Related
One recognition is all that matters:
What you remember is not frequency but relation.
Most people believe they "remembered the melody". What they actually remembered is the distance and direction between notes — which is why a melody moved to another key is still recognisable (see the side of scales that is independent of key).
It also explains why ear training starts from relations, not from note names: trying to name pitches from the first note usually breaks down by the second.
Three layers, from coarse to fine
| Layer | What you hear | What you ignore | Where most people already are |
|---|---|---|---|
| 1 contour | overall direction (up, down, arched) | exact pitches and intervals | usually fine |
| 2 landing notes | the last note of each phrase | the notes in between | the most worthwhile |
| 3 intervals | each step: leap or step, and how wide | — | needs dedicated training (see [[concept:interval-ear-training |
Layer 2 is the most undervalued. A phrase's landing note determines the phrasing and the modal identity (see phrase and period), and catching the landing notes gives you the whole skeleton — the inner notes can come later.
How to practise
Step 1: hear only the contour, draw it with your hand. On the first listen, ignore pitch and trace the rise and fall with a gesture. If you can draw the shape afterwards, you have passed.
Step 2: note only the landings. Sing the final note of each phrase and record its relative position (higher or lower than the phrase's start, and by how much). Far easier than remembering everything, and far more informative.
Step 3: fill in the middle. With the skeleton known, hear the inner notes as the path from landing to landing — asking "was that a step or a leap", not "what note was that".
Step 4: write it down to check. The test of ear training is whether you can write it out. Being unable to write it means you have not heard it, which is not the same as "it felt clear".
A common frustration and its cause
| Symptom | Real cause |
|---|---|
| forgot the first half by the second listen | trying to store every note, which overloads memory |
| can hum it but cannot name intervals | layer 3 not yet trained |
| cannot recognise it in another key | storing absolute pitches instead of relations |
| cannot follow a long melody at all | not dividing it into phrases, hearing it in one gulp |
Rows one and four are the same problem: a melody must be heard phrase by phrase, because the capacity of auditory memory aligns naturally with phrases — the aural counterpart of breathing and phrasing.
Diagram: peeling off the three layers
Listen: start from the contour
Use scale to hear a rising figure, then a falling one — that is layer 1. Hear only the direction, without naming pitches, until you can state "rising" or "falling" instantly.
Common misconceptions
- "Hearing a melody means hearing note names." You hear relations (distance and direction), which is why the
- "Not remembering a melody means bad ears." Usually the method is wrong — trying to store every note exceeds
- "If you can hum it, you have heard it." Humming shows the contour is caught; whether you can write it out
- "Ear training needs dedicated sessions." All three layers can be practised while listening normally: first
same melody is recognisable in another key.
auditory memory.
is the real test.
only the direction, then only the phrase endings, and finally the intervals. :::
Listen in the library
All tracks from midicn-lib — inline badge is the license tier (C1 commercial / C2 non-commercial / C3 study only).
Sources
- 旋律听辨的通行训练顺序（轮廓 → 落音 → 音程 → 记谱）为视唱练耳教学通则
- 听觉记忆以音高"关系"为主而非绝对频率，为音乐心理学通行研究结论
Used for fact-checking only; all prose is original.