Most learners can read a sentence they cannot hear. Chinese compresses: neutral tones flatten syllables, finals merge, and native speed leaves no room to translate. The fix is not more exposure - it is the right exposure.
Why listening is the bottleneck
Reading lets you pause; listening does not. Tone perception has to be automatic before meaning arrives, and unstressed syllables (妈妈's second ma, 了, 的) blur past untrained ears entirely. If tones are still effortful, fix that first - the minimal-pair drills target exactly this.
The drill ladder
- Dictation at your level. Play a graded story line (the library is leveled), write what you hear, check against the text. The misses show which contrasts your ear skips.
- Shadowing. Play a sentence, pause, repeat aloud copying the contour - not the words, the melody. Prosody learned physically transfers to hearing.
- Graded, then native. Stay where you catch ~90%. Below that you are practicing guessing, not hearing.
Bridge to real speech
Learner podcasts bridge the gap between drills and native audio - the podcast guide sorts them by level. For drama/film, watch with Chinese subs (not English): the subtitle is scaffolding that maps sound to characters in real time.
The part nobody tells you
Listening improves log-scale: weeks of invisible progress, then a sudden jump where a show becomes intelligible. Learners quit in the flat part. The measurable metric is not "did I understand" but "did I get my hour today" - keep that for three months and the jump arrives.
Common questions
Why is Chinese listening so hard?
Three things stack: tone must be decoded before meaning, unstressed syllables nearly vanish, and native speech runs words together. All three yield to targeted drills - they are trainable, not talent.
How do I practice Chinese listening alone?
Dictation of leveled audio plus shadowing. Play one line, transcribe or repeat it, check the text. Graded stories with audio are the ideal material.
How long until I understand spoken Chinese?
With an hour of focused listening daily: learner-speed audio in weeks, slow native content in months, natural conversation in roughly a year or two.