Is This Note Going Up or Down?

weirdly accurate.

Is This Note Going Up or Down?

Is this note going up or down?

It depends on who is listening — that is the tritone paradox. Each pair is two notes exactly half an octave apart, built so they have no clear octave, so up and down are the same distance. Yet Diana Deutsch found in 1986 that most listeners hear a definite direction: 6 of the 8 trained musicians she tested heard some notes as consistently higher than others, and different listeners heard the same pair in opposite directions. Which way you hear it seems to be learned from speech: in Deutsch’s 1991 study most listeners raised in California and most raised in southern England heard the pairs in opposite ways, and her 1996 report found children hearing it much like their mothers. So your answer works like an accent for your ears: compare it with a friend or a parent and you may find you hear the very same file in opposite directions.

This tritone paradox test runs a short version of Deutsch’s experiment. The tones are Shepard tones built as in her 1991 paper: six partials an octave apart under a fixed bell-shaped loudness envelope, 500 ms each. After two warm-up pairs of ordinary notes, you hear two rounds of six pairs, twelve in all. Round two asks the same six questions with each pair played the other way round and under a second envelope half an octave away, the balance Deutsch got from several envelope positions. Each pitch class scores by how often you heard it as the higher note, and Deutsch’s bisection method finds the top of your pitch circle, shown as a time on a clock face.

There is no right answer, and this is not a hearing test or a music test: no musical knowledge is needed. Twelve answers give a rough reading, where each listener in Deutsch’s 1991 study gave 96 over two sessions. Phone speakers drop the lowest partials and change the colour of the sound, so headphones bring you closer to the lab.

What the tritone paradox test measures

One thing: which side of the pitch circle sounds higher to you, when the notes themselves give no answer.

The task

  • Pairs of Shepard tones (Shepard, 1964) a tritone, or half an octave, apart: does the pair go up or down?
  • Tones built as in Deutsch (1991): six octave-spaced partials under a cosine bell, 500 ms each, the second note straight after the first
  • Two warm-up pairs of ordinary notes with a plain answer, then two rounds of six pairs
  • Round two flips the order of each pair and moves the loudness envelope half an octave (C5 and F#4), so every question is heard both ways under both envelopes
  • No two pairs in a row share a note, as in Deutsch’s design; replay any pair before you answer

Your result

  • The top of your pitch circle as a time on a clock face, and how clear that top is
  • How many of the six questions you answered the same way in both rounds
  • Where groups of listeners had their tops: California and southern England (Deutsch, 1991) and Vietnamese first-language listeners (Deutsch, Henthorn & Dolson, 2004) — reference points, not a ranking
  • Why it happens, with the research on language, speaking voice and family
  • A shareable card and a link that lets a friend hear the same twelve pairs and see how many you hear the opposite way

Before you start

  • Sound on; headphones are best, because phone speakers lose the low half of every note
  • About a minute
  • Go with your first impression: you can answer as soon as the second note starts
  • Your answers are scored in your browser; a share link carries only twelve yes/no marks, one per hour of the clock, and the name you add

Tritone paradox FAQ

Is this test free? Do I need to sign up to see my result?

Yes, it’s free, and there’s nothing to sign up for: no account, no email, no payment. Your result appears on this page the moment you finish, and you can share it or play again straight away.

What is the tritone paradox?

The tritone paradox is an auditory illusion described by Diana Deutsch in 1986. Two Shepard tones half an octave apart are played one after the other; physically the step is the same distance up or down, yet most people hear a clear direction, and different people hear the same pair going opposite ways.

Is there a right answer to the tritone paradox?

No. Each tritone pair is exactly as far up as it is down, so neither answer is correct. What the tritone paradox test measures is consistency: which notes sound higher to you, and whether they sound higher again when the order and the colour of the sound change.

How does the test find the top of my pitch circle?

The tritone paradox test follows Deutsch’s scoring. For each of the twelve pitch classes it counts how often you heard that note as the higher one, then cuts the circle into the two halves of six that differ most. The top is the middle of the higher half, shown as a time on a clock face with C at twelve.

Why do people hear the tritone paradox differently?

The evidence points to speech heard early in life. In Deutsch’s 1991 study, 21 of 24 listeners who grew up in California had their top in one half of the circle, against 3 of 12 who grew up in southern England. In 2004, Vietnamese listeners heard it differently from Californians, even those who came to the US as children. A 1996 summary of 15 Californian listeners, mostly children, found each one’s top about a semitone from their own mother’s.

Do I need headphones for the tritone paradox?

Headphones are strongly recommended. Each Shepard tone spreads over six octaves, and phone speakers drop the lowest partials, which changes the colour of the sound. Repp (1997) found that changing the spectral envelope moves some listeners’ answers, so a closer match to the lab sound gives a cleaner reading.

Why did my answers change in the second round?

Round two moves the loudness envelope half an octave and reverses each pair. If most of your answers flipped with it, you were following the colour of the sound rather than the note, as some listeners in Repp’s 1997 study did. Twelve answers are also a short sample: Deutsch’s 1991 listeners each gave 96.

Is the tritone paradox test a hearing or music test?

No. The tritone paradox test has no right answer and does not check hearing or musical ability, and no note names are needed. Deutsch’s results point to a hidden absolute sense of pitch: the same notes sound higher every time, even to people who cannot name a single note.

Where the numbers come from

There is no population percentile for the tritone paradox and no better way to hear it, so the result shows where groups of listeners in Deutsch’s studies had their tops, as reference points rather than a ranking. The 1996 mothers-and-children figures come from a lay-language summary, not the full paper. The site does not publish its own players’ data.

How the top of your circle is found

Shepard tone

A note made of partials an octave apart under a fixed loudness envelope. Its pitch class is clear but its octave is not, so two such notes close on the circle are heard taking the short way round (Shepard, 1964).

Tritone pair

Two Shepard tones exactly six semitones apart, the farthest two notes can be on the pitch circle. Neither direction is shorter, so whatever direction you hear comes from you.

Bisection

As in Deutsch, Kuyper and Fisher (1987), described in Deutsch (1991): each pitch class gets the share of trials in which it was heard as higher, the circle is cut into the two halves of six that differ most, and the peak sits in the middle of the higher half.