Legato on Strings

There is a very common misunderstanding and misconception about how legato on string instruments work and it is sometimes even seen in the works of seasoned orchestrators. This whole issue is something that I would put in the Top 3 of the most common errors that happen with learning orchestrators and composers.

Now even if a lot of the things that I'm going to say in this post apply to notating music and score sheets, understanding the background behind it is also benefitial in the virtual world.

So what is the issue?

The term legato refers to a playing style where the notes are connected with each other and flow as smoothly as possible from one to the next without any audible interruption between the notes. For everybody coming from a piano or guitar background, this playing style is indicated through the use of a slur:

The common understanding is that all the notes that are under one slur should be interpreted as one musical phrase. With woodwind or brass instruments, there often is the correlation between a slur and a breath, that the end of the slur also indicates the moment to breathe.

For string instruments however, slurs have a fundamentally different meaning: all the notes that are covered under one slur are to be played under one bow. With strings there are two bow directions: downbow - starting at the bottom end of the bow and drag it down over the strings and upbow doing the opposite. While this one motion happens, the left hand can play different notes on the fingerboard. If it is still unclear, here is a video that demonstrates the technique.

Consequentially whenever a slur ends, the player will change the bow direction.

This fundamental concept seems relatively simple, however there are a few more parameters that come into play:

1. The time a bow length will last is extremely depending on the dynamic:

In pp on a violin, a bow can last more than 10 seconds, at mf around 3-4 seconds and at ff it might not even last one second. These numbers get shorter the larger the instruments get. For instance, in proportion, the cello bow is considerably smaller to its instrument so it will not last as long.

2. String players prefer to play downbows on downbeats which means that they default to play a downbow with every new bar.

3. String players try to even out the intensity between down- and upbow but from a physical standpoint, downbows are more suited for a hard accent than upbows just based purely on the leverage the player has with the bow.

These three factors have a massive influence on how legato is executed on string instruments.

Obviously, the notated example above will not work as the length of the phrase will exceed the time that one bow will last. So, the next step would be to slur it bar wise:

This would technically work and would be absolutely appropriate in soft dynamics but the second point of downbow on downbeats makes this phrase a bit counterintuitve as it would need to be executed like this:

The little "V" stand for upbow, the other symbol for downbow. So particularly the last, concluding note would feel relatively awkward for the player to play on an upbow. The more natural way of phrasing this would be:

With this, we have two bowstrokes in each bar which would always land a downbow on a downbeat.

You might ask now what happens between the bow changes and whether there will be an audible gap between the notes. This is what is called a "bow change legato". String players practice their whole life to make this gap as inaudible as possible and with the reverb of the room and the slight differences between the individual players of the section, there will not be an audible gap between the notes. However, a bow change legato is indeed not as fluid as a "fingred" legato that happens under one bow, especially when it happens on the same string.

With the advent of legato samples in the virtual world, unfortunately there also appeared the "portamento" samples in string libraries: a playing technique where there is an audible glide from one note to the next note. And unfortunately, some composers found these to be their favourite string patches which results in pieces that sound like a musical analogy of a mouthful of chewing gum. The default for string players is to do their best to make the transition/slide between notes as inaudible as possible. The ocassional small slide or even portamento adds quite a bit of expressiveness and romanticism to the playing but there is a limit of when this transitions from tasteful to cheesy and it is quite low.

As mentioned above, the dynamic degree has a strong influence on how long a bow will last:

Here in this row of 16th notes, at this tempo, it is not a problem to cover one bar with one bow and as the figure is relatively static, it is also okay to not have a downbow on the second bar. However, at mf, you would most likely go to this:

Even though, half a bar at that tempo lasts two seconds and one bar would theoretically be possible with one bow length (at four seconds), it is advisable to change bows more often than what is technically necessary as this will not stress out the players to use their bow as efficiently as possible just to reach the end of the slur.

With even higher degrees of dynamics, you would need to use even shorter slurs:

Every bow change has the potential of creating a little accent on the next note, which you can use in your favor when writing in slurs:

There is an even number of 6 bows in each bar which helps to place the downbow on the downbeat. In general, it is advisable to try to get an even number of bows into each bar, even if from the possibility of the bowing, it isn't necessary.

This figure would indicate three bows in each bar:

This is of course not ideal. So there are two ways to solve that problem:

In this case, breaking up the legato group brings the player back to a downbow on the next bar. However, it might indeed be desirable to leave these 4 note groups under one slur intact, in which case, you could also go for these options:

If there is time between the notes, it is not a problem for the player to repeat the same bowstroke twice. In these cases, the marking of the bow strokes would be advisable. The difference between the two bars would be that the two downbows in the first bar have a bit more weight than the alternation in the second one.

If there is no time to repeat the bowstroke or you don't want an audible gap, you should go for breaking the 8th note group:

Of course, you could also slur the first two quarter notes but in some cases you might want them to have the weight of a separate bow stroke, like for instance in this ff passage.

No matter how well you understand these things, as a non string player, there will always be instances where your players might change the bowing. Usually this happens because it is more convenient to play. Usually the section leader will communicate these changes and all the players will transfer them to their parts. In an ideal world, all bows of one string section should visually move simulatenously up or down when watching them play. However, in session situations, especially on staccato/spiccato passages or passages where articulations switch alot, you can quite often see almost every player using a different bowing. This is usually nothing to worry about as they will try to individually do their best to translate the notes on the sheet as best as possible. It is however quite interesting to collect a few string parts after the session and look at the written in changes as it is quite educational to see how the performance differed from what you planned to. Most often however, if you as orchestrator or composer planned out the bowing properly, the only marks you will see in parts might be upbow and downbow marks as reminders for the players. In fact, even John Williams regularly meets up with his concert master before a session to go through the bowing of the string parts.

There are three common questions in this regard that I'd like to address here:

Can't I just write "legato" on a passages and let the players do their thing?

Theoretically you can, and it will probably get you a solid result, however it doesn't really show the players that you know what you're doing and feels rather unprofessional. So generally try to avoid that.

I like the "piano" way where slurs indicate a musical phrase and I want to have that also written in the strings. Can I use double slurs? Like one slur that indicates the bowing and one larger slur that indicates the phrase?

Do not use double slurs! Your players have two ears and are trained musicians, they will get the phrasing of a musical idea. Double slurs clutter the page unnecessarily and don't add any benefitial information to the music.

I often double for instance Violins 1 and Flute 1, should I really use different sluring on an identical line? Like Flute 1 sluring indicates the breathing/phrase, and violin 1 the bowing?

Yes, you should! It is quite common to use different sluring on different instruments so you shouldn't worry about "identical" parts being notated differently on different instruments.

It does pay off to invest some time getting these things to a state that doesn't need much changes as this will of course also save session time.