Slow practice- what is it good for and what is it not so good for?
If we were to take the most common piece of advice regarding piano playing, it would almost certainly relate to the value of slow practice. If you observe questions on forums, virtually any request for help (no matter how specific) will likely be met with at least one generic recommendation to simply practise slowly and be patient- as if that is supposed to amount to some unbelievable revelation. Now, let's be clear, there are plenty of situations in which thoughtful slow work may be what is missing. Those who don't regularly use it are not doing themselves any favours. However, there are also plenty of common problems that are actively caused by excessive amounts of generic slow practice! The main body of this post illustrates the basic premise that defines truly productive practice. However, given how devoted some are to a mantra that slower is always better, it's also necessary to start with appreciation of the severe problems that can also ensue.
The myth is that when starting with repeated slow work, you are using the best available approach with which to build foundations for eventually playing at full speed. If someone tries to go faster and makes a mistake, they may wrongly believe that this means that they weren't ready for any faster practice at all, and that they must clearly stick solely with slow work for longer. This might seem like an excellent display of "discipline" and a sensible way to ensure that a foundation is built. In truth, such situations might more accurately be viewed as an unhelpful retreat from necessary steps, back into a comfort zone in which there is little meaningful progress left to make. I will show you why truly efficient practice should always include some elements of faster playing, from the very first session of learning a piece. For absolute beginners, it makes sense to allow a little more leeway (so they can get used to the basic feelings of playing the instrument and listening through sounds), but that still doesn't call for exclusive slow work. I'll give examples of how a more balanced approach to practice can be applied both to very basic examples and to advanced study. To start with narrow slow work risks setting up far greater frustrations further down the line. This will surprise many, but I'll explain why the time for the most extensive use of slow practice is in the intermediate and later stages of learning a piece- and not at the very beginning!
Before going to that, let's just acknowledge another common trend amongst amateur pianists. I'm talking of the kind of player who essentially "practises" almost solely in the spirit of performance style sightreading, rather than ever slow things down. They basically just approximate a performance rendition as best as they can. This involves playing through a whole piece (or substantial section) at a tempo in which there is insufficient time to pay attention to finer detail. While they may have a sense of the main features, issues will inevitably be glossed over- including notes, fingerings and control over sound etc. The results of this could be anywhere from a reasonable enough impression of the work, to the kind of playing that would see most listeners running to buy a pair of noise cancelling headphones. Either way, the results won't improve notably further by repeating in a full performance spirit, and neither is this what I'll be recommending as the alternative to a "slow is best" approach. Although these are two opposite poles of imbalanced practise, they actually share very similar underlying problems. Anyway, if you browse around Youtube, you'll particularly come across enthusiastic amateurs (as well as some supposed "prodigies") trying to force their way through the Flight of the Bumblebee, without having been through the groundwork that would prepare for a deeper sense of success. I don't wish to make an example of anyone, so here is a deliberately ropey simulation of the kind of playing I'm speaking of:
I show first a "proper" version (although, this too contains a tiny degree of imprecision- relative to the highest standard). Then there are couple of thoroughly "faked" versions. The second is the kind of thing that might be enough to fool a few listeners, but it is clearly far too approximate to be of any real value. The third is a little more subtle. On an off day, it certainly wouldn't be the worst thing that could happen in an actual performance. Nevertheless, this is really a plan B for surviving worst case scenarios. "Practising" mainly this way does not build towards reliability, but merely guarantees the need to fake at least that much in an actual performance. I was very much this way, before I realised quite how much easier I could make things. Such a learner is exactly the type who most needs to involve good quality slow practice. So why is that only one aspect of building strong foundations, and how can excess slow work cause problems?
Where slow practice equals bad practice
it's much less obvious when it comes to the problems associated with too much slow work. In order to understand this, let's start from a slightly silly exaggeration, before looking at what happens in real life. Supposing that we take a single finger and practise the Bumblebee, using only that one finger. Anyone who is capable of sufficient patience and mental focus should be able to get the part I showed with accuracy, if they slow things down enough. Even if you couldn't read the music, you could program a keyboard to light up the keys, and then peck at them like this chicken.
If you were to offer a substantial cash prize for full accuracy at any old low speed, there would be every reason to think that concentration could lead to a positive result after a single sitting. However, what would happen if you did this many times over and then tried to speed things up to a reasonable performance speed? Well, it obviously would never work. Even though every detail will have been understood mentally, you'd have to go back to square one and start again, with movements that would make it possible to flow through those same notes.
Imagine if you'd already spent a couple of weeks practising before you first tried it quicker! In this particular case, I doubt whether anyone would have wasted so much time without becoming suspicious. However, the foundation that is produced by normal slow practise can be as problematic. Arguably, the dangers are only greater in real life situations, because the flaws that you would waste time on would not yet be easy to spot. If you only go slow, you just won't find out what is wrong. It might not involve only a single finger, but it can still be based on fingerings or movements which could be unusable at even a slightly moderate performance speed. For instance, at a slow tempo you can afford to bob the whole arm on each note, without necessarily having any trouble. At a fast speed, the same style of movement will simply become frantic and lead to a sense of locking up. There are far more styles of movement that are good enough for slow speeds, than those which are also suitable to play faster.
You might notice that my camera work starts to resemble something out of a Jason Bourne film, by the end. All those jolts of the arm send corresponding jolts back through the whole body, via the force of the reactions that come back from the piano. Not only are such movements slower but they are vastly more physically stressful. All that hard work only serves to produce a low speed (scarcely higher than the maximum for one finger) and also a lumpy sounding performance. When a technique is based on individual prods, it's actually possible to get by up to an intermediate stage of learning. An equally common fault for slow practisers is to lock the arm in a static position- causing severe tension. You can get away with either fault for a time. However, you will eventually start repertoire for which your entire technique is unsuitable- both in terms of playing the notes and giving musical expression. This is why I try never to let even beginners get stuck in a "slow practice only" mindset to a new piece. Advanced technique doesn't come from using your arm to peck at notes like a chicken. Although doing so can have occasional value as one type of practice exercise, the primary role for the arm is to produce smooth horizontal movement that guides between the fingers as they play. Far from being the easiest place to learn this, slow practice is initially the hardest place of all to learn this technique- especially if you are going slow because you don't yet know the notes well enough to go any faster.
Closely related here is also the concept of mental grouping. When we do the disciplined slow practise, it might confirm a mental grasp of details, but it doesn't mean that we are organising them into patterns. Although some will instinctively start to make larger groups, there's no guarantee. Obsessive slow practisers have a tendency to fall into the trap of simply repeating a large number of isolated thoughts, on a note by note basis. Although slow work offers plenty of time to get ready for each note, it may not be used wisely. Think of the guy at university who is the last to arrive at 11am lectures, because he lives a couple of minutes away and thus sets his alarm for 10.55. When you press a key, do you yet have any more idea about what comes next than the chicken did? Often the mind only wakes up at the last possible moment, before an urgent thrust at the next key. When the tempo is slow, you could still survive this with both full accuracy and rhythm. Raise the tempo, however, and you now have a mental and physical frenzy that will not hold together under the stress of speed- thus triggering a desire to retreat straight back to slower work. Far from creating the best foundation, retreating to a long period of slow practise may produce scarcely more learning than will be achieved by the chicken, when it pecks at individual lights. It's specifically when we step out of the comfort zone and try going faster, that we gain the necessary feedback to start improving on our thinking- especially if that causes a concealed problem to come to the surface. This is where we finally get presented with a need to plan patterns. Which raises the question:
Why waste time slowly repeating hidden flaws that you are not even aware of, for an extended period? Why not find out about possible problems as early as possible, in order to solve them before they ever turn into difficult habits?
Targeted exploration of musical units- the important precursor to repetitive slow practice
Yes, I'm afraid that the only time you can guarantee that slow practice is likely to be useful is after you could also play a particular passage faster. This is what proves that you have the sense of larger thoughts and smoother motions, to start making the slow practise meaningful. It now becomes a chance to explore details within their required context, rather than only as isolated events. If this sounds back to front (like a Catch 22 that means only the super-talented might have access to this approach) I can assure you that it isn't. When we start by targeting moderate sized musical units (and save slow work on longer sections for the second phase of our practice), we can meaningfully compare fast and slow playing side by side, from the start. This has nothing in common with the player who fakes through longer sections. The units will always be small enough to both plan precisely and then easily monitor for quality. When we make mistakes, we will revert back to a slower version for immediate corrections, before varying speeds as required. The key is to start with units that you can reliably hold in your short term memory, as a single mental image. I'll come back to how to build the Bumblebee in the final section, but let's start with a relatively easy example of this targeted style of practice.
We'll look at the basic groups for the right hand, in bars 1-4. I have drawn square brackets at the top, which show our physical groups ie groups of notes that lie under the fingers all at once. Incidentally, I generally mark the majority of my fingerings via this method- so as to force thinking in groups rather than individual fingers. In this case, each group spans five notes between the lowest and highest pitches. Assign five fingers to five notes and any detail of fingering comes automatically, based on context in a bigger thought. In the film I first play two five-note clusters to reflect this- as the most global perspective possible for those four bars. Now, it wouldn't necessarily be "wrong" to start by breaking the music into two bar groups, based on these positions, and practise them separately from each other. However, it would leave out an important rhythmic join. In this case, our primary groups also require bridging across the two physical groups. Good early work will find the important musical groups and assemble them in the very first session. The primary musical joins in the score are shown with phrase marks. There are two particular principles here that are more important than the physical positions:
1. Musical units rarely finish either at the ends of beats or bars. The natural conclusion of a musical unit is almost always on the start of a beat, and most typically the first beat of a bar. Typical groups would be a bar of music plus a note, or a beat worth of semiquavers (16ths) plus a note etc. If unsure, it's usually better to add one extra note, just in case, rather than risk cutting off an important link. By aiming towards the main rhythmic features, even work on small fragments will contribute positively to your general sense of pulse.
2. A run of shorter notes can almost never be regarded as completed until they arrive directly into the destination of longer note. When short notes continue non-stop for a significant time, it is best to introduce temporary stopping points after (*never before*) arriving at strong rhythmic moments. This divides up the work into smaller journeys towards the main rhythmic divisions, with the lowest possible risk of causing habits of hesitation.
These probably won't make sense as abstract ideas yet. However, it should clarify the premise when I show how our practice groups come directly from these. Firstly, the most common early mistake is to play the four descending semiquavers and then freeze at the bar line. I show this first in the film, before connecting them to their destination. The next note is harder to find, of course. Also, the nature of beaming in music is problematic. In theory work it is convenient to distinguish beats and bars from neighbouring notes, with clear visual separations. In practical performance, linking across beats and bars needs to be the standard way of thinking, otherwise the pulse will break down. The notes which look most separated out (due to beaming and bar lines), often require the tightest rhythmic connections of all. Seeing as the the look of printed music is something that actively hinders our instincts, we have to be constantly mindful of the need to attend to these joins.
To play the four beamed semiquavers and then pause will defy both of the above principles for building coherent musical units. The last semiquaver does not establish itself as being a semiquaver unless the next note arrives in correct time. We might have played all four notes, but the last has no rhythmic meaning yet. It won't feel remotely complete, even as a smaller idea, because the notes just vanish into nothing when a strong beat would be due. The physical movement is equally incomplete, with no sense of being aimed at any major destination. If we either go slow or allow a separation there, we may have little sense of how to produce the needed motion. The link therefore needs to be built both into our earliest mental conception of the music and into the movement that we apply. In the video I show it first with a gap. It's okay if you do that once, but only if you immediately work on connecting up the whole unit (of four short notes to a long note) before going onward to the next unit. Generally, the destination of a musical group should be the first to plan, as the primary point of focus. If you understand it well, you'll be able to try it both slow and relatively fast. Compare both until it feels familiar and easy. "Fast" is a relative concept and need not immediately be extreme. However, you have to at least go fast enough to be sure that you have got it down to a sense of one fully premeditated action. If you are figuring things out halfway through playing, you haven't yet organised your thoughts. The arm should lead downwards and then back up to the high note, in one sweep. If you feel separate thoughts or movements (or if you make a mistake) always slow the playing down to take back control. Just don't go on to the next unit until you can get both get a very directed fast version and an equally directed slow version. Neither has value unless referenced against the other. The arm needs to guide us clearly through both versions- not by pressing down on individual fingers but by drifting steadily towards them. This drift of the arm is what makes the whole unit feel like a single process, rather than a series of conscious decisions.
The second time we get the run of descending notes, the destination is much harder to get to. The trick here is not to fixate on that note alone, but on the whole shape we are moving to. I don't start by pressuring myself to play the note itself, but simply make room to align to the overall hand shape. Again, pauses are okay before a trickier note, as long as we build in a true rhythmic join during the same session. The notes have to be able to lead to the F sharp as a rhythmic destination, otherwise our units will be full of holes. Again, going slow is okay for working out/correction, but you should check the same unit fast, to see if you have truly understood it yet. It's the slightly quicker executions that tell us where we stand. This piece is certainly not so hard that to just slowly muddle through couldn't eventually work. It's just a lot quicker to learn when we work to an active plan for covering the necessary difficulties.
Notice that the semiquavers are still going as we proceed into bar 3. The strictest application of our second principle would suggest we shouldn't stop until landing into the longer quaver A. However, we don't need to be so strict here. This is a great example of how we can use a strong rhythmic moment as a point of overlap, in order to keep our initial units small and manageable. Our previous group ended in the F sharp on a clear beat. Our next group will begin here. Because it is included in both, no connection has been left out! Just a single note of overlap is enough to ensure that our rhythmic connections are being set, even in this slightly fragmented early work. Because the main thoughts are directed towards dealing with all of the trickier connections, we will achieve much more than by just going slow or by hoping to do everything all at once.
The last thing I show is how the final two bars break down. Again, we should be acutely aware of the risk of stopping before the bar line and thus robbing ourselves of the most important link. A key difference this time, is that we should not even begin with a pause before the bar line! I show first on the film with a break between bars, but as an illustration of what not to do! Previously, we had particular difficulties that would make it initially forgivable. First there was a large interval and then there was a whole change of hand position, as we passed the bar lines. Taking the pressure off these makes sense at first, as long as we attend to the missing link. This time, notice that the note across the bar line is nothing more than a continued descent through adjacent notes. If we should stop in the middle of such a simple pattern, it is not caused by any difficulty. It can only be seen as a sign of operating much like the chicken ie in isolated events, without a sense of the pattern. We have to resist any tendency to allow our units to be dictated by bar lines, unless we are happy to progress slowly and with more difficulty!
Notice that if we do divide into bars, the last bar looks more complex. Nothing could make the second bar seem much more difficult, than freezing before playing the E. When we treat the strong first beat as being the natural conclusion of the previous idea, the notes we are left with form a very neat descent through a D major chord. I have marked these with a smaller bracket, inside the larger one. We aren't changing the basic alignment of the hand from before, so the original bracket stands. However, it's a good time to calibrate the hand into the more specific shape of those particular three notes. That's why I initially play it as a single chord, on the film, before doing the literal version. If the D major shape is already felt, the arm need merely guide us down through the three fingers of the existing shape. Although it would probably have been manageable to do the whole two bars in one go, breaking them into these two memorable concepts actually makes it even easier to assemble the two bar idea.
Following these preparations, it should now be easy to assemble the whole four bars in a slow tempo Note that the bars do work rather nicely here as a self-contained idea. However, if you should worry about connecting to bar 5, you could join the final D major unit to the first note of that bar- to fully comply with our two principles from before. Anyway, when we reassemble the music, this is where the traditional style of slow practise can now be done in a truly meaningful fashion- ie to explore the details (both musical and physical) of a fully mapped out journey. High quality slow practise is like deliberately slowing down a familiar walk to make a point of looking closely at the scenery. If we go slow because we frequently have no idea as to where we are really going, this doesn't even truly count as practising yet- let alone high quality practice. Many would fall into the trap of doing a slow trudge through a whole piece or section. Doing this many times over wouldn't guarantee a sense of how to truly join even the semiquavers to the downbeat of bar 2, let alone the other important links. It is vastly more productive to take smaller units and target them for early mastery, than to simply read through slowly. The sole way to be sure that your slow practise will count for something, is if you have already been through some of those focused bursts of quicker playing.
Extended slow practise should never be treated as a display of true discipline, before this varied work has been undertaken. If you can't yet do anything quicker, I'm afraid that slow-only practise should best be regarded as "chickening" out. I hope you'll forgive the awful pun, but you may indeed be operating more like that chicken than a good practiser, if you are avoiding difficulties out of fear. You're not likely to actually avoid the deepest and most meaningful mistakes. All you'll avoid is knowing about them. This is where slow practisers and the sloppier performance style practisers are opposite poles of the same problem. Both amount to repeating much the same broadly familiar experience over and over, generally in rather long sections and with minimal active intervention to fix things up. The only difference is that the performance style practisers are constantly exposing all their faults (before doing little about them) whereas the slow practisers are constantly keeping most of them concealed (before doing equally little about them). Yes, it's a simplistically harsh categorisation and analysis, but if you want to improve your progress then it's essential to be realistic about the likely results of any narrow practice style. What good practisers do is to take units and sculpt them into a desired form.
A way to summarise good practice would be to say that it is based on constant alternation between greater risk-taking (in order to check for faults to be attended to) and extreme comfort (in order to go ahead and attend to those faults). You have to be extreme at both ends. It's like having two completely contradictory personalities that you can switch on and off at a moment's notice. Indeed, there isn't necessarily any advantage to be being someone who falls somewhere in the middle. Balance is not obtained here by being neither quite one thing nor the other. That might give slightly better results than having an extreme tendency towards one side. However, you can also get stranded in middle ground. Such a practiser might be someone who neither truly explores their limits, nor ever slows down enough to fully solve problems. The balance of good practice is chiefly earned by being sufficiently extreme at both ends. Anyway, once you have checked that you can do any unit at various speeds (with both a few wild mistakes and patient corrections along the way), that's the first moment at which slow practice of an extended passage is something that you should genuinely want to congratulate yourself for. At this point you are indeed demonstrating self control- by restraining yourself from letting the fingers fly, and instead making time to listen to every detail (as if fascinated by the quality of each musical interval) within that established bigger picture. Slow practice of a bigger idea can now become a truly musical process and a chance to cement good quality movements.
Building the Bumblebee from basic principles
The Bumblebee might be a more difficult and advanced piece. Nevertheless, if anything it actually breaks down even more simply than the Bach example. Even if you don't consider yourself anywhere near advanced enough to play it, please continue reading anyway- if you want to appreciate how relatively simple the units can be, in advanced music.