Striking the wrong note

“I’m only human.” It’s what we say when we make a mistake. Without mistakes, I’d be out of a job. When asked to explain what I do, I tend to describe classical music editing as “joining up the good bits and taking out the wrong notes.” This is, however, at best disingenuous and at worst a lie. Whilst it is a fair description of why the profession exists, it is not, as it turns out, actually what I do. So, what does the job really entail?

When you think about it, to say I’m only human is a rather odd expression. We seem to be implying that we’d rather be a computer, or at least that we aspire to the technical perfection a computer can achieve. We want to play all the right notes, in the right order, at the right time and tempo, and with perfect intonation and ensemble. We desire to be as true as possible to the score. This is the Holy Grail. But are we really on the right path? What is this truth we seek, and does it really reside in the notes themselves or does it hide behind them, between them and, most importantly, in the fragile nature of the humanity we bring to them in a performance?

In his speech on receiving the first Aspen Award, Benjamin Britten described the magic of music as being “[that] which is not in the score.” To find out what that magic is, perhaps we first need to ask the question: what is in the score? One way of finding out is to ask a computer to play it for you. With software like “Sibelius” or “Finale” this is easy, and we soon find out what Britten meant. Input the dots of a Chopin Ballade, for example, and out of the sausage-machine comes a lifeless deluge of sound that has little more to do with Chopin than Morse Code. The computer can give us rhythm, certainly, but only the kind of meaningless ensemble of a marching army, the soulless accuracy of a Swiss clock mechanism. The pulse, such as it is, comes not from the heart but from a life-support machine. There is none of the subconscious rubato and variation in colour and agogics that defines the musical outpouring of a human being. After all, the only thing consistent about a human performance is its inconsistency. (It is true that the software mentioned above can use algorithms capable of introducing a simple form of “espressivo”, but it is fake, unconvincing and, worst of all, utterly consistent.)

This reminds us that the score is not the music itself but merely a set of instructions – and a very vague set at that – formulated in order to make the music happen. Surely then, when we say we wish to be true to the score, it is not the computerised mirror-imagery described above that we aspire to, but something else; something that, as Britten points out, isn’t actually there! (Again, I would urge the reader to go to his Aspen Award speech for a much more brilliant, concise and eloquent dissertation on the subject than this one ever could be.)

So what does all this have to do with editing? To answer that, we first need to examine the structure and procedure of a recording session. For the sake of brevity and simplicity we will look at an orchestral one. Once a first “take” is recorded, the conductor and producer will listen back to it and pinpoint the things that need attending to – specific aspects like intonation and ensemble, as well as those more ephemeral elements like energy and passion. Sometimes the whole piece or movement is played again, sometimes not, but either way we soon cross the Rubicon and enter that strange, bizarre world known in the business as “patching”. Technically, a “patch” is anything less than the whole piece, or movement. It might be as much as three-quarters of it, or sometimes as little as eight bars. I have occasionally been asked to work with patches of only one or two bars in length.

Let us pause for a moment and ask ourselves what is happening here. To take a theatrical analogy, imagine recording the sound of a live performance of Hamlet. How happy would our actor be to come back the next day for repairs to deliver his soliloquy on its own, out of context? From where would he get his energy and passion? Where does the speech come from, and where does it go? And if he trips up on a word, would we really expect him to just re-record the half-sentence from which it came, or even just the word itself, in order to “patch” it? Even though, in the real world, this kind of thing does happen, we don’t need to be an actor to know instinctively that it leaves a lot to be desired. So why do we allow ourselves to do the equivalent with music? The answer is a simple one: because we can. (But see “Time is Money” for a more complicated answer.)

In a sense, we have been editing since the dawn of recording. Even with wax cylinders we could give ourselves the luxury of two or three “goes” and then choose the best one – which is a simple, but nevertheless very powerful, form of editing. However, the kind of cutting and splicing that we are now familiar with wasn’t really possible until the advent of magnetic tape. With razor blade in hand, a whole new vista of possibilities opened up before us: we could choose from different performances of the same piece and select and join together not only different movements but also different sections, passages, bars and even individual notes!

Now, if you can edit in short sections, then why not play in short sections? It was the natural next step to take. But the idea is an intellectual one, not a musical one. It makes the absurd assumption that each passage is self-contained, each bar unaffected by its neighbours, and each note an island, entire of itself. With this triumph of logic over instinct, Pandora’s box was open and the recording world would never be the same again.

In fact it is astonishing that patches work at all, on any level. And here we get to the rub: a patch is more likely to work if the playing is predictable. If the temperature is low, the tempo identical and the colour uniform, then we can successfully join one musical half-sentence or syllable to another without too much trouble. The fewer long takes we record, and the more short patches, the more likely the system is to work. All we need is to be homogenous, constant, and dependable; in other words, utterly consistent. Because of the comfort and reliability of this modus operandi, the recording world has evolved to depend on it more and more, sometimes using very few, if any, of the complete or long takes and instead to construct literally a “patchwork” of small, separately recorded sections. It is a highly efficient way of re-creating the score, but almost totally ineffective at revealing the music within it.

Advances in technology haven’t helped. With computerised digital editing almost anything technically (though not musically) is now possible. At least with a razor blade there was a sense that we were all potential Sweeney Todds and that without great care before every editing decision we might slip up and cut our own throats. Today, because all edits can be made and unmade in any order, and at great speed, there is the danger of being so cavalier with our virtual razor blade that we end up building ourselves a Frankenstein’s Monster. I will try, later, to explain how this technical freedom can be turned to our musical advantage.

For now, we must return to what I like to call the tyranny of the printed score. For it is the score, and the score alone, that is the lawmaker, the arbiter of truth, truth carved in stone, the judge and jury over all that is right and wrong. You were born a C sharp, and a C sharp you shall be forever more. Amen. In saying this, am I suggesting we should disrespect the composer’s wishes? Absolutely not. It goes without saying that good intonation and ensemble, not to mention the right notes, are fundamental building blocks of music-making.

Without them it is anarchy. But they are a means to an end, not an end in themselves, and the score is merely a map that helps us to explore, and to discover the buried treasure that we so eagerly seek: music. Not all of the composer’s wishes are contained within the printed score and the most important and meaningful of them simply cannot be, precisely because it is not possible to write them down. Even whether to play a C sharp or not is sometimes the least important thing. Perhaps composers should quote the film director James Cameron at the beginning of all their scores: “Don’t give me what I ask for, give me what I want.” That is why orchestras have conductors.

A score is two things: the tools of communication, and the message itself. The first is visual, and the second is aural. If we ignore the latter it enables us to play music in short sections; or, in other words, to make music with our eyes. It is the printed score, and our visual dependency on it, that gives us permission to record a patch of one bar.

Let us remind ourselves why we “patch” at all. As previously alluded to, the recording method has, over the last century, distilled out three main reasons for making an edit: wrong notes, suspect intonation and poor ensemble. But this is to put the cart before the horse. Correct spelling and pedantic punctuation does not a poem make. As an editor, I have been incredibly lucky to find myself in the privileged position of being able to compare two or more versions of a performance by some of the greatest conductors and performers of our generation. What is it that makes take five a little more exciting than take seven? Why does the climax of take eleven make the hairs on the back of my neck stand up in a way that take twelve doesn’t quite seem to? After fifteen years of making these comparisons, and in discussion with colleagues to whom I am indebted, a few theories and observations rise to the surface. The most simple and unsurprising one is that short takes have less musical validity than long or complete ones. (Like all rules, this one has its exception: the wise and experienced record producer can give the session itself a natural drive and architecture that short patches can pick up on and absorb, giving them some compatibility with long or complete takes; but this is an unusual technique, which, in fact, validates the idea that long takes are the most musically effective. It is also difficult to achieve.) Less obviously, a passage is often exciting despite some questionable ensemble and intonation. So, can we dare to ask the question: is it exciting because of those “questionable” qualities? (I suppose I am using the word “exciting” in the all-encompassing scientific sense of the word – in the way that temperature is transmitted through an object by one atom exciting the next – and not just to describe the obviously “big” moments in music; in other words I use it in the sense of having the power to change, move, and communicate with the listener on all levels.)

Is it those running semi-quavers in the woodwind getting slightly ahead of the pulsing cellos and basses, throwing off the shackles of simple ensemble, that actually move us? Is it the wild freedom of the melody in the first violins that raises the heartbeat and alters our metabolism, bypassing the conscious mind to touch the real human being inside us? Yes, in my opinion it is, though we need to be very careful here. This kind of argument is somewhat circular and we might short-circuit ourselves into believing that dodgy ensemble and poor intonation are inherently exciting characteristics. This would be silly and just as damaging as editing out those apparent faults from an exciting passage. Both of those decisions would be erroneous because they are led by the head and by the eye. It is much more useful to bear in mind that in the natural world there is no such thing as a straight line: good ensemble is a deeper and more complex phenomenon than simply playing vertically together. In the end, if a take is exciting, it is exciting; sometimes the playing will be precisely together, sometimes not. I’m just saying that, if it isn’t, we mustn’t cut it out as a knee-jerk reaction.

We must also not forget that when Pandora opened her box for the second time, she discovered that the Gods had concealed within the evil spirits one kindly creature: Hope.
The real power of editing lies not in the fact that we can cut between takes that are the same, but in the fact that we can cut between takes that are different. In doing so, something new is made, something that was never played at the recording session. It is an act of creation. This is the consequence of just one join. The fact that a CD may have as many as five hundred joins, each one having a consequence on the other that describes a matrix of cause and effect of almost infinite complexity (including “live” recordings; for a more detailed description, see “Live Editing”), is a measure of how awesomely powerful editing can be. But, to quote another voice from the film industry, with great power comes great responsibility. What are the things within the grasp of the classical music editor? With just one join in exactly the right place we can give ourselves a little more crescendo here, a little more rallentando there… so we’d better make sure we respond in the right way in the Da Capo in order to relate to the decisions we made in the exposition… and if we use some of take eleven followed by some of take twelve, do we create something even closer to the composer’s or performer’s vision than either take heard in isolation? And, most strangely of all, even though there are fistfuls of wrong notes, false starts and swear words in take ninety-three because the pianist was tired and fed-up and just wanted to go home, there is, in context, a wonderfully beautiful and calming effect brought about by using just the last few bars of it, creating a unique moment of sublime rightness that never actually happened in the studio.

These examples represent just the tip of the iceberg. To paraphrase that well-known adage from Chaos Theory, one edit can be the flap of a butterfly wing that causes a hurricane on the other side of the world. Many recording musicians will be running for the hills when they hear this. How much, if any, of his or her vision ends up on the finished CD? Indeed, what right does the producer or editor have in stamping his or her own personality on it? It’s a very good question. I could write another three thousands words trying to answer it, but the short explanation is that if we edit truthfully, honestly and with integrity, then the artist’s vision will not only survive but be magnified. The editor doesn’t join different takes together to manufacture a crescendo or rallentando just because he can – in other words because it’s clever – but because he has to, because the music demands it; it is where the performance takes him or her, as sure as night follows day. If we only edit for the right notes and with no other criteria, there is a serious danger that those tape-splices will inadvertently create, say, an inappropriate crescendo or a meaningless rallentando that bears no relation to the composer’s or performer’s intentions. And, discouraged by all of its wrong notes and the uncomfortable feeling we were left with at the time of recording, we wouldn’t even consider investigating take ninety-three. That is when we should run for the hills.

But we need all the right notes, don’t we? If we do, it is a psychotic dependency that invalidates most recordings made before the advent of digital editing (c.1979) and every live concert ever given. For example, Alfred Cortot’s recordings are somewhat notorious for their technical inaccuracies (wrong notes), yet he is considered by many to be the greatest interpreter ever on record of Chopin and Schumann. His white-hot interpretations are just as compelling now as they were then. It is tempting to make excuses and say that Cortot didn’t have the editing facilities and techniques available to us today (which is in any case only partially true) but to say so is not only arrogant and patronising by suggesting that Cortot, a man who, despite being obsessed with technical accuracy, didn’t know exactly what he was doing in making the choices he made, it would also be to entirely miss the point. The fact is, we do listen, and we are moved.

If Cortot were alive and working today, then it is likely that he would have enthusiastically embraced the modern editing paradigm. Could this have been done without losing the power and integrity of the recordings he actually made in the first half of the 20th century? Well, I believe it could, but I doubt that it would. It seems to me that increasingly we put in the right notes without asking ourselves what we are taking out. We have gained power at the expense of courage.

Today, we live in a world obsessed with cleanliness and perfection, suffering from allergies because we are not exposed in our formative years to their causes. This is why we have acquired an allergy to wrong notes. But right notes in themselves say nothing: they don’t inherently contain passion, energy, and truth (just ask a computer), whereas music-making that brings passion, energy, and truth will usually, if not quite by definition, have the right notes anyway. At least let us not confuse cause with effect. If we seek the former, the latter will usually follow.

Time is Money.

Actually, there is another, more pragmatic reason why we record in short sections, at least in the case of orchestral sessions: because it saves time and money. Orchestral recording sessions are astronomically expensive and they rarely pay for themselves in CD sales. We simply don’t have the luxury of recording a symphony over and over again until we get it right. Make no mistake, without the option of sometimes recording and editing orchestral music in smallish sections, we wouldn’t be able to record it at all. As for Opera, it is impossible to gather together all the necessary voices and forces to enable the work to be played from beginning to end, or even to record scenes in the right order. But the argument becomes a matter of degree: an eight-bar patch is still an eight-bar patch and we should not be distracted from considering the musical consequences of working in such an eccentric and unnatural way. It is also important to reiterate that not all artists and producers take the same approach. There are notable exceptions who are uncomfortable recording in short sections (but, with a shrug of the shoulders, mumble: “… that’s just how recordings are…”), or who only use patches effectively as rehearsals and top them off with another complete take or two, and fewer still who refuse to do it at all. Sviatoslav Richter is probably the most notable microphone-allergy sufferer and was particularly scornful of the short take – though, despite this, or, I’d like to think, because if it, he stills leaves a marvellous recording legacy behind him. There are one or two artists who, having experienced what for them was a troubling and unsatisfactory approach to recording, have come to the conclusion that it is best not to record anything at all, but in walking away from one deficient working method they leave behind all other possibilities, which is a tragedy.

Live Editing

Some artists are troubled by the moral issues raised by editing so they turn instead to “live” recordings in the belief that they represent a true and honest account of a real performance. Disenamoured by the recording process, they gravitate towards recording live concerts in a kind of back-to-basics approach, in the hope of retaining the integrity and honesty of their performance and also regaining some control over the end result. Record companies are increasingly interested in live recordings too, because they are, on the face of it, cheaper to produce (though that does not always turn out to be the case).

Some live recordings are, indeed, genuinely “live”: one performance, one recording, and no edits. But these are very much in the minority. Most live recordings are edited, and the more high-profile the project is, the more heavily edited it is likely to be. How is this possible? Well, if the concert is repeated we have two, and sometimes even three, recorded versions to work with. The rehearsal is often recorded too. In theory, this means that you can edit back and forth between each version as much as you like – every bar or note if necessary – and some live recordings are, indeed, very heavily edited. Furthermore, there will often be a “patch” session, recorded after, or between (or even before!) the live concerts. Sometimes patch sessions are so extensive that on the finished CD the entire movement of a symphony (say) may not actually be live at all. This is the safety net deemed necessary for such an expensive project. The record company is usually quite open about this: you should be able to find somewhere in the sleeve notes the dates of the two or three concerts and the names of the producer and editor; but it is, arguably, morally dubious to still call the end product “live”.

In fact we walk through a moral minefield worthy of a dissertation all of its own when we consider “live” recordings. But let’s not forget what the attraction of a live concert is: whilst we don’t want the trapeze artist to fall, it is the extra frisson of the fact that he might that is part of the excitement, and we feel a little cheated if he uses a safety net. If musicians don’t have a safety net, they play differently. Surely it is this difference we want to hear in “live” recordings. Personally, I’m always on the look-out for it in studio recordings as well – something that gives us the shuddering intensity of one of those life-changing concerts, but even better.


Benjamin Britten: “On Receiving the First Aspen Award” Faber and Faber 1964
Alfred Cortot: “Rational Principles of Pianoforte Technique” 1919


Stephen Frost is a composer. His Oboe Concerto was recorded by Chandos in 1999 and has been broadcast regularly on BBC Radio 3. Stephen has also worked in the classical recording industry for 15 years, editing for most of the major record labels including EMI, Hyperion, Warner Classics and Chandos. He was Music Producer and Editor for the BAFTA award-winning documentary “Leaving Home”, with Sir Simon Rattle and the CBSO.