel, and when you try to play it the tape machine grinds to a halt. The recording industry calls it sticky tape syndrome.

There is an answer, but it is a nail-biting technique. The sticky roll of tape is put in an oven and baked for up to 8 hours at 55 掳C. This melts the goo and seals the magnetic oxide onto the base. After cooling, the tape can be played once to make a copy 鈥� but there is usually no second chance. The process makes the tape brittle, and if it鈥檚 played again it cracks.

After numerous experiments, conservators at the NSA have invented another trick that helps speed up this lengthy process. The spools of sticky tape are mounted on a tall machine, dubbed the grandfather clock, which slowly unwinds the tape and runs it past a hot air blower before feeding it into a tape player. The technology has already proved its worth in recovering the historic recording made in South Africa in 1964 when Nelson Mandela was on trial in Pretoria鈥檚 Palace of Justice. Mandela鈥檚 impassioned speech from the dock just before he was sentenced to life imprisonment was recorded on plastic belts of the kind then used in office dictation machines. When Apartheid ended, the National Archives of South Africa passed the belts to the NSA, but they were found to be too rigid to play. So conservators at the NSA took a gamble. They heated the belts just enough to make them flexible, and played them. The recordings they retrieved are now back in South Africa and awaiting release.

All may not be well even on tapes that superficially appear perfectly solid. On some, the recordings are simply disappearing. In the 1970s, recording engineers realised that it was possible to get superb quality recordings by converting sound into digital code and storing it on Betamax videotape. This was widely used to record classical music, and live concerts and theatre performances. The National Theatre and Royal Court Theatre in London used the technique, and around the world there are hundreds of thousands of these Betamax recordings, safely stored for posterity.

But last year research at the Biblioth猫que Nationale in France brought the bad news that the signal on these tapes is slowly degrading. With video pictures this isn鈥檛 a problem, as the deterioration is so slow its effects are barely noticeable. With digital audio recordings, however, the result is far more serious. Once the signal fades below the level of noise, the recordings are lost. The British Library, with 10,000 Betamax digital audio cassettes holding 25,000 hours of theatre and music history, has conducted its own study and found the recordings are fading faster even than the French research suggested, predicting the signal level will halve in 10 years.

Going, going, gone

The British Library is now spending 拢150,000 a year on transferring its Betamax sound recordings onto other media before the signal on the tape corrupts too badly to replay. But the Betamax video players are wearing out. Worse, the supply of new Betamax machines and spares for the old ones has completely dried up. No players have been made in Europe for 10 years and Sony鈥檚 factory in China has just stopped production. When existing Betamax players stop working, the tapes will be useless.

As if conservators such as Copeland don鈥檛 already have enough to do, audio engineers have now discovered that even relatively new digital audio tapes are self-destructing. Last year, engineers at Warner Brothers records went into their vaults to pull out the treasured original master tapes for classic recordings to reissue them on a new high-quality DVD format, and were horrified to find that although stored carefully many were suffering from sticky tape syndrome. Those affected include The Eagles鈥� Hotel California, the Doobie Brothers鈥� The Captain and Me and The Band鈥檚 The Last Waltz. Even the master recording of REM鈥檚 Automatic for the People, just 10 years old, needed baking before it would run smoothly through the playback recorder. There are probably many thousands of similar master tapes in record company vaults all around the world suffering from the same problem.

There is no easy solution. At the NSA, only four people are working full-time on rescuing the past, and Copeland鈥檚 former post of conservation manager has not yet been filled. Things are little better in the US. Last year, Congress announced a grant of $750,000 to help the Smithsonian and the Library of Congress with their preservation programme. At the same time, musicians such as Pete Seeger, Ry Cooder, Bonnie Raitt and Carlos Santana are backing the call for more money.

But even if the current rescue attempt succeeds, we might need to launch yet another. There鈥檚 already evidence that compact discs 鈥� the favourite recording media of the moment 鈥� aren鈥檛 as permanent as archivists had hoped. Their aims are honourable, but today鈥檚 conservators might simply be laying down problems for their descendants.