Harold Budd is more than just a leading light in contemporary American music. His work has a natural beauty which seduces the listener with tranquil progressions, delicate warmth, and skilful arrangements that use instruments as diverse as piano, marimba, and Synclavier II.

I first heard Harold Budd's music when a classical guitarist friend lent me the American's (then) only recorded work, The Pavilion of Dreams. 'It's pretty classical, you won't like it', said the friend sceptically, but he was wrong. If I had to pick one disc to take on a desert island, The Pavilion of Dreams would stand a good chance of being chosen.

At the time (eight years ago, at a guess), the record's gentle harp, piano and marimba patterns provided an almost perfect backdrop for the breathy alto saxophone and sustained vocal harmonies that took the place of conventional 'solo' instruments. It was a fascinating combination, made more intriguing by the list of playing credits on the sleeve — Budd himself, Brian Eno, Michael Nyman, Gavin Bryars, and several of the other luminaries associated with the EG/Obscure series on which Pavilion appeared.

The album ended up being something of a cult hit, and continues to sell in reasonable quantities a full decade after its initial release. But Harold Budd has never been too interested in the commerce of modern music, and resisted the temptation to make Pavilion the launch-pad for a 'recording career' in the conventional sense.

Even now, records seem to appear with Budd's name on them almost by accident. Two albums with Brian Eno (The Plateaux of Mirror in 1980 and The Pearl in '84, both on EG) and two solo (The Serpent in Quicksilver in 1981 and Abandoned Cities three years later, neither of which have been properly released in the UK) do not a steady output make.

Live performances by Harold Budd are equally rare, and press interviews have been very few and far between.

Yet despite (because of?) his low profile, Budd has continued to write compelling instrumental music, switching comfortably from treated acoustic piano to modern keyboard instruments of every level of sophistication — a Casio CT202, an Oberheim Matrix 12, and a Synclavier, to name but a few.

And although his compositions display obvious recurring themes — long, sustained chords; delicate reverb treatments; bright, percussive timbres — Budd is just as at home penning a two-minute cameo ('The Chill Air' from Plateaux) as he is improvising his way through an extended, side-long epic such as the title-track from Abandoned Cities.

Getting hold of an interview with Harold Budd turned out to be surprisingly easy. The LA-based composer visited London for a month earlier this year to embark on his latest recorded project, an album with the Cocteau Twins — Robin Guthrie, Liz Fraser and Simon Raymonde.

A day before he was due to return to the States, Budd sat in a penthouse apartment above the Townhouse Studios in West London, and began talking. His manner was as friendly as any musician I've ever interviewed, and the overwhelming impression gained after a meeting with him is that, more than most of his contemporaries, Harold Budd has a musical mind that's open to almost any idea, anytime, anywhere.

The Pavilion of Dreams was the first record you made. Before that, what was your background?

Well, I began composing for The Pavilion of Dreams in 1972. Prior to that time I'd been teaching Music Composition at California Institute of the Arts near LA, and before I actually became a college teacher - which I didn't until 1976 - I was an avant-garde composer of the sort of music that probably wouldn't be associated with me these days. I was very much influenced by John Cage and Morton Feldman, for example, and I wrote music of a very indeterminate nature. I wrote very quiet, spacy theatre pieces, some of the music was mere verbal directions about an activity which might take place.

That was rather my role as an artist up until 1970 or so, when I really minimalised myself out of a career. There was nothing more to do and I became disenchanted with the conceit of avant-garde music.

For two years, I knew what I didn't want to do as an artist, but I didn't know what I wanted to do directionally. I slipped back into discovering something that no-one else was doing, or was likely to do in the very near future. I divorced myself from modern music in a sense, and began to develop a language which I thought was honest to God me, and totally outside of competition with my fellow composers. It was freedom from all that.

I must admit that, looking back. I'm very fond of those pieces. I really think of myself as a composer starting from about 1972. Prior to that, I was the Harold Budd that I don't recognise today.

Were you looking for an outlet to release your music at that time?

I wasn't looking for an outlet because I was teaching in college, and there was always an outlet there. There were ensembles, and I tended to write music for the ensembles that were available, because I wanted to hear what it sounded like. And my music did get played a lot in modern music venues across the country, so I didn't have any problems in getting performances.

But in terms of a record, it never occurred to me. It didn't seem that it was an alternative that had the remotest possibility of happening.

I think secretly I rather had a hunch that it was just a matter of time, and that something would occur. But I didn't work to a plan and I didn't work toward a recording career, shall we say.

How did you eventually get to The Pavilion of Dreams, then?

As I recall things, I had quit my teaching position at college, with absolutely no idea in the world what I was going to do, and brimming with confidence that something was going to happen.

I got a call from Brian Eno in London. He said he'd heard 'Madrigals of the Rose Angel' (one of the works that made up The Pavilion of Dreams) and asked me if that was the kind of music I wrote generally. I said it was, and he asked me if I'd be interested in coming to London and recording these pieces. So that's how that started.

"My work with Eno has no specific mood - that's what makes it interesting... when you can pinpoint a happy piece and then a sad piece, things get too simple."

How do you feel about your two subsequent collaborations with Eno?

The two albums with Eno are very similar. The language, the timbres and sounds are very much the same. It's curious about The Plateaux of Mirror. It came so quickly and so easily that it was kind of a phantom, you thought you could do 30 more of those with the same joie de vivre, but it's not that way.

There was an idea that The Pearl would be similar even from the outset. The surprising thing to me is that it turned out as well as it did, because I'm extremely fond of the album. In many respects I like it more than I do The Plateaux of Mirror, even though that came so easily and so gracefully and without trial.

The Pearl is more of a unity than Plateaux. All the pieces seem to belong together more. Eno's material tends to be melancholic, whereas yours sounds a bit more optimistic. The mix of the two was quite peculiar...

Yeah. I think that's one of the points that makes it interesting because it's not a specific mood. There's one there but it's very complicated. Sometimes it's a fusion of conflicting waves of emotion. I like that kind of artful confusion. I think pieces turn out not so interesting when they are so emotionally specific, when you can pinpoint a happy piece, then a sad piece. That's too simple.

That can happen very easily when people put lyrics on things.

Yeah, as soon as you have lyrics you're talking about something specific - unless it's an ironic sort of thing, surrealistic. The work that Eno and I have done together is, in a sense, not far removed from classical surrealism. It has an ambiguity of meaning which I think is an important aesthetic consideration for works that can't be placed inside of a fad or time.

How did your newest collaboration — with the Cocteau Twins — come about?

Well, I have to admit that I didn't know of the Cocteau Twins until one of the band members, Simon Raymonde, was using a piece of mine and Eno's as a cover on an album. The piece was 'Not Yet Remembered', and it was going to be the Twins themselves who were going to do it, or that's the story I heard. The publisher called me up and informed me about this, and said 'they're really a great group and I think you'd like hearing them', so I called up a record distributor friend of mine in Los Angeles and asked if he carried the Twins.

He sent me a compilation cassette, so I don't know what the album itself was, or the names of the pieces. But in any case, I was really taken with it.

Then in November they came to Los Angeles. We met and got along famously for a very brief time, and we started swapping ideas on a collaboration of some kind. They asked if I'd be free to do it, and I said 'yeah, absolutely, anytime, just give me a shout and I'll be there'. And here I am in London!

Have you spent most of your time recording?

Yes, basically. Robin has a studio in North Acton. It's a 16-track studio and we do the basic tracking there daily, and then he'll take it to a 24-track studio for transferring and mixing.

The music has been very improvisational. It's working the way most people work anyway when left to their own devices. In this instance, everyone takes the blame and the credit equally, so it's kind of composition by committee. It's pretty apparent early on in a piece that something is happening and is ringing our bell. It's a constant creative process in the studio, because nothing is written out. Very, very little is planned beforehand: it's just a matter of laying some tracks down and seeing what works.

Some very strange pieces have come up, I can tell you. That's really the thing about collaboration that fascinates me so much - the notion that you come up with music that neither one of the collaborators would have dreamed up if they were left alone. It's something that's absolutely unique to the collaboration. Consequently it's not 'pure' in the sense that it's not Cocteau Twins, and it's not me: it's an odd combination peculiar to the mixture.

Do you prefer that style of working to composing on your own?

It really depends upon the nature of the music. There is some music where I absolutely must make all the decisions myself, because I have a sense about its structure, where it's going, and what I actually want it to sound like. I'm the only one who can make those basic decisions, and an input from a second party is kind of counter-productive.

But when you know it's going to be a collaboration from square one, you go in with a totally different attitude. And of course there are two ingredients. Number one is the people with whom you're collaborating; you must like them personally because it's terribly aggravating work. Number two is that you have to like what each other person's art is, with respect to their previous work, so that you just have a hunch that it's going to work out. That's the kind of trust that's impossible until you actually put it on the line.

At the moment I take it that the music is purely instrumental. Will it have vocals on it?

"The studio I work at in Los Angeles has both a Fairlight CMI and a Synclavier II - they're a million times more than I need, though I know they do a million things."

Yes it will. I think Elizabeth Fraser's put vocals on one song so far. Apparently, her method of working - which I totally agree to - is to take the instrumental tracks home and then compose the words and melodies. She works up something that's comfortable to her and seems to work, and that's her way of working.

You're perfectly willing to accept that she will add another dimension to things that you may not agree with...

Yeah, sure. Because in a sense she's rather forced to accept what I'm doing as well. It all goes back to that trust I was talking about.

Who seems to be coming up with the initial melodies?

We have eight or nine songs - well, pieces - tracked. The interesting thing is that the ones which I come up with tend to be pieces, whereas the ones that Simon comes up with on the piano are definitely songs - they're structured that way. Mine frequently don't even change key once they begin, or even change chord for that matter (laughs). Robin's pieces tend to be somewhere in between the two, so I can't say, really.

For the first three or four days I laid down a whole bunch of stuff we had to work on, and then I took a break and Simon began some, then Robin began some and now we're back to the other way again. But it's just going to be raw tracking - I will leave an awful lot to them. We have to see how it works out. It's album length now, and it'll appear as an album.

What instruments have you been using to make these strange noises?

Well, it's somewhat restricted. I have a philosophy that one is obliged to use what's there. You don't need an awful lot of stuff, but you use fully everything that's there. What is there now, insofar as my own input is concerned, is a Yamaha electric grand piano and a Mirage sampling keyboard. Robin of course is the guitarist, and Simon the bass player, although he does occasionally play the piano, rather well too. And there's an awful lot of outboard equipment there which works very well. It's what gives them the Cocteau Twins sound - it's the stuff they use. Primarily it's the Yamaha SPX90, which is absolutely incredible... an amazing instrument.

Do you find, though, that machines like that sometimes offer you too many options?

That's a very very good point. The answer generally is yes. In this particular instance, since the style that's been developed really only uses a small fragment of the potential of the thing, it's perfectly clear that some sounds work and that all the others don't.

But yes, I feel that if you can do anything, then you're liable to end up with everything... or indeed nothing. It's kind of self-cancelling.

But where we've been recording, there is a limited amount of equipment and an even more limited amount of time. There really isn't room for very much finesse. You have to have a pretty good hunch that you're correct very early on, or else you have to abandon the idea and get on with one that's going to use the time more efficiently.

The Mirage and piano seem an odd combination: from something which is very close to a grand piano where you just lift up the lid and play, to something which you have to spend hours fiddling with to get a sound. How have you found working with something like that?

Well, I've worked with something similar before, or at least I've touched on it before. I'd say that 90% of the sounds are crap, and once you find that out, there are certain disks that you avoid, really. With the piano, I tend to work hearing a treatment with it anyway, so I improvise pretty roughly around what the sound is going to be on the tape.

So your composing is very influenced by the way things sound?

Yes. In many respects, the sort of treatment you hear on the piano influences exactly the note-to-note process: the length of time between musical gestures, and the kind of taking advantage of the ringing timbres which I'm very fond of.

Have you done any sampling of your own?

No, I haven't, partly because I don't own a sampler myself. The studio that I work at in Los Angeles - Meta Music - has both a Fairlight CMI and Synclavier II. That's a million times more than I need, though I know it does a million things... it would go out to lunch if you knew how to program it, I suppose. But really, its capabilities are rather boring to me. What I do like are certain sounds which work for certain pieces. They're rather easy to come by, and the fact that they're on these mega machines is really totally irrelevant.

You don't find it rewarding to explore a machine to find out what it's capable of?

To some extent I do, yes, but not to the extent that the engineer would have to know it, because he would have to know what its capabilities were for situations that are totally unpredictable, depending on the client, for example. For my own purposes, I think it's useful to know as much about it as you can, but I don't see the point in knowing any machine inside-out at the moment.

"We've used the Mirage for strings, organ and chorus... things that sound nice when they are sustained for a while, as a wash in the background because the guitar does all the glitter stuff better."

In what areas has the Mirage been useful?

We've used strings, organ, and chorus sounds... generally things that sound nice when they are sustained for a while, using the instrument as a wash in the background. They don't form the primary sound source, because the guitar, being heavily reverbed, adds a great deal more to it. The guitar does all the glitter stuff much better than the Mirage.

What do you have lined up when you return to the States? Will there be another solo album?

Yes. It'll be a solo work on EG. Half of it is already done, and the other half I have to do this summer. I have a lot of work to do on it because I haven't got the foggiest idea what I'm going to do.

The first side is a full 20-minute piece. It was done in LA, using the Fairlight and Synclavier, and a general array of Oberheim keyboards including the Matrix 12. I use the Matrix in a very unsophisticated fashion, which is to say that although it's multitrack, everything is a live performance. I'm not even using the Synclavier to make the notes proper: it's me, what you hear is the way I'm hitting it.

The piece's working title is 'Gypsy Violin', because that's the name of one of the factory presets I used on the Synclavier. It's a gypsy violin sound and it's excellent.

I originally did the piece for an art gallery installation in Los Angeles last November. It was called 'Blue Room with Flowers and Gong' - which describes exactly what it is. It's an environment through which you walk, or just take a quick look and walk away and think 'this guy's crazy'. It was originally the length of one side of a C60 cassette, but that version went on a bit - the sustained notes and so on...

The sustaining note and long, sustaining chord are very much trademarks of yours...

That's true. I really like to find as much life as possible in the smallest amount of material. A very simple scale, a relationship of note against note, especially a sustained note; I milk everything for all it's worth.

For example, I first found this latest piece when I was messing around, and I had it kind of scrolled away in the back of my mind. The first time I sat down and played it, I had a Matrix right above me and I was sitting at the keyboard. I used a bit of reverb - not too sophisticated - and I sat and played for 90 minutes. I was enthralled with the piece and listened to everything, so I didn't have to do any multitracking to find out if it was really going to work or not; it just happened.

I'm generally confident that I can pull this stuff off at the first sitting and get the best version. Invariably I hear two or three versions, but the first is always honest and innocent with all the mistakes, it's more human somehow.

I take it your confidence in improvisation comes from a reasonable training and a reasonable musical ear...

Well, it's nice of you to say that, but in fact I have utterly no training as a performer. I got through college without knowing how to play the piano at all. I was a traditional note composer, and I couldn't play anything on a piano.

Music theory was my strength. It was a different world. I didn't start learning to play the piano until I had to out of self-defence. For the keyboard part of 'Madrigals of the Rose Angel', for instance, I would take out a note card and write the chord down without rhythm for the ensemble, and say 'well it's just got to fit in with the harp somehow', and it was always terrible. It dawned on me that I should be responsible for showing the ensemble how the chord sounded and how it fitted in.

I can now play what I do rather well, but I can play nothing else at all. If you're writing a string quartet, why should you have to play it on the piano? One of my composition teachers demanded real proficiency on the keyboard and a good ear, which I don't have at all; I don't have perfect pitch, or even good relative pitch, for that matter.

Have you done any film soundtracks or theatre pieces? It seems to be a popular American pastime...

Yeah, but I haven't really done it. The only thing remotely like that that I've done is something that just occurred prior to my coming to London. I did two 30-second ads for an ad agency who were working for Korean Airlines. Apparently, someone had said they could do with something like Harold Budd's music, and someone else had said 'Well, why not get Harold Budd?' They said I could do anything as long as it was 30 seconds long. The only limitation I had was that the music had to sound like me!

I saw the visuals, but I didn't have to work to a SMPTE track. I did anyway just to make sure that it would time out correctly, but I didn't have to. The commercial was very open anyway. It was fairly apt for my music.

Last winter I met the head of music production at Universal Studios. I didn't solicit the appointment, he called up and invited me over. He said he'd dearly love to use an original score of mine for a movie, and I said I'd dearly love to do it. But he said that, on the other hand, I must realise that they made 10 feature films a year, none of which are appropriate for the type of music I do, but if something came up would I do it, and would I do it in Europe, in London? I said yes to both questions.

What music do you listen to besides your own?

All kinds of things. I remember that in my high school days I was a great fan of bebop. If it didn't have Charlie Parker on it, I didn't want to know. What a snob I was!

"I got through the college without knowing how to play the piano at all. I was a traditional note composer... If you're writing a string quartet, why should you have to play it on the piano?"

Now I don't listen to all that much. I really rely on friends to turn me on to something that's interesting, or something I haven't heard for a long time. I remember that a student of mine came up to me at one time with the Monteverdi 1610 Vespers and said 'You absolutely must hear this' and it was just devastating. I still think it is, but I couldn't sit and listen to the whole thing.

So I generally play music off records when someone is over who I know is genuinely going to like it, and won't insist on everyone being quiet while it's being played.

You don't seem to play live too often. What's your view of live performance of the sort of music you compose?

Well, I'm not the world's greatest performer, and to make my music sound like I really think it should needs a lot of requirements - things you'd ordinarily find in a professional recording studio. I find soundchecks cumbersome and boring, and since I really don't have much of a clue about what I'd end up playing anyway, what if I have it all set up wrong for a particular piece? I do perform occasionally and I do enjoy it, but I don't see myself in that role.

I like the democracy of recorded music: once it's out there it's not yours any longer, people have the option of bouncing the needle, going here or there, and listening to as much or as little as they please. That's why I don't own a CD, incidentally, because once it's on you're a prisoner to it. I can pick up a favourite record and flip it over and say 'oh yeah, this is the side I like without knowing the name of the tune.

You feel vinyl is still the most immediate means of communication?

I think it's ingrained in our generation. It's going to be difficult to get rid of it. Technology does an awful lot of things, but what it doesn't do is really influence our behaviour - I think we influence the technology.

I know a lot of people insist on CDs only, and everyone in America insists I'm wrong, but I think they'll take off for a short time and then be replaced by something easier to handle and less expensive.

Given that your musical ideas are very much based around studio work, would it be fair to say that you compose most of your music in the studio?

Yes, I use manuscript paper the way a lot of people use a cassette. I jot ideas down, and I use the manuscript paper as a kind of memory trigger. I often take reams of paper into the studio but I don't read off it - I just use it to jog my memory into an idea.

The engineer at Meta Music is Michael Hoenig, who used to be in Tangerine Dream. He's a good friend of mine so money really isn't a problem - he just tells me to come over when he's free.

I hate being in a studio with a strange engineer who spends most of the time trying to convince you that you're crazy, and doesn't understand what you're doing anyway. I've worked with engineers who want to put the drum track down first, you know, and I can't work that way. I like working with someone sympathetic who knows what the scene is, a good friend. I rely on someone like that a great deal.

Do you have a recording facility of your own at home?

No. Nothing at all. Just a small Baldwin spinet piano gathering dust, and a little Casio 202 which I plug into my son's guitar chorus and piddle around with. I can get as much as I need down from that.

What's the first item on the agenda when you get back to the States?

Curiously enough, a small college in the Midwest commissioned me last year for a piece that had to be 8-10 minutes long, and had to be for any combination of musicians that they have in the ensemble there. They sent a huge list of what they had, and said they'd pay me a certain amount of money and that I could do anything I wanted, but that they had the right to the first performance.

It turned out that I wrote a piece 22 minutes long and it's for a huge choir of 20 people, with percussion and keyboards, and it's the first traditional note piece that I've composed in eight years. It was a lot of fun doing it. So I have to go to the Midwest for three of four days, and be at the first performance.

Looking further into the future, do you have any unfulfilled aspirations, in any field of music, that you'd like to pursue?

No. Maybe it's a personality defect, but I generally discover things as they come along. I don't have any long-range goals about things as a basic personal philosophy, except for the obvious mundane ones. In terms of 'is there a piece out there that I've always wanted to do?', the answer is no, I don't really operate along those lines.

I'm inspired often by a situation which arises, when there's a certain task in front of me, and if I agree to do it, it's generally because I'm going to like what's happening, without having an idea what the end result is going to be. No career decisions are made.

With that sort of open attitude, you could go in any direction...

Conceivably that's true. I'm the first person to admit that I'm willing to change my mind at the drop of a hat. But I think I would be a very poor commercial composer, for instance. I'll never do anything that I don't absolutely want to do. There may be some commercial success resulting from this latest collaboration. But I don't think I would do anything else if the focus was simply on the amount of units sold.