Tubular Worlds

February, 1995
Paul White
Sound On Sound


Mike Oldfield's new album was inspired by the writings of Arthur C Clarke and features an interactive CD-ROM section. Paul White asked the composer about the artistry and technology behind the project.

When Mike Oldfield decides to create a 'concept' album, there are no half measures. His inspiration for Songs of Distant Earth came from the works of science fiction writer and visionary Arthur C Clarke, hence the album title-but not content with reading the books, Mike travelled to Sri Lanka to meet Clarke and to ask him, amongst other things, about his own musical preferences which, as it turned out extended from Borodin to Jean Michel Jarre.

"I decided to do something that didn't follow the book exactly, but rather refers to it and to other stories by Arthur C Clarke, and then I got the idea for the interactive part of the album. Since then I've quite got into the swing of interactive CD as a new art form. There are all sorts of things to explore in this interactive video-it' s not really a game, it's more of an adventure or an experience. The potential is enormous-for example, you can create a world of many dimensions where in one dimension, up is down and down is up, or gravity works in a different way-you can do anything your imagination can think of and that 3D computer software will allow you to generate."

The album was recorded by Mike in his private studio; it would be wrong to call it a home studio as it is probably one of the most sophisticated studios of its kind on the planet. Based around a Neve Capricorn digital mixing console and a Sony 48-track digital multitrack machine, the studio is almost entirely digital, with every piece of equipment sync'd to a central master clock. Tape is complemented by hard disk recording in the form of Logic Audio/ProTools, and even the external effects processors are fully digitally integrated into the system.

 

WORKING METHODS

Being familiar mainly with Oldfield's earlier works, such as 'Tubular Bells', 'Hergest Ridge' and 'Ommadawn', I notice that the new album seems to be broken down into more readily identifiable, song-sized sections rather than being one contiguous, developing piece, even though the individual pieces still flowed from one into another, with certain themes continuing to surface throughout.

All my work is made up from sections-it's just a limitation of the CD format that you have to put track numbers m somewhere, otherwise people complain that they have to play the album from the start when they might just want to hear a particular part. This album has been divided into 17 tracks, though it could have been any number, and it probably comprises 15 actual tracks or sections asfar as the working environment is concerned.

I see that you're using E-Magic's Logic Audio for sequenced parts [the program runs on a Macintosh Quadra 850 connected to a two-page colour monitor- Four sets of Digidesign's ProTools bardware give the system 16-track audio capability].

Logic is a great program and I ve used Logic Audio quite a lot on this project - It was particularly useful for synchronising bits of percussion - you can line up the beats graphically I've also used the audio facilities to loop certain musical parts, but this isn't done out of laziness; sometimes things just work better when they are looped. Looping is very much a modern thing, rather like punching a button on your video to see the same thing over and over again. I must admit that it's something I really like at the moment - though I can't say I won't go off it in the future.

One of your trademarks is your 'double speed guitar' which, in tbe days of analogue, was easily done by recording the guitar part with the tape playing at half speed- How do you manage this in an all-digital studio, where the multitrack presumably has a limited varispeed range?

I can either do it in the sampler, by sampling the guitar (and then replaying it one octave higher), or I can do it in ProTools. You can also dump backwards echo and reverb in ProTools by reversing the waveforrn, and then playing it back through the desk With an analogue machine, you get that reverse effect by turning the tape over and recording the effects onto a spare tracks (with the tape now playing backwards), but with a digital machine, you can't do that. There are always ways around things.

How do you record your guitar parts?

The guitar is DI'd through a GP8 guitar processor and goes via a couple of modules from the original Manor desk; the compressor module is used to really squeeze the Strat sound. You want the guitar to still sound like a guitar-but I might put some flanging underneath, as well as dense, long echo. I'll also use the Eventide Harmonizer on rnost of the sampled guitars. All the guitar parts are DI'd using a processor-miking up an amp is OK for some styles, but I've been using a processor for the last few years and find that I get more precision-more control. It's also more practical when the studio is being operated by one person.

One thing that our many guitar-playing readers will be interested in is your very distinctive fast vibrato style. How is that different from the techniques most people use?

I use a violinist's vibrato - well actually, a mixture of all kinds of vibrato. When most guitarists do vibrato, they'll pivot their hand around the thumb resting on the back of the neck, but I lift my whole hand up and down which means I can apply vibrato to a whole chord. I play with my fingers and I'll always stop notes that aren't being played with other fingers, which keeps the sound very clean. It' s something that' s developed from my folk days.

Folk musicians are famous for being fanatically purist. From the way that your music has developed, this obviously doesn't apply to you.

I'm not afraid of progress. This album is a stepping stone towards a completely interactive video album, which is where I believe the future lies. It's a whole new art form for creative people. But then I hope people don't neglect things like real guitar playing - I could have made this whole album on the computer, but then it wouldn't have had any soul or personality.

Did you make much use of othermusicians on the album?

There were a lot of different singers a boy soprano, two separate groups of Gregorian singers, and so on but just about the only other musician was a tabla player. Even then, I just took one small part and looped it. The world music vocal sounds came from samples; some were from Polynesia and Lapland, and some came from standard CD-ROMs. I trawl all the sample CDs I can find and I got a few samples from friends or from film soundtracks. I found some good things on the Zero G Datafile sample CD if you look through the CDs carefully enough, you can find little bits and tune them up or tune them down. These would then be further processed to make them fit the job in hand. It all sounds so boring when you describe it, but it was really quite exciting at the time!

How would you describe this album compared with your previous works?

It's actually turned out to be a very ambient album. I've tried listening to it loud through large monitors, but I don't really think it works like that. It's kind of made and mixed to he heard more in the background. The guitar sounds I've used aren't like performance guitar sounds, they're more like a force or a colour. They're not defined in space but spread out. Because I was trying to describe the indescribable, I had to do that in a very atmospheric kind of way. It was difficult - it would have been much easier to make a rock album or a folk album. I get about 10 new grey hairs every album.

Some of the tracks use very modern, (or should I say retro?) percussion sounds, which could make them popular in clubs. Have you considered the possibility of club versions or remixes?

Quite possibly. I've worked with - what do you call them - rhythmologists, and I was quite impressed by the way they took over the whole studio with samplers and little echo boxes. But the album already uses modem sounds - 10 years on from the first drum machines!

Where' s evolved a kind of science of rhythm over the past few years - you go to all that trouble and expend all that energy in search of the perfect drum loop that win make the whole world Dance - like a kind of magic pan flute. It was interesting to realise that there is a science of rhythm, and it's nothing to do with things being quantised, it's far mote complex. If anybody ever bothered to analyse the mathematical properties of the perfect drum loop they'd probably discover all kinds of interesting correlations.

 

TIME AND SPACE

On the subject of timing, do you use a lot of quantisation, or do you tend to use things as they were played?

If you're working on a drum loop, then there are certain attacks of notes that don't sound right unless they're absolutely spot on - flamming just doesn't work unless you're after a kind of tribal effect. But in between the downbeats, I'll try all kinds of syncopations or shuffles. If I'm playing a keyboard part, I'll more often than not play it straight to tape, and obviously a guitar part will be played to tape, though there may be a certain amount of digital editing afterwards.

Have you tried any of the spatial enhancement systems currently on the market to create more a 3D feel in the music, or do you rely on more conventional effects?

I don't think there are that many people who sit down exactly at the centre of a pair of loudspeakers when they listen to music, and the problem with the current 3D sound systems is that they don't work unless you're sitting in just the right place. Perhaps people used to do that once, but now I think they're less inclined to do so. I haven't been too impressed with what's available in terms of spatial enhancement, and there are samples on the market processed with things like Roland's RSS system if you want to use them.

My existing effects all have digital inputs and outputs because the studio is all-digital, other than a few front-end devices. One of the units is a TC M5000, which gives me reverb and pitch change, and then there's an Eventide 4000 Harmonizer which just sounds right, as well as being easy to use. Other than that, there are just one or two delays used for echo, flanging and that kind of thing and of course, a Lexicon reverb.

All the patching is done from within the Capricorn desk - if I want anything in a channel, I just dial it up on the monitor. It uses MADI (Multitrack Audio Digital Interface) communications, which makes things very fast. My Strat, which has single-coil pickups, is also very susceptible to picking up buzz from the computer monitors so, as well as keeping as far away from the monitors as I can, I also use the Capricorn's internal gates to clean up the sound.

So you don't miss any aspects of analogue?

No, although when we cut the album, I was a bit worried that because the signal path was all digital, the end result might sound a bit angular. We experimented by putting the whole thing onto tape using Dolby SR, and though it did something nice to the sound, the loss of clarity was also evident. It was nicer in a way, but at the same time it was disappointing. In the end, we went with the digital mix, which gives a very detailed sound absolutely everything was digital after the mic preamps and the guitar preamp.

 

THE CD-ROM

You obviously feel that the CD-ROM part of this project is particularly exciting. How does this all work?

This album is the first time that record buyers will be able to get a bit of interactive software on an audio CD. What we've done is used Quicktime II,which most people who have Macs will know about, and as far as I know, this will be the first commercial release using it. All the rendering is done on a Silicon Graphics machine - the whole thing is put together in wire-frame mode so that you can see how everything moves, and then you leave it to render, usually overnight. I love working in 3D, and the Silicon Graphics computer is probably the most interactive machine you can get.

I've just got hold of some new software which provides real-time parameters such as multiple virtual gravity sources, which enables you to make things move in very unusual and exciting ways. The particles being affected by these gravity sources can be rendered to give any 3D shape, and I could, for example, render 20 different versions of a visual sequence and provide 20 different versions of music to go with it.

At the moment, there are so many different formats of CD-ROM. How do you cope with this, or have you opted to support just one platform?

We re writing only for Mac at the moment as you say, there are too many different formats and the Mac is a very popular machine, especially in America. The CD incorporates some real-time video in with the graphics, and the uncompressed files occupied gigabytes, so it wasn' t clear how far these would compress. It was taking something like 17 hours to run the compression algorithm, and the final size ended up nearer to 70Mb, which fitted comfortably into the 150Mb of remaining disk space. Only a small part of this disk is interactive, but I'm working towards a fully interactive album for my next release. Interactive CD desperately needs to be standardised before it can really take off in a big way

Do you plan to get involved in more of this kind of work long-term?

Yes, very much so. I can do that from within my existing company, though I will need some help if I'm not to spread myself too thinly. There's been a team of three people working on this current project with me. The next project will be the completely interactive album, but at this moment, I can't predict when that will be completed.

Do you think this technology signals the end of music as a stand-alone art form?

I do. It's sad, maybe, but you can't stop progress. There will always be people who will just play, but it's a new world and the future belongs to multimedia. It may be a few years before we see really spectacular results, but things already exist which point the way things will go. Take the Back to the Future ride at Universal studios - we're talking about that kind of experience but with music.

I guess that the interactive thing will only really explode when we start to see TVs coming as standard with a CDI slot built in.

Even that will only be an intermediate stage; within five years we should have a 'pay to play' entertainment system where you can download whatever you need - music, films, interactive from a central media library over a telephone line or something similar. I don't know how the billing will work yet, but everything points to that being the system of the future.

 

THE TECHNICAL PERSPECTIVE: ENGINEER RICHARD BARRIE

Richard Barrie is a studio engineer in the mod traditional sense of the word; he is an anologue and digital electronics specialist whose role was to install the studio, and is now to keep it running smoothly. He doesn't, however, get involved in studio engineering - Mike handles all that himself. I was keen to find out more about the workings of an all-digital studio.

The studio is based around a Neve Copricorn desk and a Sony 3348 digital multitrack machine, and rather than master to DAT, we tend to master two tracks of the multitrack. The album doesn't get mixed in the usual sense because Michael will often mix as he's going along. Quite ofen, he'll even put tracks 47 and 48 (the stereo master tracks} and one other track into record! so that he's actuary paying his guitar into the stereo mix at the same time as tracking it and mixing. This means that the mix evolves as it's going along, then one day he decides the album is finished. As far as I can determine, most if not all, of the MIDI keyboard or sampled parts go down onto mubitrack, other than real off-the-wall stuff, and I've seen him pick up a guitar and play it straight into the mix without having the part recorded on a separate track. Similarly, he might want a finger cymbal or something at a specific point and it's easier for him just Play it manualy from a sampler while mixing that section. In those cases, the sound doesn't go down anywhere else other than into the mix. Micheal is very quick in the studio, and the thing that frustrates him most is being held up by something; he just wants to be able to pick up his guitar and play. He doesn't want to be sitting there for ten minutes while people are faffing around.

I would imagine that Mike is pretty well organised, because when you're working with tape, hard disk, samples, MIDI files and floppies, you really have to keep track of where everything is. It would be too easy to lose whole chunks of an album otherwise.

We have had the odd accident, and having just one digital multitrack is bad news in a fully digital studio. We bought the 48-track about five years ago, and within the first week I realized that we should have bought two 24-tracks instead, which would have allowed us to clone tapes.

So you subscribe to the view that digital information can only considered real when it exists in at least two different places?

At least two places. That's the whole thing about digital; you have to keep a very close eye on what's going on because you don't get any real warning of a problem, it just suddenly falls over. The error correction is so good that it disguises everything until it's too late. Because I'm watching error rates all the time, I notice the inconsistencies between tape batches or even within a batch. I've had so much bad tape over the bad years that I'll get a new batch and watch the first couple of tapes to make sure that the error rate is acceptably low, but even then, you might find another tape in the same batch where the oxide is falling off!

Any time we want us do a clone or an edit, we hove to hire in a second 48-track, and I'm often surprised by the condition of the heads, rollers and guides on the machines that do come in. I suspect that a lot of people don't watch what is going on in the error corection; if I see just one interpolate error on the 48-track, I'll clone the tape.

Everything in the studio is run from a master clock but how does your patching system work? I can see how the desk assigns its own internal dynamics processing, but what do you do about outboard equipment?

Michael's studio has really evolved from his previous studios and the system is very attuned to the way that he works. We don't need to keep re-patching effects, because I know that he's always going to want to have a delay line on Aux Send 2 or whatever. In other words, the stuff that's being used all the time is pretty much dedicated. Just to make the wiring easier, there is a certain amount of Patching and I have built little 1U custom interfacing boxes which provide breakout points or patch points. All the digital signals are linked using the AES/EBU interface - we don't use SPDIF at all.

Was the studio difficult to install?

Before this studio, we had a Harrison series 10 console interfaced with the Sony PCM 3348 in the anologue domain, so it was just a case of changing the desk and putting some new cabling in. I had to do it very quickly because Michael was working up to 2 o'clock one afternoon, and then we we put the Capricorn in that same night. After that I had just seven days to totally reorganise the studio, which meant planning ahead and making up little bits as I went along.

The studio runs from a master clock more for the benefit of video compatibility than for jitter. I could have taken the clock source from the Capricorn and then used that to distribute around the studio, but there is an EBU recommendation for the relationship between word clock and video sync. There are a lot of standards that haven't been standardised yet, and that can be potentially nightmarish. When I stripe tapes with timecode, I'll always be striping against the video time code so that when we do have to lock up against any video stuff, I know we've got the correct timecode.

Was the choice of equipment entirely down to Mike?

When Michael decided to go totaly digital, he decided to go for it in a big way. We trawled round all the digital processors with digital inputs and outputs and then, in effect, bought one of everything that mattered. Some of the choices were a little bit disappointing in that some of the new things are very difficult to program, and it would be so much nicer if you could use some sort of editor on your Macintosh. Instead, you end up poking about in these stupid menu-driven things.

A couple of years ago, I was taking minutes at an AES meeting when we were discussing a standard interface and control protocol for all professional equipment. That would be wonderful, because everything could then just be object mapped, and the quicker a standard is adopted the better.

Presumably going digital takes away all the headaches of ground loops and similar problems?

We never really had ground loop problems before, but some of the equipment isn't always as well designed as it should be with regard to grounds and screens. Even with AES/EBU and MADI, there are wiring conventions to follow, and even with MIDI, there are some units where the cable screen isn't lifted, and that can give rise to Problems. Ground loops are horrible things, and they don't just happen in audio - their effect can knock on all over the place. You just have to apply normal good housekeeping, though manufacturers don't make life easy for you. Even people like Neve, in their analogue interfaces, have decided that you might have some kind of Noddy wiring up your studio, so they've tried to be clever by fitting series resistors in the screen connections and all that kind of stuff, but it would be easier if they'd just present it as it's supposed to be and then document it in the manual.

I think that in digital audio, there should be a set of standard test signals, just as there are in video, so that you can easily predict a propagation delay. In a complex facility, if you're moving information around and you want to bring it back sample-accurate, it's a nightmare.

We were initially worried about the Propagation delays in the Capricorn desk, and even though you can take an advanced audio signal from the multitrack to compensate, if you're in advanced digital output mode, when you drop into record, the machine won't switch to auto input monitoring. I trust Michael's ears, and the delay doesn't worry him, but the problems we do get are much more subtle. A classic example is this: if, right at she last moment, he has a stereo mix on channels 47 and 48, and he wants to modify that slightly - Just dump a guitar in the middle of something that's already mixed - it means that those channels have to go back through the desk onto a couple of spare tape tracks, where they are delayed slightly due to the propagation delay. If this stereo recording is then transferred back to tracks 47 and 48 with a crossfade, you get a sort of flange across the crossfade, due to the small difference in timing. The delays in terms of working aren't a problem - the problems arise where you don't expect them.

Isn't it inevitable that the multitrack recording system must eventually be integrated with the desk, so that all timing delays are compensated for automatically without the user hoving to worry about them?

The full studio should be integrated. You shouldn't even know whut you're actually recording with - the studio should be intelligent. This studio is so dumb it's not true when you consider the technology. You should have something like a recognisable tape transport controlling an intelligent recording system that uses multitrack tape, RAM, hard disk, and whatever else is available. Say if you're doing a vocal overdub, you set up your monitor mix, rewind, find your cue, go into rehearse record, and then the first time you go, your muldtrack spins. It knows that you're doing an overdub, so it'll take your stereo monitor mix and dump it out to hard disk, or whatever, invisibly, so the first time through, the tape is really running, but after that, you're working into RAM or hard disk or whatever. As for as you're concerned, you're still using the seme tape controls, so you don't have to know where the data is going, and then when you've finished, the audio is automatically dumped back onto your multitrack tape.


Mike Oldfield Tubular.net
Mike Oldfield Tubular.net