How to learn flashy runs

September 4, 2017, 2:28 PM · Okay, vulnerable technique question coming up.

One thing I've consistently struggled with (on violin and in piano–hey, Chopin!) is figuring out how to play runs or fast passages that are not easily subdivided. Somehow I neatly avoided learning how to do this back when I had access to great teachers. It's not something that comes up as often in orchestral or chamber repertoire–and in my sporadic lessons since high school I've worked on Bach, Mozart, and Beethoven. (Had I stayed another year with my high school teacher, we'd have tackled this head on with Rode and Sarasate–e.g. Rode 7, the 25-note run leading up to the B between rehearsal letters C & D in the first movement...or the first page of Zigeunerweisen. But, alas, I graduated and moved away not long after she suggested those pieces as the logical next step.)

As I see it, there are two challenges:
1) getting things up to speed (as with runs or fast passages in general), for which there are great established practice techniques, e.g. slow practice, rhythms, etc.). But I can't get through this part until I solve the more fundamental issue, which is
2) figuring out how the notes map evenly on top of the beat(s). (plus, a one-bow run played at .10x the speed doesn't really fit, right?)

Seriously, how *does* one fit 22, or 25, or however many notes there are, onto a beat? Do you subdivide the entire measure into 88-100 beats on the metronome and start there? Do you play them separately before attempting to fit them into 1-3 bows?


PS: Yes, obviously working with a teacher would be super helpful here.

Replies (29)

September 4, 2017, 2:34 PM · You have to organize the run in your head. I'm assuming you're talking about the odd groupings -- a run of 11 notes, for instance. I tend to group those into threes and fours, with the bigger groups at the top of the run (so an 11 is 3+4+4). At speed, the slight difference between the triplets and quads won't be noticeable at all.

Also remember that the most important parts of the run are the beginning and the end, so ensure that you get a clean start and that the last few notes are clear.

Edited: September 4, 2017, 5:18 PM · Organizing the run is necessary, but the "flash" is all in the bow distribution.

Dividing into groups of 3s and 4s as Lydia suggests is a good start (all you need in many cases.) Sometimes you divide in other ways, e.g. turn the first note into a pickup, or if starting slow and accelerating, you might do a 4+3+8, for example. It all depends. A thorough study of K23 is a good place to start. Compare the Hubay edition to the original notation to see the rhythmic rationalization (as a former new music teacher used to call it.)

Use as many bows as it takes to get the notes under your fingers. As you use less bows, even within a single bow, accent the beginning of each rhythmic division you made up within the slur, to coordinate how you spend the bow with the fingers. Carefully distribute the notes, fully spreading them out over the divisions of the bow you choose. At first use equal distribution to utilize every inch of the bow, gradually getting rid of the added accents, but still coordinating the rhythmic groups mentally. Then, to add the flash, use more bow for the important notes (usually the first and last notes, as Lydia said, but sometimes also for important middle notes.) Also, if changing bows for the next note after the run, make sure you connect the bow change by using the same speed into and out of the bow change (unless of course a sudden change in dynamic or articulation is marked.)

September 4, 2017, 9:02 PM · Thank you both! I don't think I ever played that Kreutzer. I like how this edition breaks it down. Rhythmic rationalization is exactly what I need.

Also just found a wealth in the archives: (specific to Zig) (Bruch) (re: Bach)

September 5, 2017, 8:48 AM · There are many threads on mastering fast passages, including an active one. Otherwise you're forcing people to just keep repeating themselves.
September 5, 2017, 9:15 AM · Ah, yes. The Google. How does that work? Also,wondering if I can play Accolay for my Juilliard audition, or do I need to learn Sibelius? Can I do it in six weeks? My teacher wasn't sure but I'm pretty good, if I do say so myself. Should I use a shoulder rest or just glue my instrument to my shirt? Will fake Dominants make my Eastman Gliga Forza edition violin sing or cry? What are prunes? I have so many questions!
Edited: September 5, 2017, 9:25 AM · PS: the existence of yet another active thread on how to master fast passages, wealth of archives notwithstanding, would suggest that this is a topic of enduring interest. Jeewon and Lydia answered my fairly specific question with useful tips. I appreciate their generosity.
September 5, 2017, 9:37 AM · Katie's problem (1) is covered by the other thread. This thread seems to be about problem (2). :-)
September 5, 2017, 11:02 AM · Depending on the piece you can probably take a bit of artistic liberty and focus more on the big beats, which was what I did on Bruch first movement. I never actually counted in 7 and ended up giving more length to the first note and sometimes the following longer "actual" note so it fit with the larger scheme of things. The only slow practice I did was going trough the runs in rhythm and chunking it away from the surrounding music so I wouldn't trip.
Edited: September 5, 2017, 11:26 AM · For number 2) I believe you need to make some approximation and bring out the notes that you feel are the most important or require the most emphasis, while at the same time feeling the underlying subdivided beat.

I'm sure a computer/synthesizer could subdivide the beat evenly into decimals but that would sound mechanical rather than musical.

September 5, 2017, 11:28 AM · Katie, your satyric response above it fantastic!
September 5, 2017, 1:33 PM · Jason, are you saying glue is not advised? ;-)

I think in general I struggle with complicated rhythms–e.g. playing 2 against 3. What I hear people saying is that complete rhythmic accuracy is less important than sticking the landing, so to speak–and fitting the run into its designated space, perhaps with rubato.

Edited: September 5, 2017, 7:13 PM · Here's a thread from a few years ago which covered the same info as the other current thread:

But last time around we didn't have Charles, Erik, Bo, Christian and Christian to offer their unique points of view, and Yixi to pose the question from her perspective. We all keep repeating ourselves over and over... it's part and parcel of contributing to a board on a finite topic, and, as it turns out, kind of the job of a teacher. Thankfully there are many ways to think about and express the same thing, and cater it to the one asking about it. Also, each time I repeat myself, I might've learned something new since the last time (at least I hope so.)

In terms of figuring out how to subdivide runs it's better to do it precisely, even if you end up bending it this way or that in performance. It's good to always feel the beat and the subdivisions, and to keep track of how the run fits. You will have more control over a phrase that way.

For cross rhythms you definitely have to do the math and internalize it. That's actually the proper application of rhythmic rationalization. To figure out cross rhythms you subdivide according to a common denominator. For 2 against 3, that would be 6. E.g. for two quarters against triplet quarters, write it out using the common subdivision, 6 eighths. For the duple, you get two tied triplet eighths. For the triple, you get three tied duplet eighths. To practice play 6 eighths, accenting the 1st and 4th, to feel the larger duples (>//>//), and accenting the 1st, 3rd, and 5th, to feel the larger triples (>/>/>/). Switch back and forth: >//>// >/>/>/ with a metronome set to the eighth pulse. Speed it up. Tie the eighths to feel the actual larger beats. Set the metronome to the larger 2 and play 3 against it (feeling the eighths subdivisions in your head.) Then set the metronome to the larger 3 and play 2. In a similar manner you can get very precise with 3 against 4, 2 against 5 or 7, etc. When you play contemporary rep. you must be able to play such cross rhythms accurately. There are cases prior to the 20th C where you need it also (perhaps the first prominent piece for us is Lalo Symphony Espagnol, for 2 against 3.)

Since this thread is specifically about fast runs, there are some very specific technical things we can talk about, which I'll try to put together for next time.

Edited: September 6, 2017, 9:50 AM · But before getting back to fast runs...

Complicated rhythms have been around a long time, especially the feel of 3 against 2, a.k.a. the hemiola. In the baroque, you might recall a piece which merrily rolls along in 3 and cadences in 3 groups of duples. Haydn, Mozart and Beethoven all loved to obfuscate the regular metre with hemiolas, syncopations and sometimes a combination of the two. In Beethoven and Brahms you have to contend with metres which are displaced by a beat, often overlapping with another voice on the beat. This is all common in folk music too, and gets most complicated in African rhythms. The trick is you have to train a steady subdivision, keep track of the beat, and move on your proper subdivision while feeling the beat.

1    2    
1 2 31 2 3
1 2 12 1 2
1   2  3  

The opening of Brahms concerto is notoriously difficult to keep steady and flowing. Some people just choose a fast tempo, which makes the groups of 5 easier. But to practice carefully it's useful to set the metronome to eighth pulses and accurately feel all the cross rhythms, and the proper relation between duples and triples, quads, quints, sextuples.

To that end, playing rhythmic accelerations to a metronome, e.g. Galamian's patterns (not just the usual acceleration everyone knows, but also that insert of rhythms which most of us don't like to look at, and probably conveniently lost within a week or two) is a great way to develop rhythmic control over the fingers, which is also necessary for fast runs.

[Apologies for the big gap. Don't know my HTML very well... :-< Anyone know how to fix it?]

September 6, 2017, 8:53 AM · Katie, if you look at Kreutzer 23 (cadenza study),

You will see the first run has 31 notes. I would start really slow and just do a manageable 4 notes per bow, which leaves you with a triplet at the end. For these runs, you also eventually want to have a sort of overall shape, so you can start slow and speed up a little, and do a ritard at the end, into the next phrase (just something to keep in the back of your mind at this point).

Then you can work your way up to groupings of 8,8,8, 7 notes per bow. Then you can work up to 16 (kind of a smoothed-out 4+4+4+4), 15 (kind of a smoothed-out 4+4+4+3). In any case, working slowly from small groups until you hear it, and then make bigger groups which don't sound like you are playing to a metronome. These runs still need to sound musical (make sense to the listener), and as far as I can tell (and this is my personal musing), we more easily process some big -uplet as a bunch of more digestible chunks grouped together.

From there, you kind of do that with all the runs. The triplets are obvious, and some other runs are a little interesting. You have to always approach each run from a musical standpoint, not just as an abstract number of notes to be played.

If you look at the first run of the second page, I divide it (2+3+3), then (4+4), then (4+6), where the last 6 notes is faster than any of the preceding notes of the run. There don't seem to be a ton of options that make sense to me on that one, but I like for the phrase to start slow and crest at the top note and slightly ritard.

The next run I shape towards the top notes, then build up some speed going down, and then slow down and group the last 5 notes.

Thinking this out, I'm not sure how rigorous my "process" is, but I think that working in smaller chunks and making sure that the run makes musical sense allows you build up to a whole run where all the notes are well-placed and still organic.

Consider also subdividing your scales in a way where you practice with 5-tuplets, 7-tuplets, etc. Now that I think about it, this is probably where you do the work of figuring these -uptlets out, because you can play them dividing the beat evenly across all notes, or you can play them, chunking them into groups of 2,3,4.

Anyhoo, I hope my word salad has some nutrition today!

Edited: September 6, 2017, 9:17 AM · My skill level is surely lower than Katie's but it's a challenge to play 12 notes in the same bow and to hear them all clearly (to say nothing of twice as many), so the prospect of emphasizing, subdividing, or distributing my bow unevenly seems especially daunting. I don't know why it should be so, but the severity of the problem depends greatly on what my left hand has to do. The latter could simply be a mental bandwidth issue -- after taking care of my left hand, not enough brain cells left to deal with tone.

I think Kreutzer No. 23 would probably put me in my grave, so I appreciate Christian's permission to subdivide. I have worked on No. 29 and that's probably the level where I need to be now. I even subdivided that to learn the LH.

September 6, 2017, 9:14 AM · One idea I like: you can fit more notes at the frog than you can at the tip. :)
September 6, 2017, 10:07 AM · Actually, I think you can fit the most notes in the middle of the bow. Hence the general bowing rule, spend-save-spend, for long runs.
September 6, 2017, 3:12 PM · Hahaha, Paul - Granted. I have trouble playing all those runs on one bow (and it will be a while before I could put up a recording of myself on youtube that I wouldn't completely cringe at), so I think it makes sense to divide as necessary, with the eventual goal being playing the runs on one bow (or at least on two). But if you can make it sound better dividing, then why not? No reason to fetishize long bows if they actually make it sound worse.

That's a bear of a study (I must have spent at least two months on it before my teacher let me move on), but I could see a progression of skills leading up to K23. K1 and K23 seem to be very intimately linked, and before that, practicing son file, and then grafting that long bow onto Schradieck studies would seem to be training the two independent skills of really fast, light fingers together with a heavy, but relaxed bow arm.

September 6, 2017, 3:18 PM · Son file has been a revelation for me.

Christian, I like the idea of playing scales in odd numbered segments per beat. Will try it this afternoon.

Jeewon, I do know to subdivide in theory–-indeed, it was the only way I was able to get right and left hands in sync in Debussy's "Reverie"!--but haven't had to use that technique on the violin yet...much. I will dive in.

Thank you all!

Edited: September 6, 2017, 3:43 PM · Jeewon's advice on spend-save-spend for long runs is insightful as usual. Most teachers I've met would tell us to save from the frog when a down bow starts. I notice that, during last July's three-week summer string academy and boot camp at our local conservatory, it's common even among advanced young players started too fast at the beginning of a down bow. So teachers would usually call out "slow!" when students just started from the frog. However, comes to think of it, save at the frog wouldn't sound good and that's why we tend to move too fast at the beginning of the down bow. What needs to be remembered is to save at the lower part of the bow immediately beyond the frog.
Edited: September 7, 2017, 11:13 PM · "What needs to be remembered is to save at the lower part of the bow immediately beyond the frog." Yes! That's a better way to describe it. Almost like a fp without the p or the f, necessarily.

Katie, I forgot you mentioned Chopin off the top :) These days, now that I don't really care much what it looks like, and having been inspired by the greatest toe tapper I've ever seen live, Roger Tapping, former violist with the Takács (yes, for reals,) I just stomp my foot to feel the "counter" to my rhythm whenever I want (most recently for Britten's 1st quartet... well, it wasn't quite a stomp, more of an internal toe twitch... but I couldn't have done without feeling the beat somehow, anyhow.) But... and it's a doozy of a but... the foot must be rhythmic (gotta practice tapping in time too, if you're gonna go there) And I think it all starts with simply paying attention to synchrony. If one fails to develop the sense of "being with," no amount of metronome practice will help play in time, rhythmically, and with others.

The way I introduced it was to simply get students to say "1" and move the bow precisely as they spoke. Usually, the bow would move first for most, those not used to executing with the mind. Though difficult to do, counting beats and &s of beats while playing is a useful exercise, even if done slowly and piecemeal. The action must follow the command, not move on it's own, or by rote. Some exercises my old teacher used to get us to count:

1) shifts (large and small): play initial note on 1, on 2 release pressure and start to glide on the 'surface' of the string, counting over 8 beats, arriving at the target note precisely on the next 1, and shifting back down; you can also add a free vibrato on the 1's

2) finger lifts: start with the finger on the surface of the string; on 1 lift finger (into it's curl, not straight) as quickly and forcefully as possible from the base knuckle; then slowly, gradually, continuously lower the finger back onto the surface of the string over the remaining 2, 3, 4, then on the next 1, "bam" lift explosively (without having pressed the string at all)

3) string crosses: play lower string on 1 and slowly, gradually, continuously cross to the next string over 2, 3, 4, playing the higher string precisely on the next one (you can also use double stops as an arrival point, cross to lower strings, cross 2, 3, all strings, see Ysaye scales, use on arpeggios, Dont Op. 37 string cross exercises, etc.)

4) we were never made to do this, though it was implied, but as an extension of the "commanding the bow to move on 1" exercise mentioned above, you can practice divisions of the bow, in stopped bowing, parlando, legato, all the while counting the divisions of the bow

You can do all sorts of counting exercises, ultimately using a metronome to count for you (but to reiterate, metronome work is useless without rhythmic control.) The next step is to count, feel, grunt, lurch, on e.g. tied notes, dotted rhythms, cross rhythms, etc.

1 2& 3 (count, grunt, lurch, stomp, walk on 1 and on & to feel in 2 while playing in 3)

1 unh!2 (count, grunt, lurch, stomp, walk on 1 and unh! to feel in 3 while playing in 2)

All this to say, to internalize rhythm (without the advantage of the other hand in similar motion as on a keyboard) requires a lot of mental control (which you probably already knew but I thought I might spell out.) But once you have such control you truly feel the rhythm of every motion, which is what you need for complicated patterns, and speed.

September 7, 2017, 9:12 PM · Short answer, about fast runs, to quote my orchestra manager; "Play the first note then add more notes, land on the last note"
Don't try to chop it up accurately. What works often enough, I am thinking of that Brahms opening, is to start a little slow then accelerate and crescendo into the target note, which is usually on a strong beat.
Edited: September 7, 2017, 11:23 PM · We'll have to agree to disagree, Joel. Touring artists are often at liberty to do whatever they want but that doesn't mean their choices are always convincing. I suspect there are more who play the opening quintuplet arpeggio strictly. Of those who add rubatos, few manage to pull it off without sounding all scrambly. (There is no scrambling in Brahms. And let's not even mention those who continue the rubato into the next measure!)

Regarding your manager's strategy, it might do for a section player hiding behind others who can articulate every note, but doesn't cut it for a solo or chamber performance, or an audition.

September 7, 2017, 11:58 PM · Thanks Jeewon, I was being partly facetious. Anyway, sometimes we over-think things, forgetting that that paper in front of us is not real music, but a (rather excellent) notation system. After the technical problems are worked out, after some years of experience and study, a musician can start to trust their own instincts, hear the piece in their head , and go for it. Regards, jq
Edited: September 8, 2017, 6:06 AM · Ultimately I suppose that's all we strive for. But the question remains how do we get from here to there. I'm pretty sure it's not by the seat of our pants.

As for the score, it will forever be debated whether one's own musical instincts take precedence over the encoded musical ideas the composer left for us. But you have to take into consideration who you're decoding. Brahms could sometimes be infuriatingly vague in his markings. Yet at other times he was pointedly explicit (if I remember correctly he may have been the first to indicate a metric modulation, buried in the middle of a score.) Here we have a D minor context which gives the fiddle, not a straight D minor scale or arpeggio, but a double harmonic, quasi scale/arpeggio. He could've written it 4 + 6, or used more notes, turning it into 2 + 3 + 6, or even 3 + 4 + 6 or more. He could've indicated tempo rubato, or some indication to take liberties in this clearly cadenza-like solo entrance. But I don't think he was that kinda guy. Though he wasn't a stickler for tempo in terms of metronome markings, he sure liked to add modifiers to his tempo indications, and make changes in indications and metre within a movement. And rhythmically he liked to put a straight jacket on the ensemble as a whole. So sure, the music is not in the notation per se, but where do we as performers draw it from.

P.S. by "opening of the Brahms" I was referring to the whole developy section from the entrance of the violin leading to the restatement of the 1st theme proper.

Edited: September 8, 2017, 6:54 AM · Today we think of rubato almost solely as bending time, giving and taking, playing around with tempo freely. We think evenly distributing an odd number of notes is reserved for 20th century, weird stuff. But back in the day, the ability to play rubato "contrametrically," e.g. 11 against 2, was considered a sign of mastery. Furthermore:

Use of freely shifting contrametric rubato continued into at least the first half of the 19th century. Louis Adam described it without its name in his influential Mithode de piano de Conservatoire. In a section on orchestral playing in Ludwig Spohr's Violinschule, published in 1832, the reader is advised that when a soloist plays in tempo rubato "the accompaniment must continue its even, measured course;" and the well-known critic Henry Chorley described soprano Giuditta Pasta's admirable sense "for the measurement and proportion of time." This included "metronomic correctness" balanced by an "artful licence in giving and taking."

See also the interesting stuff about Chopin's contrametric rubato on pp. 41 and 42. In some ways, considering their tastes and temperaments, Brahms may have been more of a classical classicist than Beethoven, his musical forebear.

Brahms treated tempo with more restraint than Wagner, but he also applied some agogic rubato when performing. His conducting score of Ein deutsches Requiem contains many handwritten signs for pushing or holding back the tempo and, in a letter of 20 [?] January 1886 to his close friend Joseph Joachim, Brahms discussed the similar tempo modifications he had entered into the score of the Fourth Symphony. However, the gulf between the performance styles of Wagner and Brahms was large and Brahms's tempos were described as "painfully dry, inflexible, and wooden" by Wagner, who would have preferred to hear Brahms's technique "anointed with a little oil of Liszt's school." According to Louis Kohler, himself a respected pianist and teacher, the virtuosic imitators of Liszt used a "senseless accelerando and ritardando" and "constant rubato, which leaves one uncertain whether it is the player or the hearer whose head is turned. [p. 49]
Edited: September 8, 2017, 10:30 AM · Jeewon, how I wish you lived within 40 miles of Menlo Park, California!
September 8, 2017, 10:00 AM · Or move to BC!
September 8, 2017, 11:38 PM · I've always felt more like a west coaster :) So much closer to Hawaii!

So getting back to runs...

In Katie's examples I think we can identify two kinds of runs, those which occur as part of passagework, in the main body of the music, and those which kind of stand alone, as in Zig, flashy outbursts in between the main passages. It's more important to organize passages according to a pulse, whether into triplets and duples/quads, or spread out more evenly, when the runs are part of a passage.

In the stand alone runs, it's still useful to group notes, especially when practicing under tempo, but rather than grouping according to a steady pulse, keep the note value steady and change the metre within the run, as it were. Up to speed I think it really is however many notes in 1 beat. And when you're comfortable playing changing metres you get used to irregular groupings, so odd number runs feel as normal as ones which divide neatly into beats.

Either way, there are two elements we need to stitch together position work + shift work. For position work we need to smooth out the string crosses, keeping track of how many notes there are per string, making a smooth cross. It's useful to do open string exercises, creating a rhythm according to the string cross pattern. Start with bowings as needed, then as written. Use added accents to mark the crosses, then make them as seamless as possible, crossing smoothly and without any change in bow speed, i.e. use even subdivisions for control, then with whatever uneven subdivisions you need for phrasing and dynamics.

For shift work it's useful to figure out the shift pattern which forms the skeleton of the shifting portion of the run. Any scale can be reduced to such a skeletal pattern, a 1 finger arpeggio of sorts, which helps organize middle and final positions, arm v. pivot shifts, position playing v. extensions/contractions.

Next time, an example.

Facebook Twitter YouTube Instagram Email is made possible by...

Shar Music

Pirastro Strings

Coda Bow

Violin Lounge Academy

Yamaha V3 Series Violin

Dimitri Musafia, Master Maker of Violin and Viola Cases

Metzler Violin Shop

Gliga Violins

Corilon Violins

Leatherwood Bespoke Rosin

Anderson Musical Instrument Insurance

Bobelock Cases


Nazareth Gevorkian Violins

Los Angeles Violin Shop

Pro-Am Strings

Wangbow Violin Bow Workshop