Weeks 5-6 · ear · audiation across delay
Hear a phrase. Hold it. Sing it back five seconds later.
The rule
pass: each note ≤ ±50¢ · ≥75% of notes · 6/8
The grey wedge on the left is the 5-second silence between hearing the phrase and singing it. Each dot to the right is one note of the recalled phrase: pitch error vs. target. Four land inside the emerald ±50¢ band — pass. The third sits ~80¢ flat: the classic delay-drift failure mode, where the singer loses the tonic during the hold and the phrase slides off-key.
Hear it first
First clip: a 5-note phrase, then five seconds of silence, then a recall that lands every note. Second clip: the same phrase recalled, but the third note slides flat — the tonic anchor wasn’t held during the silence. Third clip: contour right, starting note wrong by a semitone (transposition error — separate fail mode).
What’s happening. Audiation — holding a melody silently in your head — is a separate skill from immediate echo. Working memory for melodic contour decays fast: most people lose the absolute pitch of a 4-note phrase within 3–5 seconds, even if they keep the shape. The drill trains pitch retention across delay, which is the engine of singing without a piano cue and the foundation of recognising songs in your head before performing them.
Why a 5-second delay. Two seconds is just immediate echo; ten seconds is a memorisation task. Five seconds sits in the working-memory window where most untrained singers start to lose the tonic — long enough to test recall, short enough that the phrase shape stays intact. The ±50¢ tolerance is the same band as single-note-match, so we’re isolating the “held it through silence” variable.
Why this is the foundation for song. song-from-memory downstream asks you to recall a full 8-bar verse after one playback. That’s impossible if 4-5 notes through 5 s of silence is still wobbly. two-phrase-memory and a-cappella-resolve-tonic both depend on this skill being settled.
Session ahead
8 phrases · 4-5 notes each · 5-second hold · 6/8 recalls promotes.