Transposition and Relative Hearing: Why Absolute Pitch Is Not the Point
Recognising a transposed tune is free. Producing one is not, and that gap is where the training goes.
Short answer: transposition means moving a passage so that every note shifts by the same interval, and the fact that it still sounds like the same tune is the single most important fact about human pitch perception. Melodies are stored mostly as patterns of distance, not as lists of absolute frequencies, which is why a song is recognisable when a different singer starts it a fourth higher. Absolute pitch, the ability to name an isolated note without a reference, is rare and is not what musicians use to play in tune together. Relative hearing is. That is the good news for anyone who has decided they missed the window.
What transposition is, and where you meet it
Grove's definition is the plain one: shifting a piece or passage to a different pitch level while preserving the intervals between the notes. In practice it arrives in four everyday forms.
| Situation | What moves | What the ear has to do |
|---|---|---|
| A singer takes a song down a third | Every note, by the same interval | Nothing. Recognition is automatic. |
| A guitarist uses a capo | The whole instrument | Nothing, unless you are matching a recording. |
| A clarinet in B flat reads a written C | Notation against sounding pitch | Hold two pitch names for one sound. |
| An accompanist transposes at sight | Notation, in real time | Read function rather than letter. |
The transposing-instrument case is the only one that is genuinely hard, and it is hard for a notational reason, not an auditory one. A B flat instrument sounds a major second below what is written, an F horn a perfect fifth below, an E flat alto saxophone a major sixth below. Nothing about the sound is unusual. The difficulty is holding two naming systems at once, which is exactly the problem that movable-do solfège was designed to dissolve.
The evidence that melody is stored relatively
Dowling proposed in 1978 that memory for melodies has two separable components: contour, the pattern of ups and downs, and the scale the melody is drawn from. Contour survives transposition intact, which is why an unfamiliar tune moved to a new key is still judged the same tune, and why listeners will accept a contour-preserving imitation that gets the exact intervals wrong.
Attneave and Olson had already made the point from a different direction in 1971. They asked listeners to transpose a familiar signal pattern into new pitch ranges and found that people reproduced it in terms of frequency ratios, consistent with pitch operating as a logarithmic medium in which a pattern can be slid up and down without deforming.
The developmental evidence closes the case. Trehub, Bull and Thorpe showed in 1984 that infants treat transposed melodies as equivalent and are sensitive to melodic contour. Relative hearing is not an advanced skill acquired through training. It is the default mode, present before any training is possible.
But absolute information is retained too
The relative story is often told too strongly, as though absolute pitch level were discarded. It is not. Levitin asked people with no absolute pitch training to sing familiar pop songs from memory and compared their productions against the recordings. As summarised in the later replication paper, 44 per cent of his participants sang the correct pitch on at least one of two trials and 12 per cent were correct on both. A six-laboratory European replication with 277 participants found the effect in the same direction but weaker: 25 per cent exact on at least one song, 4 per cent on both.
Schellenberg and Trehub found the same thing on the recognition side. Untrained listeners heard five-second excerpts of instrumental soundtracks from familiar television programmes, either at the original pitch level or shifted by one or two semitones, and correctly identified the originals. Listeners given comparable excerpts from unfamiliar recordings could not. Their conclusion is worth carrying around: adults' reportedly poor pitch memory is largely a by-product of being unable to name isolated pitches, not of failing to store the information.
So the real divide is naming, not hearing
Put those two bodies of evidence together and the usual framing collapses. It is not that some people hear absolutely and others relatively. Almost everyone hears relatively for structure and retains a surprising amount of absolute information for familiar material. What absolute-pitch possessors have is a fast, reliable label attached to that information.
This matters practically because labelling is not what music-making requires. Tuning a chord, matching a section, playing in a transposing ensemble, harmonising by ear, singing a line back: every one of those is a distance judgement. Our separate guide to relative pitch versus perfect pitch goes through the trade-offs, including the awkward cases where absolute labelling actively gets in the way, such as reading for a transposing instrument.
How to train the transposition half
Recognition across transposition is free. Production across transposition is not, and that gap is where most ear-training effort belongs.
- Reproduce, do not repeat. Hear an interval, then rebuild it from a different starting note. Repeating the same two notes trains nothing new.
- Move a short phrase deliberately. Take four notes, sing them from one starting pitch, then from a starting pitch a whole tone higher, checking the last note against a reference.
- Use degrees, not letters. A phrase remembered as 1, 3, 2, 5 transposes for free. The same phrase remembered as C, E, D, G does not.
- Check the drift. Transposed reproduction usually fails by accumulating a small error per interval. Sing the phrase twice through and compare the final note with the reference; the size of the gap tells you where you are.
What our games do here, honestly
This is the one page in this group where the Hearkle games line up directly with the topic. Hearspan plays a reference interval, gives you a new starting pitch and asks for the same distance again, which is transposition reduced to its smallest possible unit. Hearecho measures how long a reference survives, which is the resource transposed reproduction spends. Heardrift tests whether you can hold a target while a tone slides away from it.
What the games still cannot do is present a key, a chord or a harmonic context, because all eleven of them are monophonic. Transposing a melody inside a key is a bigger job than transposing an interval, and the tonal half of it needs a keyboard or a drone.
Transposition, reduced to one measurement. Hearspan gives you a reference interval, then a new starting pitch, and asks for the same distance. That is precisely the skill that lets a musician play in any key without naming a single note, and it is the one game we have that maps onto this page directly.
Frequently asked questions
What does transposition mean?
Moving a passage so that every note shifts by the same interval, leaving the distances between notes unchanged. The tune stays recognisable because listeners store melodies largely as patterns of distance rather than as lists of absolute frequencies. A singer taking a song down a third and a guitarist using a capo are both doing the same thing.
Why can I recognise a song in any key?
Because melody memory is built on contour and interval structure. Dowling (1978) separated the two components, contour and scale, and showed that contour survives transposition. Trehub, Bull and Thorpe (1984) found that even infants treat transposed melodies as equivalent, so relative hearing is the default rather than a trained skill.
Do people without absolute pitch remember pitch levels at all?
Yes, more than the folklore suggests. Levitin (1994) had untrained listeners sing familiar pop songs from memory; as reported in the later replication paper, 44 per cent produced the correct pitch on at least one of two trials. A six-lab replication with 277 participants found a weaker but consistent effect, 25 per cent exact on at least one song. Schellenberg and Trehub (2003) found the same for recognising the pitch level of familiar television themes.
Is relative pitch enough to be a good musician?
It is what the job actually uses. Tuning, blending, harmonising by ear, playing in a transposing ensemble and singing a line back are all distance judgements. Absolute pitch adds a fast label for isolated notes, which is useful for some tasks and an active nuisance for others, notably reading for a transposing instrument.
Why are transposing instruments confusing?
For a notational reason, not an auditory one. A B flat instrument sounds a major second below its written note, an F horn a perfect fifth below, an E flat alto saxophone a major sixth below. The sound is ordinary; the difficulty is holding two naming systems for the same pitch. Thinking in scale degrees rather than letter names removes most of the problem.
How do I practise transposing by ear?
Reproduce rather than repeat. Hear an interval or a short phrase, then rebuild it from a different starting pitch and check the final note against a reference. Remember phrases as scale degrees rather than letters, because degrees transpose for free. Errors usually accumulate a little per interval, so comparing the last note is the fastest diagnostic.
Keep reading
- Relative pitch vs perfect pitch: what each one is actually good for.
- Can adults learn perfect pitch?: what the training studies found about the naming half.
- Tone memory training: the resource that transposed reproduction spends.
- Interval recognition training: the evidence behind the recognition half of the skill.
Sources: Dowling (1978), Scale and contour: two components of a theory of memory for melodies, Psychological Review 85(4); Attneave and Olson (1971), Pitch as a medium: a new approach to psychophysical scaling, The American Journal of Psychology 84(2); Trehub, Bull and Thorpe (1984), Infants’ perception of melodies: the role of melodic contour, Child Development 55(3); Levitin (1994), Absolute memory for musical pitch: evidence from the production of learned melodies, Perception & Psychophysics 56(4); Frieler, Fischinger, Schlemmer, Lothwesen, Jakubowski and Müllensiefen (2013), Absolute memory for pitch: a comparative replication of Levitin’s 1994 study in six European labs, Musicae Scientiae 17(3); Schellenberg and Trehub (2003), Good pitch memory is widespread, Psychological Science 14(3); McDermott, Keebler, Micheyl and Oxenham (2010), Musical intervals and relative pitch: frequency resolution, not interval resolution, is special, Journal of the Acoustical Society of America 128(4); Transposition and Absolute pitch, Grove Music Online, Oxford University Press; Deutsch, Absolute pitch, University of California San Diego.
Understand it, now train it. The daily games stay free forever. A Personal Pass adds progress tracking (trends + export); Teacher / School tools let you assign it and follow a whole class. For teachers & self-learners →