Twelve-Tone Equal Temperament (12-TET)

12-TET is an engineering compromise: divide the octave into 12 exactly equal steps, each with a frequency ratio of 2^(1/12) ≈ 1.05946. No interval except the octave is acoustically pure. Every interval is slightly “wrong” — but equally wrong everywhere. This uniformity is the point: it buys unlimited transposition at the cost of tuning purity.

🎯 Simple version: 12-TET is a deal: you make every note slightly imperfect so that you can play in any key equally well. It’s like rounding every fraction to a decimal — close enough, and much easier to calculate with.

The Problem 12-TET Solves

The fundamental mathematical impossibility underlying all tuning systems:

You cannot simultaneously have pure octaves (2:1) and pure 7-step-intervals (3:2) in a closed cycle.

The proof, expressed in step arithmetic:

Stack twelve 7-step-intervals: 12 × 7 = 84 chromatic steps = 7 octaves. Stack seven octaves (the 12-step-interval): 7 × 12 = 84 chromatic steps.

Both paths traverse 84 steps — they should land on the same pitch. But the frequency math disagrees:

Twelve 3:2 ratios:  (3/2)^12 = 129.7463...
Seven octaves:      2^7      = 128.0000

The difference — the Pythagorean comma — is:

(3/2)^12 / 2^7 = 531441/524288 ≈ 1.01364

In cents: approximately 23.46 cents — roughly a quarter of a chromatic step.

This means: if you tune 12 pure 7-step-intervals in a row (84 chromatic steps = 7 octaves), you overshoot the target octave by 23.46 cents. The cycle doesn’t close. The integer arithmetic says 12 × 7 = 7 × 12, but the frequency ratios say (3/2)^12 ≠ 2^7. You can’t have it all.

Every tuning system in history is a different strategy for distributing this inevitable error. 12-TET distributes it equally across all 12 steps.

The 12-TET Solution

12-TET defines every chromatic step as exactly:

Ratio per step = 2^(1/12) ≈ 1.059463
Cents per step = exactly 100

This means:

• Every the 7-step-interval (the “fifth”) = 700 cents, vs. just 702.0 cents → error of -2.0 cents
• Every the 5-step-interval (the “fourth”) = 500 cents, vs. just 498.0 cents → error of +2.0 cents
• Every the 4-step-interval (the “major third”) = 400 cents, vs. just 386.3 cents → error of +13.7 cents
• Every the 3-step-interval (the “minor third”) = 300 cents, vs. just 315.6 cents → error of -15.6 cents

The fifths and fourths are excellent approximations (2 cents off — barely perceptible). The thirds are noticeably compromised (14-16 cents off — clearly audible in sustained chords to trained listeners). This is the central trade-off.

See the complete comparison in the Intervals table.

What 12-TET Sacrifices

1. Pure thirds. The 5:4 ratio (386.3 cents) is approximated by 400 cents — 13.7 cents sharp. In a sustained {0,4,7} chord on a piano, the step-4 interval beats at a rate determined by the 13.7-cent mistuning. This is why barbershop quartets and a cappella groups naturally drift toward just intonation on sustained chords — the singers’ ears pull them toward the pure ratios.

2. Harmonic series alignment. The 7th harmonic (ratio 7:4 = 968.8 cents) falls 31.2 cents below the 12-TET step-10 (1000 cents). The 11th harmonic is 48.7 cents off the nearest step. 12-TET is optimized for 3-limit and 5-limit intervals; higher-prime harmonics are poorly served.

3. Timbral variety between keys. In historical temperaments (meantone, well-temperament), different keys had different interval sizes, giving each key a unique character. In 12-TET, all keys sound identical — a feature for transposition, but a loss of color.

What 12-TET Gains

1. Universal transposition. Any melody, chord, or progression can be shifted to any starting pitch without changing its internal interval structure. This enabled the explosion of modulation and key-change techniques from Bach onward.

2. Instrument compatibility. All 12-TET instruments are compatible with each other in any key. An orchestra, a rock band, and a jazz combo can all play together without retuning.

3. Simplified arithmetic. Transposition = addition mod 12. Interval = subtraction. No lookup tables or special cases. This is the foundation of the PhizMusic naming system.

Historical Context

Equal temperament was not a Western invention. It was independently discovered by:

• Zhu Zaiyu (朱載堉), Chinese prince and scholar, who published the precise mathematical formulation in 1584, computing 2^(1/12) using an 81-column abacus
• Simon Stevin, Flemish mathematician, who independently derived the same result around 1585 (though his calculations contained small errors)

Both arrived at the same solution because the mathematical problem is universal — it doesn’t depend on cultural musical preferences. The adoption of 12-TET in Europe was gradual, becoming standard only in the 19th-20th centuries. Many cultures continue to use non-equal temperaments that optimize different priorities.

The Railsback Curve

A revealing real-world complication: piano tuners do not actually tune pianos to 12-TET.

The Railsback curve, documented by O.L. Railsback in 1938, shows that professional piano tuners systematically stretch octaves — tuning the upper register slightly sharp and the lower register slightly flat relative to mathematical 12-TET.

Why? Piano strings are thick, stiff wires, not ideal mathematical strings. Their stiffness causes inharmonicity: the overtones are slightly sharper than perfect integer multiples of the fundamental. The 2nd partial of a bass string might be at 2.003× the fundamental rather than exactly 2×.

To make octaves sound pure on a real piano, tuners match the 2nd partial of the lower note to the fundamental of the upper note — which means tuning the upper note slightly sharp of the mathematical 2:1 ratio. This stretching accumulates across the keyboard.

The Railsback curve is a beautiful demonstration of Sethares’ timbre-tuning coupling (see timbre.md): you don’t tune to abstract mathematical ratios — you tune to the actual spectrum of the instrument. The “correct” tuning depends on the instrument’s physical properties, not just number theory.

The 14-Cent Question

Is the ~14-cent error on thirds audible? It depends on context:

• Melodic intervals (notes played sequentially): most listeners don’t notice — the brain processes sequential pitches categorically
• Sustained harmonic intervals (notes held together): trained listeners hear beating, especially on thirds
• Fast passages: errors are essentially inaudible
• Choral/ensemble music: singers and string players naturally adjust toward just intonation
• Piano: the attack transient and rapid decay mask the beating

The perceptual threshold for interval mistuning is roughly 5-10 cents for trained musicians in sustained harmonic contexts. The 2-cent fifth error is essentially imperceptible. The 14-cent third error is above this threshold — barely, and context-dependently.

This is why 12-TET won: the errors are just small enough to be tolerable in most musical contexts, while the transposition freedom is indispensable for modern music’s harmonic complexity.

Hear the Difference

Compare just intonation intervals (pure ratios) with their 12-TET approximations. Root = 220.00 Hz (La3 / octave 3, step 9). The difference is most audible on sustained step-intervals 3 and 4.

Translation Table

Connections

• Frequency Ratios — just intonation ratios that 12-TET approximates
• Intervals — complete table showing just vs. 12-TET for every interval
• Scales — all scale subsets are drawn from the 12-TET grid
• Harmonic Series — the natural ratios that 12-TET approximates
• Consonance & Dissonance — how temperament errors affect perceptual fusion
• Instrument Physics — string inharmonicity and the Railsback curve
• Timbre — Sethares’ timbre-tuning coupling

Suggested References

• Sethares Dissonance Curves (aatishb.com) — Interactive demonstration of how 12-TET intervals approximate just ratios, with audible comparison