In jazz, when the music calls for a Cmaj7 chord, this almost never implies a pianist should play “C E G B”. Usually, the pianist will choose some other way of playing the chord, even if it is simply an inversion of the basic root position chord. There have been entire books written on the subject of chord voicings. The discussion here only scratches at the surface of the possibilities. I have loosely categorized the voicings described here as 3/7 voicings, quartal voicings, polychord voicings, close position and drop voicings, and other scale based voicings.
It is somewhat of a shame that the most common type of voicing used by most pianists since the 1950’s has no well established name. I have seen these type of voicings called Category A and Category B voicings, Bill Evans voicings, or simply left hand voicings. Because they are based on the third and seventh of the associated chord, I call them 3/7 voicings.
The basis of these voicings is that they contain both the third and seventh of the chord, usually with at least one or two other notes as well, and either the third or the seventh is at the bottom. Because the third and the seventh are the most important notes that define the quality of a chord, these rules almost always produce good sounding results. Also, these voicings can automatically produce good voice leading, meaning that when they are used in a chord progression, there is very little movement between voicings. Often, the same notes can be preserved from one voicing to the next, or at most, a note may have to move by step.
For instance, consider a ii-V-I progression in C major. The chords are Dm7, G7, and Cmaj7. The simplest form of the 3/7 voicing on this progression would be to play the Dm7 as “F C”, the G7 as “F B”, and the Cmaj7 as “E B”. Note that in the first chord, the third is at the bottom; in the second chord, the seventh is at the bottom; in the third chord, the third is at the bottom. Also note that, when moving from one voicing to the next, only one note changes; the other notes stay constant. This is an important characteristic of 3/7 voicings: when they are used in a ii-V-I progression, or any progression in which root movement is by fourth or fifth, you alternate between the third and the seventh at the bottom. An analogous set of voicings is obtained by starting with the seventh at the bottom: “C F”, “B F”, “B E”.
Normally, you would use more than just the third and seventh. Often, the added notes are the sixth (or thirteenth) and ninth. For example, the C major ii-V-I could be played as “F C E”, “F B E”, “E B D”, or as “F A C E”, “F A B E”, “E A B D”. The added notes are all sixths or ninths, except for a fifth in the first chord of the second example. When playing these four note voicings on guitar, any added notes will usually be added above the third and the seventh, or else your voicing may end up containing several small intervals, which is usually possible to play only with difficult hand contortions. Thus, the C major ii-V-I might be played with four note voicings on guitar as “F C E A”, “F B E A”, “E B D A”.
Note that none of these voicings contain the roots of their respective chords. It is assumed that the bass player will play the root at some time. In the absence of a bassist, pianists will often play the root in their left hand on the first beat, and then one of these voicings on the second or third beats. Actually, you can often get away with not playing the root at all; in many situations, the ear anticipates the chord progression and provides the proper context for the voicing even without the root. It is not forbidden to play the roots in these voicings, but it is neither required nor necessarily better to do so.
These basic voicings can be modified in several ways. Sometimes, you may wish to omit either the third or the seventh. Often, a minor of major chord that is serving as a tonic will be voiced with the third, sixth, and ninth, and these voicings might be interspersed with regular 3/7 voicings. Also, voicings with the fifth or some other note at the bottom can be interspersed with true 3/7 voicings. This might done for any of several reasons. For one thing, when played on the piano, note the voicings described thus far all tend to slide down the keyboard as the roots resolve downward by fifth. The normal range for these voicings is in the two octaves from the C below middle C on the piano to the C above middle C. As the voicings settle downward, they will start to sound muddy, at which time you might want to jump up. For instance, if you have ended up on a Dm7 as “C F A B” below middle C, and need to resolve to G7 and then Cmaj7, you might want to play these two chords as “D F G B” and “E A B D” respectively to move the voicing upward while preserving good voice leading. Also, roots do not always move by fifths; in a progression such as Cmaj7 to A7, you might want to voice this as “G B C E” to “G B C# F#” to preserve good voice leading.
One thing to note about these voicings in the context of a diatonic ii-V-I is that, because the chords imply modes of the same scale (D dorian is the same as G mixolydian is the same as C major), a given voicing can sometimes be ambiguous. For example, “F A B E” might be either a Dm7 with the seventh omitted, or a G7. In the context of a modal tune like “So What”, it clearly defines the Dm7 or D dorian sound. In the context of a ii-V progression, it probably sounds more like a G7. You can use this ambiguity to your advantage by making one voicing stretch over several chords. This technique is especially useful when applied to the more general scale based voicings discussed later.
Another thing you can do with 3/7 voicings is alter them with raised or lowered fifths or ninths. For instance, if the G7 chord is altered to a G7b9 chord, then it might be voiced as “F Ab B E”. In general, the notes in the voicing should come from the scale implied by the chord.
These voicings are well suited on the piano for playing in the left hand while the right hand is soloing. They can also be played with two hands, or with all strings on a guitar, by adding more notes. This provides a fuller sound when accompanying other soloists. One way to add more notes is to choose a note from the scale not already in the basic voicing and play it in octaves above the basic voicing. For instance, on piano, for Dm7 with “F A C E” in the left hand, you might play “D D” or “G G” in the right. In general, it is a good idea to avoid doubling notes in voicings, since the fullest sound is usually achieved by playing as many different notes as possible, but the right hand octave sounds good in this context. The note a fourth or fifth above the bottom of the octave can often be added as well. For example, with the same left hand as before, you might play “D G D” or “G D G” in the right hand.
The 3/7 voicings are perhaps the most important family of voicings, and many variations are possible. You should try to practice many permutations of each in many different keys.
A style of voicing made popular by McCoy Tyner is based on the interval of the fourth. This type of voicing is used most often in modal music. To construct a quartal voicing, simply take any note in the scale associated with the chord, and add the note a fourth above, and a fourth above that. Use perfect fourths or augmented fourths depending on which note is in the scale. For instance, quartal voicings for Cm7 are “C F Bb”, “D G C”, “Eb A D” (note the augmented fourth), “F Bb Eb”, “G C F”, “A D G”, and “Bb Eb A”. This type of voicing seems to work especially well for minor chords (dorian mode), or dominant chords where a suspended or pentatonic sound is being used.
These voicings are even more ambiguous, in that a given three note quartal voicing can sound like a voicing for any number of different chords. There is nothing wrong with this. However, if you wish to reinforce the particular chord/scale you are playing, one way to do this is to move the voicing around the scale in parallel motion. If there are eight beats of a given chord, you may play one of these voicings for the first few beats, then move it up a step for a few more beats. The technique of alternating the voicing with the root in the bass, or the root and fifth, works well here, too. On a long Cm7 chord, for instance, you might play “C G” on the first beat, then play some quartal voicings in parallel motion for the duration of the chord.
As with the 3/7 voicings, these voicings are convenient left hand voicings on the piano or three or four string voicings on the guitar. They can also be made into two handed or five or six string voicings by stacking more fourths, fifths or octaves on top. For instance, the Cm7 chord can be voiced as “D G C” in the left hand and “F Bb Eb” in the right, or “Eb A D” in the left and “G C G” in the right. The tune “So What” from the album Kind Of Blue used voicings consisting of three fourths and a major third. On a Dm7 chord, the voicings used were “E A D G B” and “D G C F A”.
The basis of a polychord voicing is to play two different chords at the same time, such as one in the left hand and one in the right on a piano. The relationship between the two chords determines the quality of the resultant chord. These are always two handed voicings on a piano, or five or six string voicings on the guitar. They produce a very rich, complex sound compared to the voicings presented so far.
The simplest style of polychord voicing is to play two triads; for instance, a C major triad in the left hand on a piano, and a D major triad in the right. This will be notated D/C. This notation is overloaded in that it is usually interpreted as meaning a D triad over the single note C in the bass; it is not always clear when a polychord is intended. Polychords are seldom explicitly called for in written music, so there is no standard way to notate them. You must normally find your own opportunities to play polychords.
If you take all the notes in this D/C voicing and lay them in a row, you will see that this describes either the C lydian or C lydian dominant scales. Therefore, this voicing can be used over any chord for which those scales are appropriate. If you experiment with other triads over a C major triad, you will find several combinations that sound good and describe well known scales. However, many of these combinations involve doubled notes, which can be avoided as described below. Among the polychords that do not involve doubled notes are Gb/C, which produces a C HW diminished scale, Bb/C, which produces a C mixolydian scale, Dm/C, which produces a C major or C mixolydian scale, Ebm/C, which produces a C HW diminished scale, F#m/C, which also produces a C HW diminished scale, and Bm/C, which produces a C lydian scale. These polychords may be used as voicings for any chords that fit the corresponding scales.
You may have noticed that Db/C, Abm/C, Bbm/C, and B/C also involve no doubled notes and sound very interesting, although they do not obviously describe any standard scales. There are no rules for when these polychords may be played as voicings. When your ear becomes accustomed to the particular nuances and dissonances of each, you may find situations in which you can use them. For example, the last polychord listed, B/C, sounds good when used as a substitute for Cmaj7, particularly in the context of a ii-V-I progression, and especially at the end of a song. You may resolve it to a normal Cmaj7 voicing if you wish.
You can construct similar polychords with a minor triad at the bottom. Db/Cm produces a C phrygian scale; F/Cm produces a C dorian scale; Fm/Cm produces a C minor scale; A/Cm produces a C HW diminished scale; Bb/Cm produces a C dorian scale; and Bbm/Cm produces a C phrygian scale. In addition, D/Cm produces an interesting, bluesy sounding scale.
I mentioned before the desire to avoid doubled notes. One way to construct polychords that avoid doubled notes is to replace the triad at the bottom with either the third and seventh, the root and seventh, or the root and third of a dominant chord. Voicings constructed in this fashion are also called upper structures. They always imply some sort of dominant chord.
For example, there are several possible C7 upper structures. A Dbm triad over “C Bb” yields a C7b9#5 chord. A D triad over “E Bb” yields a C7#11 chord. An Eb triad over “C E” yields a C7#9 chord. An F# triad over “C E” yields a C7b9b5 chord. An F#m triad over “E Bb” yields a C7b9b5 chord. An Ab triad over “E Bb” yields a C7#9#5 chord. An A triad over “C Bb” yields a C7b9 chord.
You will find it takes a lot of practice to become familiar enough with these voicings to be able to play them on demand. You may wish to choose a few tunes and plan ahead of time where you will use these voicings. It is well worth the effort. The richness and variety introduced by these voicings can add a lot to your harmonic vocabulary.
The simplest voicing for a four note chord is the close position voicing, in which all the notes in the chord are arranged as close together as possible. For example, a C7 chord might be voiced in close position as “C E G Bb”. This is referred to as root position, since the root, C, is at the bottom. The chord might also be voiced in close position as “E G Bb C”, which is also called the first inversion, since the bottom note has been inverted to the top. The second inversion is “G Bb C E” and the third “Bb C E G”.
A drop voicing is created from a close position voicing by dropping one of the notes down an octave. If the second note from the top is dropped, the voicing is called a drop 2 voicing; if the third note from the top is dropped, the voicing is called a drop 3 voicing. For a C7 chord in root position, “C E G Bb”, the corresponding drop 2 voicing is “G C E Bb”. The second note from the top, G, has been dropped down an octave. The corresponding drop 3 voicing would be “E C G Bb”. Drop 2 and drop 3 voicings can be constructed from any of the inversions of the chord as well. On the piano, the dropped note must normally be played in the left hand, so these are almost always two handed voicings. The intervals in these voicings make them perfectly suited for guitar.
Close position and drop voicings are effective when used to harmonize a melody, particularly in a solo setting. Each melody note may be harmonized by a different drop voicing, with the melody note on top. Pianists and guitarists often use this type of approach in their own solos. A phrase in which every note is accompanied by close position or drop voicings is said to be harmonized with block chords. Red Garland, Dave Brubeck, and Wes Montgomery all regularly played block chord solos.
There are other logical ways of constructing voicings; too many to describe individually here. Most approaches are similar in that they they associate a scale with each chord and construct the voicing from notes in that scale. By using a scale approach, you can devise your own patterns for voicings. For instance, a second with a third stacked on top is a somewhat dissonant but not too cluttered sound that many pianists use extensively. For a chord such as Fmaj7, you can apply this format at any position in the associated F lydian or F major scale. Since the F major scale contains an avoid note (Bb) in this context, one would normally opt for the lydian scale and the B natural, so that none of the generated voicings would contain any avoid notes. The particular pattern described above yields “F G B”, “G A C”, “A B D”, “B C E”, “C D F”, “D E G”, and “E F A” over the F lydian scale.
Most of these voicings are very ambiguous, in the sense that they do not readily identify the chord. As with the 3/7 and quartal voicings, however, you will find that the presence of a bass player, or just the context of the chord progression being played, will allow almost any combination of notes from a given scale to make an acceptable voicing for the associated chord.
You may wish to experiment with different patterns and different scales to see if you can find any voicings you particularly like. Often, the goal is not to find a voicing that completely describes a given chord, but rather to find a voicing that conveys a particular sound without seriously corrupting the chord. You may find that at a given point in the music, you may wish to hear the characteristic authority of a perfect fifth, or the characteristic dissonance of a minor ninth or of a cluster of several notes a second apart, but without the characteristic wrong note sound of a completely random selection of notes. Thinking of the associated scale and putting your sound into that context gives you a logical and reliable way to get the sound you want without compromising the harmony.
An accompanist may occasionally reharmonize a chord progression to sustain interest, introduce contrast, or create tension. This involves replacing some of the written or expected chords with other unexpected chords. Substitutions such as the tritone substitution are one type of reharmonization.
Some musicians spend a lot of time trying different reharmonizations when working on a tune. However, unless they tell the soloist what they doing beforehand, many of the reharmonizations they may come up with are not suitable for use in accompanying, since the soloist will be playing from a different set of changes. There are some simple reharmonizations that can be used without disturbing the soloist too much. The tritone substitutionis one example; at any time a dominant seventh chord is called for, the accompanist may substitute the dominant seventh chord a tritone away. This creates exactly the same type of tension that is created when the soloist performs the substitution. Another simple reharmonization is to change the chord quality. That is, play a D7alt in place of a Dm, and so forth.
Another common reharmonization is to replace a dominant chord with a ii-V progression. This was already demonstrated when discussing the blues progression; one of the progressions replaced the F7 chord in bar 4 with a Cm7 – F7. This is especially common at the end of a phrase, leading to the tonic at the start of the next phrase. Most of the scale choices the soloist may have been using over the F7 chord will also work over the Cm7 chord, so this reharmonization doesn’t usually create too much tension. This technique can be combined with the tritone substitution to create a more complex reharmonization. Rather than replace the V with a ii-V, first replace the V with its tritone substitution, and then replace that with a ii-V. For example, in bar 4 of the F blues, first replace the F7 with B7, and then replace that with F#m7 – B7.
Another type of reharmonization involves the use of alternation. Rather than play several measures of a given chord, the accompanist may alternate between it and the chord a half step above or below, or a dominant chord a fifth below. For instance, on a G7 chord, you might alternate between G7 and Ab7, or between G7 and F#7, or between G7 and D7. This is especially common in rock based styles, where the alternation is performed in rhythm. If the alternation is performed regularly, such as throughout an entire chorus, or even the whole tune, the soloist should be able to pick up on it and control the amount of tension produced by playing along with the reharmonization or by playing against it. That is, the soloist can lessen the tension by changing scales as you change chords, or increase tension by keeping to the original scale.
Once you have decided what notes you want to play, you must decide when to play them. You do not want to simply play whole notes or half notes; your accompanying generally should be rhythmically interesting, although not distracting to the soloist or listener.
There are few guidelines that can be given for playing comping rhythms. Because there is very little theory to fall back on, the first piece of advice I can give is to listen to other accompanists. Too often we tend to ignore everyone but the soloist anyhow. Be sure to choose albums that have solo instrumentalists other than the accompanist on them. Pianists to listen to include Bud Powell, Thelonious Monk, Horace Silver, Bill Evans, Wynton Kelly, Herbie Hancock, and McCoy Tyner. Pianists should also listen to guitarists and mallet players; often the constraints of those instruments can lead to ideas you might not have thought of otherwise.
Guitarists should listen to pianists, but also to guitarists such as Herb Ellis, Joe Pass, and Wes Montgomery. Often, guitarists work in tandem with pianists, and their style when there is a pianist in the group may differ from how they play when they are the only chordal accompanists. For instance, some guitarists play only short chords on every beat if there is a pianist providing most of the rhythmic interest. Others will lay out (stop playing) entirely. For this reason, it is especially important to listen to guitarists in several different types of settings.
You should also listen to recordings that do not have any chordal accompaniment, such as any of several Gerry Mulligan, Chet Baker, or even Ornette Coleman quartet albums. Try to play along with these. This will often be difficult, since the music was recorded with the knowledge that there was no chordal accompaniment, so the soloist and other accompanists generally left little room for a piano or guitar. Practicing accompanying in this type of situation can help you avoid over-playing. Most beginning accompanists, like many beginning soloists, tend to play too much. Just as space can be an effective tool while soloing, it can be even more so when accompanying. Let the soloist work with only the bassist and drummer for a few measures, or longer, every so often. Laying out and leaving the soloist with no chordal accompaniment is sometimes called strolling. McCoy Tyner, Herbie Hancock, and Thelonious Monk often laid out for entire solos.
Sometimes it helps to imagine yourself as a background part in a big band arrangement. When you are comfortable with a particular chord progression, and no longer are having to concentrate fully just on playing the “right” notes, you can concentrate on the rhythmic and even melodic content of your comping. Listen to the horn backings in some big band recordings, such as those of Count Basie, to see how melodic accompaniment can be.
Certain styles of music call for particular rhythmic patterns. For instance, many forms of music before the bebop era used the stride left hand pattern, which consists of alternating a bass note on one and three with a chord voicing on two and four. Many rock based styles also depend on rhythmic patterns, often specific to the individual song. While the Brazilian derived styles such as the bossa nova and samba, as played by most jazz musicians, do not have well-defined comping patterns, other Latin jazz styles, particularly the Afro-Cuban forms sometimes collectively referred to as salsa, use a two measure repeating motif called a montuno. A typical rhythmic pattern is “and-of-one, and-of-two, and-of-three, and-of-four; one, two, and-of-two, and-of-three, and-of-four”. These two measures may be reversed if the underlying drum pattern (see below) is reversed as well. A full description of the role of the piano in Latin jazz and other styles is beyond the scope of this primer. A good discussion can be found in Mark Levine’s The Jazz Piano Book.
The most important aspect of accompanying in most styles is to communicate with the soloist. There are several forms this communication can take. For instance, there is call and response, in which you essentially try to echo back or answer what the soloist has played. This is particularly effective if the soloist seems to be playing short, simple phrases, with pauses between them. If the soloist is working on a repeated rhythmic motif, you can often anticipate the echo and actually play right along with the soloist. Sometimes you can also lead the soloist in directions he might not have tried otherwise. For instance, you might start a repeated rhythmic motif, which might encourage the soloist to echo you. Some soloists like this type of aggressive comping, and others do not. You will have to work out with each soloist how far you may take him.