Compare tested groups, not a universal percentage
The chart compares an on-site test at two music conservatories. For students who began musical training at ages 4–5, roughly 60% in Beijing and 14% at Eastman met the strict criterion. These are approximate readings from the published figure, not exact case counts.
The groups were specialized: Mandarin-speaking first-year students at the Central Conservatory of Music in Beijing, and first-year Eastman students whose own and parental language backgrounds were nontonal. Neither group represents the population of its country.
What was tested?
Deutsch and colleagues’ 2006 study played 36 piano tones across three octaves and asked students to write down the note names. Success meant at least 85% correct. Successive notes were more than an octave apart to reduce the usefulness of relative-pitch cues.
The second scoring rule allowed an answer one semitone away. Changing the scoring rule changes some percentages without changing who took the test. The selector lets you compare those two definitions directly.
Why training age matters to the comparison
The figure separates students by the age at which musical training began. The full groups comprised 88 Beijing students and 115 Eastman students, but the displayed age groups include only subgroups of at least nine. The sample beside each result is that subgroup’s size, not the school’s total.
The paper found higher success in the earlier-training groups. This comparison does not isolate a single cause: recruitment, language, education and other differences cannot be removed by merely comparing two bars.
What does a zero on the chart mean?
No student in that displayed subgroup met the criterion. A small sample with no observed cases cannot establish that the ability is impossible in a broader group. It also cannot establish a worldwide frequency such as “one in ten thousand”.
Deutsch’s university research overview discusses the distinction between naming an isolated pitch and recognizing relationships between notes. This explorer compares the published evidence; it does not administer a pitch test. A related perception question has its own evidence on the synesthesia page.
Good questions
How rare is perfect pitch among music students?
It depends on the group. In a 2006 study, about 60% of Beijing students and 14% of Eastman students who started music training at ages 4–5 met the strict criterion. These are approximate readings from Figure 1.
What did the study count as perfect pitch?
Students named 36 notes. The criterion was at least 85% correct. The researchers also reported a second scoring rule that allowed errors of one semitone.
Do these percentages describe China and the United States?
No. They describe selected groups of first-year conservatory students, with different language backgrounds. They are not national population estimates.