Mix Rescue: Russell Simon

No, it's not the Naked Chef, just SOS getting busy in the kitchen to salvage some old tape recordings...

This month's mix rescue is a little different from the norm. With not so much as a CD in sight (never mind an MP3!), I was approached by guitarist Russell Simon, who had a handful of quarter-inch tapes he'd made with rock bands he'd played in during the late '70s and early '80s. He wanted them transferred to CD before the tapes became too old to be salvageable, and he wondered if I still possessed anything so antiquated as a tape recorder that was capable of doing the job.

Fortunately, I still keep my old Tascam 32 quarter-inch machine in the studio for such jobs, so I said I'd give it a try. However, as you may already know, older tapes can sometimes present problems when you come to play them.

Cooking up a feast: tapes suffering from 'sticky shed' are placed in boxes in the oven. The oven has to have very good temperature control and the average kitchen model will not be up to the job. If you have a plate-warming section like this oven, then you can keep the temperature more stable.

Different Breeds

There are two main types of tape. The older one is based on an acetate backing film and the more recent type uses a polyester film. You can often recognise acetate tape by holding the reel up to the light so you are looking at the wound tape edge on. If you can see the light passing through the tape from edge to edge, it is probably acetate, as the film material itself is fairly transparent. Polyester, on the other hand, is opaque, so you won't see light through it.

Acetate tapes were commonly used in the '50s and '60s, so you don't often come across them these days, unless you deal with archive material. If you find an acetate tape, try sniffing it, and if you detect a whiff of vinegar (acetic acid), that's a pretty sure sign the tape is breaking down chemically. As the tape ages, it actually loses weight and also tends to curl, so to make it play adequately well, it is often necessary to improvise some way of increasing the tape tension in order to ensure good head contact. Fortunately, acetate tape doesn't tend to stretch until it reaches breaking point, so you can use quite a lot of tension on it without deforming it.

Sticky Shed

Polyester tape was developed as a superior replacement for acetate, but I first encountered a problem with it in the 1980s. I had no idea what was going on, and nobody else at the time seemed to know either. It is now well known that certain brands of tape dating back to the '70s and '80s suffered from a problem known as 'sticky shed syndrome'. It is most often associated with Ampex 456 back-coated tape from those decades but user groups have also reported similar findings with certain old Audiotape, Agfa and Scotch/3M formulations.

Part of the cause of sticky shed syndrome is that the polyurethane material binding the oxide particles absorbs water, a process that occurs over a period or months or years, depending on the tape formulation and storage conditions. What happens is that when you try to play the tape, oxide sheds onto the tape heads very rapidly and, unlike normal oxide shedding, this becomes sticky and impedes the passage of the tape over the heads and tape guides, often causing audible squealing. Sometimes it may even cause the tape machine to slow down or stop entirely. Playing a tape in this condition damages it further by dragging away part of the oxide layer, so at the first sign of sticky shed you should definitely stop playing the tape.

A Reel Wind-up

To find out if Russell's tapes were suffering from sticky shed syndrome (as I suspected they might be, given their age and origin), I wound them directly from one reel of the Tascam 32 to the other, threading the tape to bypass the tape guides. This meant holding the tape tension arm with my finger to keep the machine running, as the angle of this arm senses the absence of tape, using a microswitch to stop the motors when the tape reaches the end.

Bypassing all the tape guides and heads prevents the vulnerable oxide coating from coming into contact with anything other than the back of the tape as it winds. If the tape unwinds cleanly, it may well be alright to try playing it, but if it is suffering from sticky shed syndrome you'll be able to see and hear it peeling away from other layers on the reel as it winds. Some people also allow the oxide face of the tape to run over their finger to see if any residue builds up, though I've never tried this approach myself.

Even though Russell's tapes had been stored at home in dry conditions, the winding test confirmed that there was a degree of stickiness that needed addressing.

The Tascam 32 quarter-inch tape machine that was used for this salvage job.

Bakers Beware!

There are two causes of sticking tape: binder degradation, due to water absorption, and a loss of lubricant from the tape formulation. Binder degradation can be reversed temporarily by baking the tapes, but it cannot help with loss of lubricant. So far, all the problem tapes I've encountered have been made playable by baking, so it seems the binder breakdown problem has been the predominant cause. It also is worth noting that baking should only be used on back-coated polyester tapes, as it may actually cause further degradation in other types. The process is a temporary one, as the tape binder will start to absorb water from the atmosphere again once you've dried it out. It is essential to transfer the recording to a new medium as soon as is practical after baking: under normal circumstances, the benefits of baking last for around two to four weeks.

If the tapes are of historically important sessions, or have commercial value, you should really get the baking done professionally. However, if they are simply old demos that you'd like to transfer for nostalgic reasons, you might want to attempt it yourself. The recommendation for quarter-inch tapes is that you bake them at around 50 degrees Celsius for between four and eight hours. It is important that the temperature remains fairly stable, which means that a domestic baking oven is not appropriate. due to inaccuracies in the thermostat at the lower end of the scale. The first time I tried baking tapes, I borrowed an electric egg incubator, which worked fine, but for this current project I used the plate-warming oven in my kitchen stove, as this is designed to operate at temperatures around that at which tape baking is carried out, and it has its own thermostat, separate from that of the main baking oven. Ideally, you should use a thermometer to check that the temperature is correct to within a couple of degrees. The type available from a photographic store should be fine for this purpose. Alternatively, you can do what I did and put some plates in the oven, guess the temperature, then adjust it until you can just hold the plates without burning your fingers (this can use up a lot of plates!). You should only use electric ovens, as gas produces water vapour as it burns and that's exactly what you're trying to drive out! It is important to get the temperature as close to 50 degrees Celsius as you can, as if it is too cool the process will be ineffective, whereas too high a temperature will encourage more magnetic print-through, which can result in an audible pre-echo at the start of songs and during quiet sections.

To keep the temperature more stable, I left the tapes in their cardboard boxes (but removed the polythene bags that were in the boxes) and stacked them with an air gap of around an inch between each tape. I decided to leave the tape in the oven overnight, which meant that they were cooked for around eight or nine hours, after which they were allowed to cool naturally before I attempted to play them. It is beneficial to leave the tapes for a few hours after baking, so that the lubricants in the oxide layer can find their way to the surface. Steady cooling can easily be achieved by simply switching off the oven and then taking the tapes out after the oven has cooled.

The next process was to wind the tapes completely through, to separate the windings and to check for splices that may have been falling apart. Any suspect splices should be remade using proper splicing tape, a simple splicing block and a single-edged razor blade.

Soundsoap Pro is good tool for cleaning up old audio tracks, and was used here to remove low-frequency amp buzz.Having wound the tapes and ensured that they were the right way round, the tapes were ready to play. To play them back, your tape machine must run at the same speed as the original tapes were recorded at and the track layout must be the same. Most serious stereo machines were two-track, with no 'other side' to the tape, whereas many consumer machines used a four-track layout, with one stereo pair of tracks running in one direction and one in the other, so that by turning the tape over, you could double the playing time. As each track on the consumer systems was half the width of those on the simpler two-track layout, the noise performance of the consumer machines was slightly worse. It is important to remember that you can't play back a tape made on a four-track stereo machine on a two-track stereo machine, as you will hear both 'sides' at once, one playing forwards and the other playing backwards.

There's also the question of the equalisation used by the recorder (NAB or IEC) when the tape was made, and if it was made in a pro studio, there may also have been some form of noise-reduction system used, such as Dolby A. There's no accurate way of fudging the noise-reduction system, so you need to hire in a compatible system or have the tape transferred professionally, which may be the most cost-effective option. The equalisation part of the equation is less of a problem, as the subjective difference between the two is fairly subtle — NAB and IEC are essentially two different EQ compensation curves applied during recording and playback.

This EQ system maximises high-frequency headroom and also takes account of the non-linear way in which electrical signal levels relate to magnetic flux intensity within the record/playback chain in order to produce a linear frequency response. If you have a machine with switchable NAB or IEC EQ, such as the old Revox A77 I had back in the '70s, and if the engineer noted the type of EQ on the box, all you need do is switch to the right one. If you can't switch EQ types or you don't know which one was being used on the recordings, then you'll have to use conventional EQ to get the sound as close to how you'd like to hear it as you can. This was the position in which I found myself, as my Tascam machine uses a fixed EQ system and there was no info on the original tape boxes other than tape speed, which was 15ips (inches per second) in this case.

My procedure was simply to patch the tape recorder into my interface, then record the result into my computer at 24-bit resolution, ensuring that I achieved healthy peak recording levels without risking clipping on any of the loud sections. As Russell wanted to make a few CDs from the final recording, I stuck to a 44.1kHz sample rate. Whatever you do subsequently, it is essential to keep an unprocessed backup of those original 24-bit files, so that you can visit them again if better restoration technology becomes available in the future.

Restoration

If you're rescuing recordings with a high commercial value, you may want to have them professionally restored using high end systems, such as those made by Cedar Audio, as well as professionally baking and transferring the original recording tapes (a service offered by FX Copyroom), but for the majority of us living in a more frugal world, there's a lot that can be done using more affordable software plug-ins or outboard processors.

Tape hiss can be reduced to a useful degree using de-noising software of the type that takes a noise fingerprint from material just before or just after the recording, although, if the engineer has been really tight with his splices, you may not find a long enough piece of noise to sample. Typically, you'll need at least half a second of isolated noise. If you don't have an isolated noise sample, you can try playing a blank piece of new tape and taking a noise fingerprint from that, but in my experience there's usually a lot more noise added by the recording process and by noisy gear than by the tape itself. The other main tool is a good equaliser, as the tonality of old recordings is often less than ideal and it is often possible to make a considerable subjective improvement using a basic parametric EQ.

Noise and tape hiss can usefully be reduced by denoising software (Waves X-Noise and Steinberg's Denoiser from Wavelab 6 are shown here). The best results can be achieved using a plug-in that can analyse the noise, and you typically need about half a second of isolated noise, without music, to obtain a good noise 'fingerprint'.With Russell's tapes, which contained mainly high-energy rock music, tape noise wasn't a significant problem, though some guitar amplifier noise and buzz was audible during exposed sections, especially on the intro of the piece we chose to use as an example. Amp buzz can be tricky to fix, and dedicated software that notches out the fundamental hum/buzz frequency along with a series of harmonics is often most effective. You can, however, do this manually, using a series of parametric EQs set to create narrow, deep notches at 50Hz, 100Hz, 150Hz, 200Hz and so on for 50Hz European mains frequencies, or multiples of 60Hz for US records. You may also have to move these frequencies up or down slightly if the tape speed between record and playback is slightly different. I settled on using the hum-filtering section of BIAS' Soundsoap Pro, setting the lowest frequency to 150Hz and then engaging all the harmonics above that. This, in combination with silencing the intro right up until the first note, made an adequate subjective improvement, so I decided that broadband denoising would not be necessary.

Little was needed in the way of EQ, so I tried the Focusrite Liquid Mix, running its 'huge analogue' equaliser set to give a gentle mid-range dip augmented by a hint of boost at 90Hz and 10kHz. This created a very subtle smile curve that added a sense of loudness and clarity without changing the overall feel too much.

The song clearly hadn't been mastered, so I opted to give it a bit more density using the Bus Compressor from SSL's Duende system. Although Liquid Mix offers what appears to be a setting based on the same device, I chose the SSL version partly for variety, and partly because the SSL version isn't an emulation, as it is based on the same code used in SSL's own digital consoles. With a minimum ratio of 2:1, this can be a bit fierce as a buss compressor, if you don't keep an eye on the gain-reduction meter, so I tweaked the threshold to give a gain reduction of three to four dBs on peaks, using a 0.3ms attack and the auto release setting.

Final limiting was done using the TC Powercore Brick Wall limiter, which worked exceptionally well, and allowed me to trim a couple of dBs off the peaks without compromising the sound. Comparing the before and after files, showed that the second appeared much louder, even though the peak levels are actually half a decibel less.

TC Electronics' Brick Wall limiter was used to trim the peaks without compromising the sound.

As an experiment, I also tried to process the whole mix using a little Altiverb 'Wooden Room' reverb with the low end rolled out, as I felt the original recording sounded a little as though it was still locked in the studio in which it was recorded. I felt it needed a sense of a 'bigger' performance space, and I managed to achieve this with Altiverb without making the recording seem processed. In this case Altiverb went directly before the limiter, but it could also have come earlier in the chain.

At the end of the session I felt that the processing had brought these old demos more up to date and the apparent extra level makes them more comparable with commercial records, which is a consideration if Russell intends to listen to them on his MP3 player or to play them alongside other commercial CDs. At the same time, the original character of the recordings was preserved, but with just a hint of modern polish. Russell said he was pleased I'd been able to rescue his recordings at all, so the processing was just a bit of icing on the cake!

Summary

Tape baking isn't something you should do yourself, unless you're working on tapes that you can afford to lose. That said, unless you're very unlucky you shouldn't have any problems, and it is a good way to get those old demos onto CD or hard drive before the original tapes deteriorate too badly. Once you have copied the material into a DAW, you can use a variety of restoration tools to improve the subjective quality, but you should keep backups of the original files so you can try out new restoration software as it becomes available. Subtle compression, EQ and limiting may be all you need, but there may also be occasions when denoising and de-buzzing software can help.

If you can make direct comparisons with commercial material in the same style and played back at the same subjective level when working, it will help you focus on what needs doing and will help prevent you from overprocessing the material. Russell was glad to be able to hear his music again after so long, and I learned some valuable lessons from the process of improving the sound without doing anything too radical to it.