Tag Archives: ballet music

Diamond fairy variation: new piano arrangement


My new version of the Diamond Fairy from Sleeping Beauty. Click to download free piano reduction.

I had to play this yesterday at a competition, and surprisingly, it’s the first time I’ve had to do it in public. It’s vile to play. Nowadays, if I’m faced with something like this, I go back to the orchestral score to see if there’s anyway I can make the job easier for myself, or better for the audience. Click here for my new version

Siloti’s pianistic homage versus a workable ballet reduction

The first thing I noticed about the difference between the orchestration and Siloti’s arrangement is that while Siloti’s hovers up the top end of the piano within the span of two hands, in the orchestra, those left hand Gs are in fact octaves, an octave lower: forte bassoons, arco bass and celloThe cost of his accurate representation of detail in the flutes and clarinets is the loss of the off-beat chords played by oboes, cor anglais and three, sometimes four, horns. 

Siloti's arrangement of the Diamond Fairy

Siloti’s arrangement of the Diamond Fairy from Sleeping Beauty (Act 3 No. 23, Var. 4)

Siloti’s transcription works both as a piano piece, and as a credit to what is most compositionally interesting about Tchaikovsky’s work here. But as the accompaniment to a variation, and for the dance accompanist, so help me God, it doesn’t work at all. You feel so utterly ungrounded, and so focused on the wrong things: to accompany a variation you first of all need a beat that is so strongly and safely grasped that if you need to change it, you can. Without it, it’s like trying to throw a pot with one hand; trying to steer your way out of a skid with only one hand on the wheel. 

The flutes and clarinet figure in the Diamond Fairy reduced to a manageable handful.

When I make arrangements like this, I do a constant accounting exercise: how much is lost if I take this out, how much gained? What’s the trade off between having a bass at the right pitch, and hearing the clarinet? I’m fairly convinced that you could get away with reducing it right down to the example on the left, and no-one would be any the wiser. Then it’s literally safe in your hands, rather than your hands being preoccupied with precarious detail, and you can use the other hand to play the bass at the right pitch, or give an impression of the horn chords; give it some weight, some “floor” in the music. 

Forget the clarinets: that’s a pretty thumping offbeat accompaniment in the oboes, cor anglais, bassoons and horns.

Less is more—except when it’s not

Considering how many times pianists around the world have to play the Tchaikovsky ballets in rehearsals and at vocational schools, it’s astonishing that we are still stuck with the first piano reductions, with all their inadequacies and problems and unsuitabilities. To my knowledge, my version of the Black Swan variation is the first publicly available reduction of one of the most famous solos in the repertoire. We all struggle along in our corners, doing our own ill-informed thing, assuming the score is right or the best possible, and only thinking about alternatives when problems occur.

Galina Bezuglaya, head of the Vaganova Academy music department is one of the few people to have committed anything to print about this   Amongst other things, she points out that it’s mainly other pianists rather than composers (or ballet accompanists) who make arrangements, which will bring a particular perspective to the reduction; Glazunov piano reductions are difficult because he tends think orchestrally, not pianistically (on the other hand, sometimes less is less: in the Raymonda Act 3 Hungarian coda, you really want to hear a good thumping bassline in the correct (low) octave); Tchaikovsky spent half a summer simplifying Taneev’s piano reduction of The Nutcracker, because—as he said in a letter to Ippolitov-Ivanov—”Taneev’s is so difficult that it’s impossible to play” [сделал облегченное полное переложение балета, ибо С. И. Танеев настоящее сделал до того трудно, что нельзя играть]. I’ve been typesetting a lot of Nutcracker recently for a job, and every time I go to put back in something that Tchaikovsky took out of Taneev’s arrangement, I end up taking it out again when I try it out on the piano.  Piotr Ilich knew what he was doing. 

Tchaikovsky and Franco-Italian hypermeter once again

On a different point, what continues to flummox me (which I can do nothing about) is trying to find the harmonic, melodic shape of the opening phrase. If you place the centre of it in the wrong place, you can wrong-foot yourself badly, and be tempted to miss out a beat. I am increasingly convinced that what’s happening here is a factor of Tchaikovsky’s tendency to write in what Rothstein calls Franco-Italian hypermeter . There is a very subtle interplay here of meter and grouping that will fall apart if you try to think only of a single metrical accent. There are (at least) two, and they are in counterpoint with each other (see also this post and the one’s branching out from it). I still haven’t worked out a fail-safe way to think of this phrase, I can only get through it safely by not thinking about it. All offers of advice gratefully received. 


If you’ve suffered at the hands of the Diamond Fairy variation before, I’d be interested to know what you think of my arrangement. I deliberately didn’t post this until I’d actually done it in performance. It seemed to work for me, the best proof being that I felt able to adjust the tempo from the corner of my eye, something that I’d not been able to do with Siloti’s. Don’t take the notes in the right hand too literally: anything that approximates the harmony will do. You can steal and copy some notes from the harmony in the left hand, leave things out. I have no idea what I really played in the heat of the moment. 







A year of playing cards #5: An operatic adage by Dvořák

"Song to the Moon" by Dvorak

Click to download

The curse of the operatic adage 

I think I only have about three of these in my repertoire, which is why it was high time I got another. The way that some ballet teachers mark adages, you’d think the world was just full of voluptuous music that went “and one and a two and a.” I guess my worst fear is when you’re thinking of what to play, you settle on something fairly plain that will work, and then the  teacher does that inclined head thing, gives you a knowing smile, and says “Something inspiring.”  You have to hope they don’t add “…for a change”. This is the stuff of nightmares, because it usually wipes out what you’d decided to play (which is another reason not to decide what to play until the last minute. You never know what tempo or adjective is going to hit you in the few nanoseconds before you play the first note of the introduction).

This aria from Rusalka is just about perfect. The tune really does go “one and a two and a” so there’ll be no fumbling about while the class finds the beat, and half way through, it goes all Maria Callas. I’m afraid I’ve had to do inexcusable metrical surgery on the first part, leaving out a whole 8 bar phrase in order to make it regular, but it’s hard to hear the joins unless you know the aria really well.

Playing tips

You have to have heard this before trying to translate it into piano music. The opening muted strings are hard to reproduce on a piano, and you have to do a lot of work to get the tune out on top, but If you’re lucky, you won’t have to fill it out with semiquavers, though that’s a possibility if you don’t have a very good piano or nice acoustics.

Watching this video is a rather fascinating lesson in how to play for adage well. Listen to the elastic, free, fluid vocal line in the “chorus” bit, and look how the harp accompanies it with almost metronomic rhythmic precision. It must be really precise, because in fact, the last semiquaver that you hear in the bar (part of a single group in my piano reduction) is not the harp (which is silent on the last semiquaver of the bar), but the last note of the pizzicato string figure (quaver, quaver, semiquaver semiquaver) that accompanies the harp.

Pianists tend to be “expressive” and pull the timing around in the bar, but for adage you need to choose your moments very carefully. To provide the right kind of support for a dancer who is doing the equivalent of the vocal line, you have to be as rhythmically solid as that harp and those strings, but at the same time hint at the elasticity of the vocal line. It’s something like the Chopinesque rubato where the accompaniment remains steady while the right hand floats free, but somehow conceptually different. Hard to put into words, but easy to see in this clip.

Metre issues

I’ve put this in “Spades” (Adage) because it’s quite definitely an Adage (see here for an explanation), but on the other hand, it’s about as truly triple metre as metre gets, which is common in some Czech music. Yet more proof that “three” is a big subject in music: so many ways to be triple.


Organizing music for ballet class: problems and solutions


Organizing music for ballet class: where to put stuff, and where to find it again?

I’ve had a blog post brewing for years  on the topic of organizing music for ballet class, but I’m glad to see that someone has saved me the job. Trevor Hewer’s The best way to organize music for free ballet class sums up the problems, but he’s also got a nice picture of a folder with coloured tabs. For the blog post I never wrote, I wanted to collect pictures of people’s folders and discuss their approach to categories and ordering. It’s a fascinating topic.

I agree with all Trevor’s points – memorization seems the best way to me, because sometimes things work for reasons that you can’t categorize. Conversely, once you start categorizing, you build boundaries around the objects you’ve categorised. As I’ve written about before, time signatures (for example) can blind you (deafen you) to the possibility of hearing something in 12/8 as a waltz, or a waltz as being effectively 12/8 or 4/4 with triplets. The other day I got myself out of two difficult corners when I suddenly realised I could use a tango instead of a slow march, and a tarantella instead of a quick one. The less you know in these circumstances, the better, because you don’t rule out things on the basis of their category.  I also found that as much as I like ForScore on my iPad, I find it very difficult and time consuming to catalogue my music on it – and it would take me forever to do it well; and as Trevor points out, “extreme ease of adding new music means less emotional connection with it” — quite. What I’ve found with my own “year of ballet playing cards” project is that the act of choosing, inputting, arranging and practising and blogging about each piece embeds it much better in my memory. You have to love your repertoire into practice.

None of this makes the problem of categorization go away, however. You still have to categorize your memory, or find ways to deliberately overrule it. Sometimes, when I’m stuck, I think of music titles beginning with a letter of the alphabet, or in a certain key; or I think of a word or word-type (songs about love, or food, or places). Sometimes, a particular teacher will get repertoire out of me that I’d forgotten about, and I couldn’t tell you why. Some miniscule difference in the way a teacher marks, or what you know about them, or the way they interact with you. More worryingly, some teachers seem to elicit the same narrow repertoire from me every time, and I have to find ways to avoid the rut.

The thing about memorization is that it potentially allows you random access, whereas a book requires you to think about order and category, titles, subtitles, genres, and so on. I had to make decisions about it several times in my previous job. One of the things I learned is that if anyone else is going to use your system, your plan has to align as best it can with the way that the user thinks about the topic, otherwise it will never get used, or will fail.

String and categories

It’s interesting to see how supermarkets deal with the same problem. For example, the other day I helped a woman in Wilkinson’s who couldn’t find a plain ball of string anywhere. She’d looked without success in “household” (where you’d think it might be), she’d settled for a ball of overpriced, fancy, “hobby” string in the toy section (for making necklaces and so on). I eventually found what she was looking for (plain string, 50p a ball) in the stationery area, on the packing and parcel materials shelves.  We had to try to understand the “mind” of the shop in order find it.

Categorizing music is no different:  what are the chances of someone else working out where you have put a piece of music? If someone else has devised the system, you have to be able to understand their mind and their culture in order to retrieve an object. Trevor makes a good point – you might as well put your repertoire in alphabetical order, since the thing you’re most likely to remember is the title.

Faceted navigation

But that only gives you one chance of finding it – if you don’t know the title, you’ll have to go through all your library piece by piece. The best chance you have is to assign multiple categories to each one, so that you and others  have several chances of finding what you want. On different days you might want the same piece for different reasons: because it sounds French; because it has a long anacrusis; because it’s in 3/4; because the teacher likes it; because you didn’t play it last week; because it’s in E flat and in a related key to what you’re playing right now; because you can adapt the rhythm and feel easily to the exercise (which is an affordance somewhere between you, your imagination, the music, what you’re playing it on, and the exercise – it’s not a property of a sound object). I have a piece in mind right now, the Petite Valse by Joe Heyne which has been all of those things to me.

I came across faceted navigation in Paul Lamere’s “Social Tagging and Music Information Retrieval” (article, free to download) when I was writing about similar issues in training ballet teachers to deal with music (given that the immediately relevant categories of music-as-heard and danced-to are not the same as the notational ones, how do you start talking about music, especially when the music can be categorised in various ways?).  Faceted navigation means being able to locate or select a piece of music by using any of a number of its attributes (“facets”) not all of which are “inherent’ in the sound at all, but cluster around the object through discourse, cultural conventions, use and personal experience.

I’m now writing about similar things in my thesis, but with the additional theme of boundary objects. Perhaps the strangest and most elusive thing about musical “objects” in the ballet world is in the context of use, a piece of music becomes “something I can do this to” where this is gestural, and does not become particularly musical (rhythmic, metric, dynamic, etc.) until the music is drawn into it. Try categorizing that.


Confessions of an anxious pianist #26: Same or different music on the other side?


Although the “anxious ballet pianist” series is officially over, I’m adding one more post now, because I realised today that after thirty years of playing for class, I still often ask myself the same question: shall I play the same or different music for the other side of an exercise?

Sitting on the fence about music on the other side

Same or different music for the other side: cat sitting on a fence

Sitting on the fence about the “same or different music on the other side?” issue.

What’s made me think about it is that I’ve just played for a teacher that I first worked with maybe 28 years ago, who made me bristle (back then, that is) by saying “Please play the same thing on the second side” after one of the first exercises at the barre. I bristled for a long time, because variety was my shtick, and was what I believed you were supposed to aim for in class: avoid boredom and sameness at all costs (see previous post on fear of repetition). I remember crying into my beer with another teacher, who cheered me up by saying “But it always feels different on the other side anyway — it’s not the same thing.

That was 28 years ago. A few months ago,  I played for that teacher again, and with the wisdom of experience, I remembered that he liked the same music on both sides, and so that’s what I did, without any bristling.  Experience had also taught me that he was a highly respected teacher with a securely individual approach and style, and that he had known exactly what he was doing when he asked for the same music on both sides. Looking at him and his class again, I realised that I had been lucky to have the correction, because it had given me something to think about for thirty years: only problems generate solutions.

Playing for him again more recently, I reminded myself not to alternate at the barre, but this introduces another anxiety: I know why I’m repeating the same music, but the class doesn’t. Do they care? Does it matter? Will they think I’m dull, or lazy? Part of me thinks that nobody probably gives a damn, they’ve got other things to worry about. And particularly in this case, the exercises are hard enough that the music needs to be there to help, not distract.

“Same or different music for the other side” is a constant dilemma (literally, a choice between two unpleasant alternatives). In another class recently, after I’d played the same music for three groups in adage in the centre, I decided maybe I could do better, so I changed the music. The teacher (one of the most experienced and musical I know) stopped me and said something like “You’ve lost them. Play what you played before, they can’t find what they need in the music.”

Now that’s an even more difficult dilemma: what I was playing wasn’t great, but it at least had the virtue of familiarity after a couple of groups. Possibly, what I was going to play would have been better had I played it the first time round, but now it was too late: better the devil you know. It’s the wise choice, but it runs counter to the pervasive idea that progress and change are unquestionably a Good Thing.

From both sides now

The trouble is that there is no right or wrong about this issue:  you just have to make a reasonable guess about what’s right in each situation, and risk getting it wrong. I probably got the idea that changing the music was a good thing because I learned my trade playing for syllabus classes where any diversion from the set music was a welcome relief. The teacher who said “It feels different on the other side” was right, and there are other occasions when changing the music has a positive effect. But there are other times when you have to let the music listen to the exercise, so to speak: when it’s new, difficult, or to achieve a very particular thing. As I’ve discovered, that might not only be with children: it can be at company class level as well, but you have to know when and where what is appropriate.

I got it wrong last week, I realised half way through pliés that the tiny rhythmic hint that the teacher had given in the marking was not just incidental or accidental, it was in fact exactly what she’d wanted. I changed the music to something more suitable halfway through the exercise, and she smiled and nodded at me.  I felt great for a moment, and then thought “Why didn’t I just do that the first time around?”  Was I clever to have sorted it mid-exercise, or stupid for not getting it right at the beginning? I don’t know.  That’s another anxiety to add to the list.

A year of ballet playing cards #50: A chameleon-like march by Granados (DJ)

Screen-grab of the Marcha Militar by Granados. Free piano music for ballet class

Click to download the score of this chameleon-like music

The march that isn’t a march: one of the perennial problems of music for ballet class

Another problem that I could have added to my “Confessions of an anxious ballet pianist” series is the search for something march-like  that can nonetheless immediately adapt to its surroundings without losing its identity:  a musical chameleon that can be staccato and legato, slow and  fast, up and  down on the beat, but not too much; loud and  soft, rhythmed and even; even but not mechanical, strict but not rigid. You need this for a certain kind of battement jeté exercise that tries to be all things to all women, and is neither fish nor fowl, musically. Enter the Military March by Granados (Marcha Militar).

Originally for piano duet, this little march is great for those occasions when you start playing and then realise, horrified, that you misread the exercise in the marking: it turns out to be slower/faster, louder/softer, more down on the beat, more up on the beat than you thought, and so on. With the Military March by Granados, you can pick various levels of the meter and emphasise them. There are different sections that vary from soft and fluid to sharp and detached, but within those sections, you can also alter your articulation and dynamics without causing any life-threatening injuries to the music.  I found it thanks to Susie Cooper, who recommended as something for a children’s piece in a school on a Facebook thread. I heard two bars, and fell in love with it. Thanks, Susie.

How fast is a march in music for ballet class?

The published score is marked allegretto: poco lento which would give this a warm, demure, leisurely, slightly pastoral feel: a parade in a country town after lunch, not the Red Army Choir or The Dambusters.  In fact, it’s more of a literary march than a military one, to borrow a concept from Raymond Monelle, who talks about the “cheval écrit” — the literary horse.  If there’s anyone marching here, it’s not an army, it’s  the pianists, dressed up in toy soldier uniforms. It was written in 1904, and  dedicated to King Alfonso XIII of Spain.  However indirectly, the favour was returned later: when Granados and his wife died in 1916 as a result of the torpedoing of the SS Sussex , King Alfonso set up a collection to raise money for the orphaned Granados children.

There’s a nice performance of it here, (as the original piano duet) and another one in the clip below, for brass band. Both are faster (at least to my mind) than necessary, and lose some of the potential for elegance and subtlety —but they demonstrate how it could be played fast, as well as slow.


The arrangement

In making the arrangement, I’ve tried to keep almost everything in, so you can see what the chord voicings should be, but it would be impossible to keep that up all the way through. I’ve shown an ossia at the beginning to give an idea for what it could be, when simplified.  I find myself that even when I know that an orchestra would double the bass at the octave, I’m nervous to actually do it unless I see it written down, so my principle in reductions is to put it all in and let the player decide.

The manuscript of the Military March by Granados

For some details about the composition/publication history (in Spanish) see this short article.


A year of ballet playing cards #35: A mazurka by Hubay


Click to download score

Slow, mazurka-like exercises from the corner for multiple pirouettes are a staple of all the company classes I play for, and if you haven’t got the right kind of music, it’s the longest 10 minutes of your life (see earlier posts on the “dreaded slow mazurka and “think mazurka, not waltz for pirouettes“). This has been a problem for me for 30 years, and it’s only recently that I’ve come close to solving it. Once you’ve experienced playing the right kind of music for this, you realise just how wrong a waltz is.  An eminent teacher recently said in company class that it wasn’t until he was 50 that he realised that the difference between a waltz and a balancé is that a waltz goes down-up-up, and a balancé goes down-up-down. When he said that, a light went on for me: I realised that this probably explains why waltzes tend to be wrong for an exercise with a balancé in it—the third beat of the bar will have the wrong gravitational feel (see my article on meter, ballet, and gravity if you haven’t already).

By who? By Hubay, that’s who

I first heard of Hubay when I was researching music for another project, and came across Hullàmzò Balaton, which was remarkable in that it contained one of my favourite bits of the Grand pas Hongrois in Act 3 of Raymonda (see earlier post), that I had always believed to be by Glazunov. I guessed from this that Hubay probably wrote some other good dance tunes. What I wanted most was something polka-mazurka-ish, but with oomph. Of all the “playing cards” I’ve created so far, the most useful one for me has been the polka mazurka by Verdi.

Mazurka or polka mazurka?

Hubay calls this a mazurka, but rhythmically it’s got that characteristic rumpty-tumpty-tumpty of a polka mazurka, yet has none of the tweeness. It’s the same rhythm as the middle section of the Coppélia mazurka, which is also useful (as long as you’re not playing for a company class, where you may get shot for playing it). Incidentally, the original of the Hubay is remarkably similar to this, with the change of rhythm prefaced by four bars of fifths on the violin, as here. It’s interesting to note, however, how subtly different they are below the surface: Delibes’ appears to be more markedly in 4-bar phrases compared to the 2-bar units of Hubay. But harmonically, Delibes’ change of chord on every bar makes it more markedly more truly triple meter than Hubay, who moves from G major only after the fourth bar: those two-bar units are beginning to look suspiciously like 6/8 in disguise. The longer you play for ballet, the more you appreciate how details like this can be a tipping point for choosing one piece rather than another for an exercise.


Extract from the mazurka in Coppélia


From Hubay’s mazurka – same key, same fifths, similar rhythmic patterns

Hubay’s mazurka works well for pirouettes if you play it slow and large. At a faster speed (the crotchet = 172 that I’ve marked) it also works for a certain kind of grand battement. Once you’ve played it a few times and the rhythmic patterns and conventions are in your fingers, you can use it as a basis for improvisation. Another convention that is good to bring in is the huge leaps across two octaves, which would be out of place in vocal music and counterintuitive when you’re thinking pianistically.

I’ve done a lot of messing around with this to get it into a format that will work for class. In the original—though I didn’t notice until long after I’d input it—there are several 12 bar phrases (or rather an 8-bar antecedent followed by a 4-bar consequent), and 8 bar interludes. Better to work on the assumption that there will be 32 counts per dancer, and then you don’t get left hanging mid-phrase.  However, the original is lovely to listen to, so here it is without the straightening out and the cuts:

Because it’s a concert piece for violin, there isn’t a recording of this that gives a sense of what it could be like when it’s butched up on the piano for a ballet class, so I’ve quickly recorded a rough version to give an idea of what I think it can do. It could go slower than this, and there’s plenty of room for rubato and pauses and stretches to allow for multiple pirouettes and other contingencies. Forgive my mistakes, but it’s better than nothing.

PS: There’s a small octave mistake that I’ll correct when I have the will to live — it’s in the repeat of the G minor section near the end on page 3. The D-Eflat-D motifs should be up the octave, as they are the first time around on page one.

See also: 

A year of ballet playing cards #53 (Black Joker): Reel from La Sylphide

Image of piano score of The reel from La Sylphide, Bournonville's ballet with music by Løvenskiold

Click to download the score

I promised that I’d be posting a couple of “jokers” in the pack of ballet playing cards, and here’s one of them: the Reel from La Sylphide.  It’s a joker, because you should only play this one with care: I’d advise against playing it for class anywhere where La Sylphide has recently been in the company repertoire, but it’s such great music for little jumps, and such fun, that it almost always raises a smile, or gets people trying to remember the steps while they wait for the next group.  A crotchet/quarter-note beat of 108 is only a guide, but it’s not worth playing if it’s slower than that, frankly – the whole point is that it’s fast, furious, and jolly, as a reel should be. To talk shop for a moment, it’s also great for little jumps because it has a proper four-on-the-floor feel, rather than falling into a see-saw pattern of oom-pahs. Also rather interesting how often it has a strong beat at the end of the bar. Explain that one to your music teacher.

It’s handy because it goes on for pages and pages (I’ve included everything, but the most useful part stops at the end of page 3 – after that, go carefully, although it’s in 8 bar phrases almost right up to the end. There’s something really exhilarating about playing for an exercise that goes on for a long time and not running out of material – because people keep think you’re going to, and then you don’t. I’ve done this joker as a favour to myself, so I can remind myself of how it all goes. I’m also rather fond of a post I wrote about it when I had to play it after slicing the end of the fourth finger of my RH with a food processor blade, and discovered interesting new fingering methods.

A reel life experience

This piece brings back so many memories too – one of my earliest jobs was sight-reading it from a terrible hand written score for Peter Schaufuss and Festival Ballet (ENB) at a stage rehearsal back in the 80s. It was then that I realised how much more fun Bournonville ballets were than the turgid Tchaikovsky ones.  Probably the two best ballet nights out I’ve ever had were watching Matthew Bourne’s reworking of La Sylphide (Highland Fling!) in London when it first came out, and then in 2013 when it was restaged for Scottish Ballet.