Swing Timing: What the Notation Does Not Say
Jazz notation asks for swing and leaves the ratio to the players, because the timing is tempo-dependent and a fixed written value would be wrong at most speeds.
A written instruction like swing or medium swing asks a player to change the placement of notes that are, on the page, perfectly even. That looks like an instruction with missing information, and it is one, deliberately. The missing information is the ratio, and the reason it is left out is that the ratio is not a constant.
What the notation specifies
A jazz chart in common time usually writes eighth notes as eighth notes and prints swing once at the top, or above the first line, or in the part. What the player is expected to know is that each pair of eighths inside a beat becomes unequal: the first note takes more than half of the beat, the second less.
Where the feel has to be exact, scores write it out with the one symbol standard notation already has for dividing a beat in three, the triplet bracket. A written triplet pair with its first two notes tied gives the model most often taught: the first eighth occupies two thirds of the beat, the second one third. Some educational material uses a dotted eighth followed by a sixteenth instead, which is a different and more literal figure, and reading it as swing notation is a source of confusion rather than a shorthand worth adopting.
There is a matching instruction in the other direction. Charts that want the pairs played equally, mostly in Latin and funk contexts and in passages taken from them, mark the music even eighths or straight so that the swing default is cancelled.
What performance does
Recordings are the evidence here, and they agree on two things. The first is that the pair is genuinely unequal, which is why swing is not a myth of notation. The second is that the inequality is rarely the written two-to-one: the first note commonly takes somewhere between roughly 55 and 70 per cent of the beat, and the figure varies with tempo, with era and style, with the player, and with the position of the phrase.
The variation inside one performance matters as much as the average. A soloist can sit close to even for four bars and widen the pairs at a phrase ending, where the extra weight lands on the first note of a pair. Two players in the same band can disagree by a few per cent without anyone calling it out of time, because what they are sharing is a placement of the beat rather than a length of the eighth. The lengthening of an individual note is not the only mechanism either: the first note of a pair tends to be the accented one and the second tends to be lighter and shorter, so the feel is carried by weight as much as by ratio.
Where the ratio moves
Tempo is the variable with the clearest effect. As the beat gets faster the pairs flatten towards even, because the time available for inequality is disappearing, and at the fastest tempi a bebop line notated in plain eighths is played close to straight. The lilt does not vanish; it moves into the accent pattern, the articulation and the phrasing across the barline.
| Approximate tempo | Typical first note of the pair | What the pairs sound like |
|---|---|---|
| 60–90 bpm | near or beyond two thirds | a rolling limp, the basis of a shuffle |
| 100–160 bpm | roughly two thirds to three fifths | the familiar swing feel, both notes clearly present |
| 170–220 bpm | narrowing towards even | the lilt survives in the accents while the pairs even out |
| 230 bpm and above | close to even | the eighth line reads almost straight; swing lives in phrasing |
The boundaries are approximate and move with style. A medium-tempo blues and a medium-tempo bebop head will not place the pairs identically even when a metronome agrees about the tempo, which is part of why the same written line can sound like two different styles.
Three levels of division
Swing eighths are only one case of a general habit, and knowing which level is in play keeps the analysis honest.
At medium tempo the pair sits inside the beat, and the beat is divided into two unequal halves. When the music wants the division to be explicitly three-part, as in a shuffle or some blues feels, the triplet model becomes literal and can be written. At fast tempo the eighth is no longer the expressive unit: pairs of sixteenths inside the eighth carry the inequality, and the eighths themselves are close to even. In all three cases the underlying question is the same, which unit is being divided and how unevenly.
Confusing the levels produces a specific and recognisable error. A line of fast eighths played with a wide medium-tempo ratio sounds staggered rather than fast, because the ratio is being applied where the ear is not subdividing. The polyrhythm article describes the neighbouring case, where the division is not merely uneven but genuinely conflicting.
Eight pitches. Written as eighths they are equidistant on the page, and the swing instruction is a claim about where each pair sits inside its beat, not about the pitches.
Three notes that fill one beat evenly are a triplet. This is the shape the written model uses, and it is why the model is teachable: it converts a timing judgement into a countable figure that players can agree on before they have developed the judgement.
Why a fixed offset sounds mechanical
Sequencers offer to do the work automatically. Many express swing as a percentage of the pair, with an even division at 50 per cent and the triplet model near 67, and applying a value shifts every second eighth later in the beat.
Three things are missing from that operation. The first is tempo dependence: the same percentage at 80 and at 240 bpm produces two different feels, and only one of them is idiomatic. The second is accent and articulation, since a played pair is unequal in weight as well as in length, and moving a note without changing how it is attacked leaves the shape without its emphasis. The third, and the largest, is the reference. A player places the pair against something: the ride pattern, the bass line, the click of the hi-hat on two and four. The ratio is a by-product of leaning against that reference, and quantising to a global percentage removes the leaning while keeping the arithmetic.
The reference, not the ratio
The practical consequence for anyone writing or programming the music is that the ratio is the wrong thing to specify. What a chart needs to make clear is where the beat is and which subdivision is in play; what a player needs is a shared reference to place the pairs against, which is why practising with a metronome on two and four rather than on every beat produces a different and more useful result.
For a composer, three options are honest. Mark the feel and write even eighths, which trusts the idiom and keeps the page readable. Write triplets where the feel must be exact, and mark the passage even eighths where it must not swing. Or, if the music is going to be performed by a machine, express the feel as a groove parameter and accept that it will be a rendering choice rather than a specification, since the same passage played by a person will place the same notes differently on a different evening. The distinction between a written ratio and a played one is the same distinction that separates notated meter from heard meter: the page holds the category, and the performance supplies the value.
Questions
Does swing always mean a triplet feel?
No. The triplet is the model most often taught because it can be written and counted, but performances put the second eighth anywhere between a third and a half of the beat, and the value moves with tempo and style.
Why does a fixed swing setting in a sequencer sound mechanical?
The setting changes note lengths while leaving accents, articulation and the relationship to a groove untouched. Played swing is unequal in placement and unequal in weight at the same time.
