The human auditory system can separate simultaneous sounds into distinct perceptual streams — different instruments, different voices. Training this ability is essential for harmonic dictation, orchestral listening, and chamber music performance.
AUDITORY SCENE ANALYSIS
Albert S. Bregman, a psychologist at McGill University, coined the term auditory scene analysis to describe the process by which the auditory system decomposes a complex acoustic signal into its component sources. His book Auditory Scene Analysis: The Perceptual Organization of Sound (MIT Press, 1990) is the foundational text on this topic. Bregman showed that the brain uses multiple cues — frequency proximity, common onset, harmonic relationship, spatial location, and temporal continuity — to group acoustic components into perceptual streams representing different sources. When you hear a string quartet, your auditory system automatically separates the first violin, second violin, viola, and cello into four distinct streams, even though the acoustic signal reaching your ears is a single complex mixture.
Musicians with training have enhanced auditory stream segregation abilities. Diana Deutsch at UC San Diego has documented what she calls the scale illusion: when two interleaved scales are presented simultaneously (one to each ear), most listeners group the pitches by register rather than by ear, perceiving a different pattern entirely. The ability to hear polyphonically — to follow multiple independent voices in a complex texture — is trainable and is one of the primary skills developed through ensemble playing and score-reading.
THE SOPRANO-BASS FRAMEWORK
In four-voice harmonic dictation (the standard training format used at most music schools), the standard approach is to focus first on the outer voices — soprano (top) and bass (bottom) — before adding the inner voices (alto and tenor). The reason is functional: in tonal music, the soprano and bass carry most of the harmonic information. The bass defines the chord root and inversion; the soprano defines the melodic line and the voice-leading gesture. The inner voices (alto and tenor) typically fill out the harmony and move smoothly between chord tones, following strict voice-leading principles that make them somewhat predictable once the outer framework is known.
For dictation purposes: write the bass first (it moves most clearly and defines harmonic roots), then write the soprano (the most prominent melodic voice), then infer the inner voices from the chord quality and voice-leading conventions.
FOLLOWING A FUGUE VOICE BY EAR
A fugue is perhaps the most demanding polyphonic listening exercise in Western music. In a Bach three-voice fugue, three independent melodic lines proceed simultaneously, each with its own melodic contour, rhythmic profile, and harmonic role. The ability to follow a single voice through a fugue — tracking it as it enters, disappears, reappears in a new register, is imitated by another voice, and finally arrives at a cadence — requires precisely the auditory stream segregation skills Bregman describes.
A practical exercise: choose a Bach two-voice invention (simpler than a fugue but using the same principles). Listen to it repeatedly. On the first listening, follow only the right hand (soprano). On the second, follow only the left hand (bass). On the third, try to hold both in your attention simultaneously. Then sing one voice while listening to the other. This contrapuntal listening practice — more than almost any other exercise — develops genuine musical understanding of polyphonic texture.
HARMONIC DICTATION FROM INNER VOICES
Advanced harmonic dictation requires not just identifying chord roots and qualities but tracking voice leading across multiple voices. Which notes resolve? Which notes leap? Which notes are doubled? In SATB (soprano-alto-tenor-bass) dictation, the standard exercise at Year 2-3 level, students must notate all four voices correctly. The voice-leading rules studied in harmony class (avoid parallel fifths, resolve the leading tone, double the root) serve as constraints that make this task more predictable — and violations of these rules appear in dictation exercises precisely so students learn to identify and correct them.
VOICE-CROSSING AND VOICE EXCHANGE
In skilled polyphonic writing, voices sometimes cross each other — the alto rises above the soprano momentarily, or the tenor drops below the bass. Voice crossing creates moments where the auditory stream segregation requires extra effort: the listener may need to follow a voice by timbre or register continuity rather than by absolute pitch position. Bach uses voice crossing strategically in his fugues for textural variety.
Voice exchange is a related phenomenon where two voices swap notes: one voice moves from A to B while another voice simultaneously moves from B to A. The total harmonic content stays the same, but the individual voice motion creates an internal restlessness. Voice exchange is a basic technique of Schenkerian analysis and appears throughout tonal music.
IMITATION AND CANON
When one voice presents a melodic idea and another voice presents the same idea slightly later, the result is imitation. If the imitation is exact (same intervals, same rhythm) and continues throughout, the texture becomes a canon. Pachelbel's Canon in D is the most famous example, with three violin parts entering in succession on the same melody. Bach's Goldberg Variations include nine strict canons at various intervals.
Recognizing imitation by ear requires holding the first voice's melodic shape in memory and identifying when a second voice presents the same shape. This is a fundamental skill for analysing Renaissance and Baroque polyphony.
THE FUGAL TEXTURE
A fugue is the most complex polyphonic structure in Western music. The texture consists of three to five independent voices, each presenting a distinctive subject and various countersubjects, modulating through different keys, and recombining in different relationships throughout the piece. Bach's Well-Tempered Clavier (two books of 24 preludes and fugues each, totaling 48 fugues) is the standard repertoire for studying fugue.
For ear training, the goal of fugue listening is not to track every voice every moment but to follow at least one voice from beginning to end while remaining aware of the others. Choose one of Bach's three-voice fugues from WTC Book I and listen 5-10 times, each time following a different voice. After this process, you will hear the fugue as three independent threads rather than as undifferentiated polyphony.
ORCHESTRAL LISTENING SKILLS
Following voices in an orchestra is a related but distinct skill. An orchestra has dozens of instruments grouped by family (strings, winds, brass, percussion). Each instrument family has a characteristic timbre, and the timbral differences make polyphonic listening easier than in a homogeneous texture (like a string quartet, where all four instruments share the same timbre). For orchestral listening, focus first on instrumental color: where are the violins, the woodwinds, the brass? Then track how a melodic line moves between instruments — orchestrators often pass a melody from one instrument to another, requiring you to follow the line across timbral changes.
Rimsky-Korsakov's Scheherazade (1888) is a standard introduction to orchestral listening: clear melodies, distinctive instrumentation, and identifiable thematic transformations make it accessible for developing the skill.
PRACTICAL EXERCISES
Exercise 1: Choose a Bach two-part invention. Sing one voice while a recording plays. Then sing the other voice while a recording plays. Then alternate between voices in your singing. Exercise 2: Listen to a string quartet (Haydn's Op. 76 No. 3, the "Emperor," is ideal for beginners). On successive listenings, follow violin 1, then violin 2, then viola, then cello. Notice when each voice has the melody and when each plays accompaniment. Exercise 3: Use score-following with a recording. Choose any piece of polyphonic music with score available on IMSLP. Read the score while listening, identifying each voice's part. This trains the audio-visual correspondence that underlies all formal music analysis.
SOURCES & FURTHER READING
Bregman, Albert S. Auditory Scene Analysis. MIT Press, 1990. — The scientific foundation of polyphonic hearing. Karpinski. Aural Skills Acquisition. Oxford University Press. Chapter 9: Four-Voice Dictation. Aldwell and Schachter. Harmony and Voice Leading. Schirmer/Cengage. Chapters on voice leading. Open Music Theory — Voice Leading (openmusictheory.net, CC BY-SA 4.0). teoria.com — Chord Identification in Context.