Audio Memory Test

The sequence memory test lights each square as it plays. Watch that and the sequence arrives as a list of places, which is why it measures what your eyes kept hold of.

Light a pad here and exactly the same thing would happen. The melody would become a row of flashes, you would remember the flashes, and the notes would be decoration on top of a visual task.

So the pads stay dark. A note sounds, and the only way to know which pad it came from is to hear how high it was. That conversion is part of the job, and it is the part that makes this an audio number rather than a visual one.

It also makes the test unusually forgiving about equipment. Each pitch is computed rather than recorded, so a cheap speaker plays it at exactly the right frequency. Only the volume changes, and volume is not what you are being asked about.

Read someone seven digits and most people can say them straight back. Play seven notes and almost nobody can. The gap is large, and musical training closes very little of it.

A digit has a name. You say it under your breath, and the sound of your own voice keeps it alive for a few seconds. That rehearsal loop is most of what a digit span measures, and how working memory works walks through the machinery. The figures themselves sit on what counts as a good digit span.

A note has no name unless you gave it one, and very few people did. The trace has to survive as a pitch rather than as a word, and pitch traces fade faster and get in each other's way.

Deutsch showed that interference directly in 1970. Fill the gap after a tone with spoken numbers and the tone survives. Fill it with other tones and it does not. Two stores, doing two jobs, sitting side by side.

So a tone span of five is not a poor digit span. It is a different measurement, and there is no exchange rate between the two.

The card reports the average position of your mistakes, and it is often the more interesting of the two figures.

Slips cluster in the middle. The first note is protected because nothing was competing with it when it arrived. The last note is protected because it is still ringing when your turn starts. The notes between have neither advantage.

That shape turns up in almost every list task ever run. Finding it in your own three or four slips proves nothing. Finding it across a dozen runs is a different matter.

Slips at note one mean something else is wrong. Either the melody starts before you have settled, or the pace is quicker than you can follow. Slow it down and take it again.

Four pads laid out low to high hand you a rough naming scheme. Top, upper middle, lower middle, bottom. That is nothing like absolute pitch, and it is a long way from nothing.

It is why the reference curve here sits above the plain capacity figure for material you cannot name. Cowan puts that near four items. Miller's famous seven belongs to material you can put words to. A tone ladder sits between the two, and so does the curve.

Close the pads up and the shortcut stops working. The narrow setting squeezes the whole set inside a fifth. Telling the third pad from the fourth becomes a pitch judgement rather than a glance at a shape. Most people lose a note or two.

Widen the pads again and add more of them instead. Six notes across an octave and a half are easy to tell apart and hard to keep in order. The two settings break the task in different places, which is why both are here.

• Fewer pads, wide apart: the easiest setting, and the one the reference curve is drawn for.
• More pads, wide apart: the notes stay easy to name and the order gets crowded.
• Any number of pads, close together: naming fails first, and the order barely matters.

A span is a snapshot. It moves with sleep, with the noise in the room, and with how many melodies you have already sat through today. One run is a data point rather than a verdict.

It is not a musical ability score either. Plenty of trained musicians land in the middle here, and plenty of people who have never played anything do well. Getting better at this page means getting better at this page.

Nothing here detects or rules out any condition. There is no hearing problem and no memory problem hiding in a number between two and twelve. If your memory has changed in a way that worries you, that belongs with a doctor.