Chinese Listening Practice That Starts Below the Sentence
Almost every Chinese listening exercise starts at the dialogue, which is far too high if you cannot yet cut the stream into syllables. The beginner unit is the syllable, then the word, then the phrase. Tingyin drills exactly that ladder and nothing above it; when you are ready for conversations, HelloChinese or a tutor takes over.
Every listening exercise starts one floor too high
Search for Chinese listening practice and the results are remarkably uniform. A short dialogue plays. Two people order food, or ask directions, or discuss the weather. Then you answer three comprehension questions about what was said.
This is a good exercise. It is also, for a beginner, an exercise in a different skill from the one being advertised. If you can follow the dialogue, you were already able to hear Mandarin. If you cannot, the questions do not tell you why not, and the honest way most people pass them is by catching two recognisable words and reasoning about the rest from context. That is a useful life skill and it is not listening practice - it is guessing, with a score attached.
The gap is easy to miss because it has no name. Between knows no Mandarin at all and can follow a slow dialogue there is a long stretch where the sound arrives as an undifferentiated stream, and nothing in the standard curriculum is aimed at it. Courses jump from the pinyin chart to conversations in about two weeks, because that is what feels like progress.
Tingyin sits in that gap deliberately. It is a tone-listening trainer: a clip plays, you pick which tone you heard, it tells you immediately. There are no dialogues in it at all, and there never will be. Level 1 is free with no account, level 2 is free once you sign in, and levels 3 and 4 are paid.
You cannot understand a stream you cannot cut up
The reason dialogues fail beginners is a problem called segmentation. Speech has no gaps in it. The spaces you see between written words are a convention of writing; what reaches your ear is continuous, and the fact that you hear separate words in your own language is something your brain does for you, using knowledge it spent years acquiring.
In Mandarin the cues your brain uses for that job are partly different, and one of them is tone. A syllable's pitch contour is part of its identity, in the way a vowel is - not a layer of expression on top. If you cannot yet hear the contour, you are trying to segment the stream with one of its main signals switched off. Words blur into each other, and the ones you do catch feel like they arrived by luck.
This is also why the standard advice - listen to more Chinese, watch shows with subtitles, get exposure - works so much better for people six months further along than for people at the start. Exposure builds a category only if you can tell the instances apart. Before that it is just noise you are being patient with. The mechanism, and why production and perception come apart so sharply, is the subject of why Mandarin tones are hard to hear.
The ladder: one syllable, then two, then three
The unit of beginner listening practice is the syllable. Then the word. Then the phrase. Then - much later than anyone wants - the sentence. Each rung is a genuinely different listening problem rather than a longer version of the one below it, which is why skipping a rung does not work.
| Rung | What you are learning to do | What is new |
|---|---|---|
| One syllable | Identify a tone with nothing else to go on | The four contours, in clean isolation |
| Two syllables | Hear two tones in sequence, and their interaction | Sandhi, the neutral tone, relative pitch |
| Three syllables | Hold a pattern across a whole word | Rhythm, stress, and where the word boundary falls |
| Four syllables | Track a contour through a phrase | Speed, and running out of working memory |
| Sentence | Follow meaning in real time | Grammar, vocabulary, everything at once |
Tingyin's four levels are that ladder, and the corpus can be counted. Of 637 clips, 280 are a single syllable, 237 are two, 107 are three and 13 are four. By level it comes out as 80 clips at level 1, 200 at level 2, 237 at level 3 and 120 at level 4 - which is to say the first two levels are the single syllables, level 3 is the two-syllable words and level 4 is everything longer.
Level 1 is narrower still: single-syllable minimal pairs, the same syllable presented in each of its tones. It is the smallest useful listening exercise that exists in Mandarin, and it is where most people should spend their first fortnight. It is also free without an account, which is the least awkward way to find out whether your ear is the actual bottleneck.
What changes when the second syllable arrives
The jump from one syllable to two is the largest on the ladder, and it catches people who thought they had finished with tones. Three things arrive at once.
- Tones stop being what the page says. In connected speech Mandarin systematically substitutes tones - most famously a third tone before another third tone, which is pronounced as a second. Tingyin applies sandhi in 55 clips and labels every syllable with the tone as spoken, which is why not one of its 637 clips is tagged third-then-third: a word written 3+3 is stored as 2+3, because 2+3 is what you hear. The sandhi rules go through the rest.
- The neutral tone appears. 80 clips contain one, and all of them are at least two syllables long, because a neutral tone has no contour of its own - it takes its pitch from the syllable before it. You cannot practise it in isolation. It only exists attached to something.
- Pitch becomes relative. On its own, a first tone is simply high. In a word, what you are judging is where each syllable sits compared with the one beside it, which is a harder perceptual task and a different one.
None of this is visible in a comprehension quiz. You will simply find the dialogue hard and have no way to tell whether the problem was the vocabulary, the grammar or the fact that 可以 did not sound like kě yǐ.
What a beginner listening week looks like
A workable shape, assuming you have perhaps forty minutes a week to spend specifically on listening rather than on vocabulary or characters:
- Five sessions of eight minutes, not one of forty. Perceptual categories consolidate between sessions. Frequency does more work than duration here, and it is one of the few places in language learning where that is straightforwardly true.
- Always commit to an answer before the feedback. A forced choice with an immediate correction is the whole engine. Listening without answering is barely practice at all.
- Stay on single syllables until you are reliably right. Not perfect - reliably. If you are still coin-flipping between second and third tone, moving up to two-syllable words adds difficulty without adding learning.
- Keep some unstructured listening running alongside. A podcast you barely follow, a show with subtitles, a conversation partner. It does a different job - voices, speed, rhythm - and a drill on a fixed corpus cannot replace it.
The thing to protect against is the temptation to measure progress by content difficulty. Moving up to harder dialogues feels like advancement; going back to single syllables feels like regression. The scoreboard that matters is your hit rate on the rung you are actually on.
A week like this survives contact with a real schedule better than it looks, because the drilling half of it fits into gaps rather than needing a desk - and it keeps working in a tunnel, which is where a lot of those gaps are. Practising with no connection covers how to set that up and what the audio actually weighs.
When dialogues become worth doing
Dialogue practice is not wrong. It is a later rung, and it becomes efficient at roughly the point where you stop having to work out each syllable and can spend your attention on meaning instead. A practical test: play a two-syllable word you know and see whether the tones arrive on their own, without a deliberate act of decoding. If they do, dialogues will start paying back the time you put into them.
At that point Tingyin has largely done its job, because it does exactly one thing and stops. It teaches no characters, no grammar and no vocabulary, it has no dialogues, and it cannot hear you speak - there is no microphone in the exercise, so it will never tell you whether your own fourth tone is too gentle. HelloChinese and Duolingo run a full beginner course with conversations in it; Pleco is the dictionary you will want for words you meet in the wild; an italki tutor is the only thing on the list that can actually hear you. For a sense of how the pieces sequence over months rather than weeks, see how long it takes to learn Chinese by skill, and for the mechanics of the drill itself, how to practise tones and what does not work.
If you want to start at the bottom rung, level 1 is 80 single-syllable minimal pairs and needs no account at all. Level 2 opens when you sign in. Levels 3 and 4 - the two-syllable words and everything above them - are paid, at $5.99 a month or $15.99 once on the web, which also adds progress sync across devices and offline downloads. Buying inside the app instead goes through the store, where the prices are higher.
Frequently asked questions
How should a complete beginner practise Chinese listening?
Below the sentence. Start with single syllables and tone identification, move to two-syllable words, and leave dialogues until the tones arrive without effort. Tingyin covers only that lower stretch - 280 single-syllable clips, then 237 of two syllables - and level 1 needs no account. HelloChinese and Duolingo will give you the dialogues once you can use them, which for most people is a couple of months in, not a couple of weeks.
Why can I not understand spoken Chinese even though I know the words?
Usually because you know them as written words rather than as sounds. Speech has no gaps in it, and Mandarin uses tone as one of the cues that mark where one syllable ends and the next begins - so if the tone is not yet audible to you, the stream will not come apart into words you recognise. Reading vocabulary does not fix this. Only identification practice on audio does.
Are Chinese listening exercises with comprehension questions useful?
Yes, once you are past the segmentation problem. Before that they mostly measure how well you guess from context, because two recognised words plus a plausible situation will get you through most beginner quizzes without hearing the sentence. They also give you no diagnosis: a wrong answer does not tell you whether the vocabulary, the grammar or the tones defeated you.
How long should I stay on single syllables?
Until you are reliably right rather than perfect - most people need a couple of weeks of short daily sessions, and the second-against-third contrast takes longest. Moving up early is the common mistake, because harder material feels like progress while the hit rate quietly stays at chance. Tingyin's level 1 is 80 clips of exactly this, free without an account, so there is no cost to staying there.
Does Tingyin teach vocabulary or characters?
No. It plays a clip and asks which tone you heard - that is the entire application. There are no characters to learn, no grammar explanations, no vocabulary lists and no microphone, so it will never comment on your pronunciation either. It is a supplement for one specific weakness. Pleco handles the dictionary side and Anki the vocabulary side far better, and a tutor handles everything Tingyin cannot.