At the 2014 BISA (British and Ireland Sound Archives) Conference I was very fortunate to run into Will Prentice, the head of technical services in the sound and vision department at the British Library. Will deals with preservation of sound and moving image and had plenty of useful info about the task of digitising the audio collections from Arthur Tapp and the New York Radio Festivals that Birmingham City University has acquired (discussed in this earlier post). I managed to grab 10 minutes with him at the end of the conference to speak about the best way to approach the archiving of audiocassettes, and here’s a transcription of that discussion.
Will also provided some useful links – which I’ve included at the end of this post.
How do you get the best out of cassette playback?
“Relative to other formats, which is a handy way to talk about them, (cassettes) are pretty robust and they’re fairly straightforward to play. There are set parameters that you need to handle. The worst, or the simplest case scenario is that you stuff it in the machine and you press play and you’ll hear stuff, it’ll be okay. Two main things you want to adjust to get the best out of them: one is make sure you’ve got the right noise reduction setting. So if it was recorded with Dolby, Dolby B usually, maybe Dolby C, you want to switch on Dolby B on the playback device, if it has that. If it wasn’t that you don’t want to use that. So, that makes a significant difference to how it sounds. If you want it to sound as it did when it went on the tape, you want to use the right noise reduction, or, not use it if it wasn’t used. The other thing is you want to make sure you’ve got the azimuth, which is the angle of the replay head, adjusted correctly. Most machines, certainly professional machines, will have a little screw, a little cross head screw, that you can adjust below the head which will allow you to do that. You can do it by ear, pretty straightforward. It sounds intimidating and it sounds a bit obscure but if you have someone demonstrate it for you, they’ll teach you in about a minute and a half. It’ll be perfectly straightforward and you’ll be amazed by the results. So, I would strongly recommend that people find a machine that is semi-professional, shall we say, that allows them to adjust azimuth and noise reduction.”
“There are certain things that can confuse that though, one is – it may be muffled for other reasons, it might be an nth generation copy of a cassette, in which case the damage may have happened in an earlier generation. It may be that it was made on a poor quality machine which didn’t capture high frequencies, it may be that the record heads were dirty when it was recorded, or the replay heads were dirty when it’s being played back, so those kind of things can affect it – but what you would do is stick the tape in, listen to it, does it sound a bit muffled, doesn’t it? Adjust the azimuth, are you making it better or worse? Tweak it left and right – until you have as much high frequency information as possible. Job done, 20 seconds.”
How stable are cassettes as a format?
“They’re fairly stable, so, in terms of, relative to other carriers, they are robust and probably long lasting, so I don’t expect them to degrade in a hurry. Notionally, they can become sticky, some brands will become sticky, but I’ve never come across that in the UK. I’ve heard North Americans talking about it occasionally, and it certainly happens with certain brands of quarter reel tape, but I’ve never really seen it in cassettes. Cheap ones can have all kinds of friction issues inside, you know, you stick it in and the tape won’t turn round properly or it will squeal as it’s going round because it’s a poor quality shell or it’s degraded inside, but really, it’s the obsolescence of the equipment that’s the issue. Getting hold of decent equipment will be the challenge, and once you’ve done that, it should be fairly straightforward.”
What’s the best way to go about capturing cassette audio for archiving purposes?
“It’s partly about the originating format, which is cassette, but partly also about the content, and your intended archival purposes for it. So, because it’s cassette, if it’s a cassette of music I would argue, if you have the resources, to do that at a high resolution, so, 24 bits, 96 kilohertz. For spoken word material, I would probably still wanna go for 24 bits for capture and maybe, you know, 48 kilohertz for example. So, 24 bit, 48 kilohertz. The advantage of going to 24 bits rather than 16 is that you have more headroom, more dynamic range, shall we say. So, there’s always a risk with a non-professionally made cassette which is that the output levels will be highly variable, so you may think you’ve got a very quiet cassette and you start off by recording at a high level but then something really loud occurs and you go right over and it distorts appallingly, so you need enough headroom for that. At 16 bits you don’t have as much headroom, you have to sort of push it as close as you can and take a risk. At 24 bits you can have it a lot lower. You’ve got 48 decibels more headroom at 24 bits than you do at 16. So if something is at minus 5 DB full scale at 16 bits that’s the same as minus 53 DB at 24 bits, so you’ve got masses of headroom. So what you could do, if you’re nervous about that but you still want to be efficient about the space is that you could digitize at 24 bits with loads of headroom, then adjust- normalize the recording for loudness up to minus 3 DB so make it a louder up to a certain level and then dither, convert it to 16 bits, so, 24 bits, as the number implies, is a third more data per sample and therefore per recording than 16 bits. So, if you wanted to dither it down to 16 bits you would ensure you’ve got the right head room and so on, but you would then have a smaller recording. The important caveat there is that you do document any processes you’ve carried out any subjective processing such as normalizing the recording, dithering, sample reconversion and so on, so do document that in your metadata. But that’s one way you can do it, or you might just want to keep it at 24 bits.”
Are there any “tricks” to cut corners?
“There’s no magic bullet, no, but cassettes are good because they’re relatively uniform, I mean, you know as it says on the shell, it’s either a C60 or a C90 or whatever so you know that if you have 4 C60s they’re gonna start and stop at the same time. We tend to do them in batches of 4, where we have 4 that have some degree of uniformity. So if those four cassettes, for example, were made by the same person, perhaps it’s a long oral history interview, we know that all the tapes in that interview, if they’re made at the same time, were made on the same machine with the same azimuth settings, roughly the same quirks that the interviewer had, you know, were they typically too loud or too quiet when they did it, did they typically use a (incomprehensible) cassette or not necessarily. Those kinds of things. So we know that the fewer decisions you have to make per cassette, the quicker, the simpler, and the better job you’ll make. So we do 4 cassettes at once going into an 8-track machine, a digital audio workstation to an 8-track analog to digital converter, and so that’s roughly how we work. It means that you’re not listening to the entire content all the way through but you can spot through them all and you can see the four way forms or the eight way forms for the eight channels appearing on the screen, so that will tell you if things are going too loud, or if there’s any kind of problems. If the cassette is muting, if there’s nothing on it, you can see that kind of thing and so on. So that’s a relatively straightforward way of doing it.”
Any final thoughts about cassettes as an audio format?
“Techie people get a bit wound up about formats but the reality is – it’s all about the content, isn’t it? The great thing about cassette is that it democratized sound recording at a certain time in the 70s and 80s. It means we have collections of stuff on cassette that, had cassette not existed, we never would’ve had. So, it’s a gateway into a material. As a format it’s not hi-fi but it doesn’t need to be; the collection (New York Radio Festival submissions) you’ve been talking about sounds really fascinating – but it’s spoken word so it’s not like orchestral music that is completely contingent on super high quality. It’s really about intelligibility, capturing the moment or retaining the capturing of the moment. It’s important that you create wave files, not mp3s because at some stage in the future, inevitably, they’ll need to be converted to another format. If you were to archive mp3s only, they would sound okay, but then you convert them to another format in the future and the chances are they will not sound okay, if you capture wave files now, when you convert them to something else, they will sound okay. That’s a great way of future proofing. It seems like a larger file and more expensive but actually it’s saving in the future because it means you do it once, only… forever.”
And finally, here are some links, recommended by Will, which provide useful advice for anyone wanting to archive audio;
Bradley, Kevin Ed. (2009), Guidelines on the Production and Preservation of Digital Audio Objects (IASA TC 04, 2nd Edition).
“A brief, relatively easy to read and essential. A comprehensive guide to Best Practise in audio archiving”
Wright, Richard (2012) Preserving Moving Pictures and Sound
“An excellent primer, and a succinct survey of the current landscape, covering everything from access technology to long term storage issues. A good place to start, though it does cover video and film too. With authoritative suggestions for further reading.”
Casey, Mike & Bruce Gordon (2008) Sound Directions: Best Practices for Audio Preservation
“Sound Directions is a descriptive, detailed account of two institutions testing and applying emerging standards (Including the IASA TC documents) on the ground.”
Schuller, Dietrich (2008) Audio & Video Carriers
“Provides a useful overview of technical and passive preservation issues, and a summary of an upcoming more comprehensive IASA publication.”
“The PrestoSpace project aimed “to provide technical solutions and integrated systems for a complete digital preservation of all kinds of audio-visual collections”.