How to understand spoken Chinese: listening practice for HSK 3.0

On this page

Why spoken Chinese is hard to follow
The listening part of HSK 3.0
Learn each word by its sound first
Hear tones in whole words
Let the context tell you the word
Practise with recordings you almost fully understand
Practise like the exam
How Hanzi By Heart helps with listening
Questions and answers
Sources

Short answer Spoken Chinese is hard to follow for three reasons: it is fast, many words sound the same, and the tone decides the meaning. Read aloud, Mandarin runs at about five syllables per second, and of the 5,400 words of HSK 3.0 levels 1–6, one in eight sounds exactly like another word on the list. In the exam, listening makes up 42 to 50 percent of the questions at every level. What helps most: learn every word by its sound, listen for tones in whole words, let the context tell similar words apart, and practise with recordings in which you know almost every word.

Why spoken Chinese is hard to follow

It is fast. In a study of seven languages, Mandarin speakers read short texts aloud at about 5.2 syllables per second. A two-syllable word like 知道 (zhīdào, to know) is over in less than half a second.
There are few syllables. All 5,400 words of levels 1–6 use only 375 different syllables if you ignore the tones, and 993 if you count each tone separately. The same sounds come back again and again.
Many words sound the same. 670 words on the list sound exactly like at least one other word, tones included. Within level 1 it is only 16 of 300 words; across levels 1–6 it is one word in eight.
Many words differ only in tone. We found 213 sets of two-character words with the same syllables that differ only in their tones, such as 老师 (lǎoshī, teacher) and 老是 (lǎoshì, always).
Reading does not train your ear. If you met a word only in writing, you may know it on the page and still miss it in a sentence. Reading and listening are separate skills, and each needs its own practice.

The listening part of HSK 3.0

The table follows the official structure and sample questions for levels 1–6, published by Chinese Testing International in December 2025. The last column shows the share of listening questions in the whole written exam.

Level	Questions	Minutes	What you hear	Share
1	20	≈ 12	short phrases and sentences, very short dialogues	50%
2	25	≈ 17	sentences and short dialogues	42%
3	30	≈ 23	short and longer dialogues, short talks	43%
4	32	≈ 20	dialogues of two to four lines, short talks	46%
5	35	≈ 25	longer dialogues and stories, several questions per recording	49%
6	40	≈ 30	reports, talks and an interview	49%

At levels 1 and 2, the sample paper prints pinyin above the answer options. From level 3 on, the options are in characters only, and from level 4 on, each question has four options. From level 5 on, several questions often belong to one longer recording, so you have to keep the content in mind while you listen.

Learn each word by its sound first

When you learn a new word, play the recording before you look at the meaning, and say the word aloud. Then test yourself the other way round from most flashcards: hear the word and recall what it means before you see the characters. A word you can only recognise in writing is not yet ready for the listening part.

Learn the sound of the whole word, not only of each character. In 你好 the first syllable changes from the third to the second tone, and in 不是 (búshì) 不 changes to the second tone. Our guide on pinyin tones explains these changes.

Hear tones in whole words

Practise tones in two-syllable words, because that is how you will hear them: 4,101 of the 5,400 words have two characters. Words that differ only in tone make good training:

Words	Pinyin	Meanings	Levels
老师 · 老是 · 老实	lǎoshī · lǎoshì · lǎoshi	teacher · always · honest	1 · 4 · 6
时间 · 实践 · 事件	shíjiān · shíjiàn · shìjiàn	time · practice · event	1 · 5 · 5
知道 · 直到 · 指导	zhīdào · zhídào · zhǐdǎo	to know · until · to guide	1 · 3 · 5
眼睛 · 眼镜	yǎnjing · yǎnjìng	eyes · glasses	2 · 4
教室 · 教师	jiàoshì · jiàoshī	classroom · teacher	2 · 4
手机 · 收集	shǒujī · shōují	mobile phone · to collect	1 · 5

The falling fourth tone is the most common one: it carries 36 percent of all syllables in the 5,400 words, followed by the second tone (22 percent), the first (21 percent) and the third (18 percent). If one tone gives you trouble, collect words with it and listen to them side by side.

Let the context tell you the word

Some words you cannot tell apart by sound at all. Then the words around them decide:

Sound	Word and meaning	You hear it in
gōngyuán	公园, park	去公园散步 (go for a walk in the park)
gōngyuán	公元, CE (the calendar era)	公元前221年 (221 BCE)
lǚxíng	旅行, to travel	去云南旅行 (travel to Yunnan)
lǚxíng	履行, to carry out	履行合同 (fulfil a contract)
yàoshi	要是, if	要是下雨，我们就不去了 (if it rains, we won't go)
yàoshi	钥匙, key	我的钥匙找不到了 (I can't find my keys)
jìnlái	进来, to come in	请进来 (please come in)
jìnlái	近来, recently	近来工作很忙 (work has been busy lately)

This is why it pays to learn words in short phrases and example sentences: you then hear a word together with the words it usually comes with.

Practise with recordings you almost fully understand

A study of English as a foreign language found that learners understood informal spoken stories well once they knew about 95 percent of the words. At 90 percent, some still understood a lot and others little. For Chinese, that gives a rule of thumb: if you miss more than about one word in twenty, choose an easier recording or learn its new words first.

Then work through each recording in steps:

Listen once without the text and note what you understood.
Listen again with the transcript and look up what you missed.
Listen a third time without the text. Now you should hear every word.
Repeat short sentences right after the speaker, at the same speed (shadowing).

Use short recordings of one to three minutes, so you can hear each one several times in a few minutes. Good sources are the recordings of your textbook, the official sample papers and mock tests, podcasts for learners and, later, TV series with Chinese subtitles.

Practise like the exam

Read the options first. Use the seconds before each recording to read the answer options. They tell you what to listen for: a time, a place, a person or an opinion.
Keep going when you miss something. The next question comes soon. Mark your best guess and move on.
Do whole listening parts under time pressure. Chinese Testing International publishes sample questions and runs an official mock test site, hskmock.com. Our guide on preparing for the HSK shows when to start with mock exams, and the guide on the HSK 3.0 exam lists all parts and times.

How Hanzi By Heart helps with listening

Hanzi By Heart has all 5,400 words of HSK 3.0 levels 1–6 and trains each one three ways: reading, listening and writing. In listening mode you hear the word first and recall its meaning before you see the characters. Listening has its own review schedule, so a word you read easily keeps coming back until you also understand it by ear. Every word comes with example sentences and audio.

Questions and answers

How many listening questions are in HSK 3.0? 20 at level 1, 25 at level 2, 30 at level 3, 32 at level 4, 35 at level 5 and 40 at level 6, in about 12 to 30 minutes.

Why can I read Chinese but not understand it when I hear it? Reading and listening are separate skills. If you learned a word from the page, you also need to practise hearing it, at normal speed and in sentences.

How fast is Mandarin spoken? In one study, speakers read short texts aloud at about 5.2 syllables per second. In everyday speech the pace depends on the speaker and the situation.

Are there many homophones in Chinese? Yes. Of the 5,400 words in HSK levels 1–6, 670 sound exactly like another word on the list, tones included. The context tells them apart.