Practising Chinese Tones With No Connection
Tone listening is the one part of learning Mandarin that needs nothing from the internet once the audio is on the device. A clip plays, you pick a tone, the app says whether you were right - no server is involved in any of that. Tingyin’s whole corpus is 637 clips and thirteen minutes of audio, which is smaller than one podcast episode.

The places you would actually do this have no signal
The honest answer to “when will I practise” is almost never at a desk. It is the twenty minutes on a train, the walk back from somewhere, the queue, the middle of a flight. Those are the slots that exist, and they have two things in common: your hands are busy or your eyes are, and the connection is bad or absent.
Most language practice copes badly with that. Reading needs your eyes. Speaking needs a room you can talk in. Anything with a server in the loop - a grammar exercise that posts your answer, a dictionary lookup, a video lesson - stalls in a tunnel. Tone identification is the exception, and not by accident. The whole exercise is: hear about a second of audio, choose one of four or five answers, find out immediately whether you were right. Everything that loop needs is a sound file and the correct answer, and both are small enough to sit on the phone.
The entire corpus is thirteen minutes of audio
This is the part people do not expect. A tone-training corpus is not a library of lessons; it is a pile of very short recordings. Tingyin ships 637 of them, and the average clip runs 1.2 seconds. Added together, every clip in the product - all four levels - comes to about thirteen minutes of audio.
| Level | What it holds | Clips | Audio |
|---|---|---|---|
| 1 | Single syllables, minimal pairs | 80 | 1.5 min |
| 2 | More single syllables | 200 | 3.7 min |
| 3 | Two-syllable words | 237 | 4.9 min |
| 4 | Three- and four-syllable words | 120 | 2.9 min |

Thirteen minutes is less than a single podcast episode, and you will listen to those 637 clips hundreds of times between them. That ratio - tiny download, enormous reuse - is what makes tone drilling the natural thing to do with no connection, and it is why a download-once model fits it so much better than streaming.
How this works in Tingyin
Tingyin is a Mandarin tone-listening trainer: a clip plays, you pick the tone you heard, it tells you immediately whether you were right. There are no characters to write and no vocabulary lists. Level 1 is 80 single-syllable minimal pairs and runs without an account at all; level 2 adds 200 more and is free once you sign in. Every one of the 637 clips is a recording of a named human speaker carrying its source, speaker, licence and checksum, rather than synthesised audio.
Offline downloading is part of the paid tier, and there is no button for it. If your account has it, the app quietly fetches the whole corpus into the browser’s own cache shortly after you open it, and from then on training works with the connection off. Three things are worth knowing about how that behaves:
- It stores everything, not just the level you are on, so moving up a level offline does not hit a wall.
- It only reports itself as ready when every clip is actually stored, so a download that was interrupted resumes rather than pretending to be complete.
- Signing out clears the stored audio. That is deliberate - the files belong to the account that has the entitlement, not to the device.
On the free levels, training asks for the connection. A clip you have already played is usually still in the cache, so it is not that the audio is gone - it is that offline training is the thing the paid tier turns on, rather than something that happens to work if you happened to play enough clips first. Levels 3 and 4, progress sync across devices and offline downloads are what the payment buys: on the web that is $2.49 weekly, $5.99 monthly or $15.99 once for lifetime access. The web is the only place to pay today; the native apps are not published yet.
What to do with twenty minutes and no signal
A commute is short and repetitive, which suits this better than a long study session does. The thing to avoid is treating it as revision. Tone perception improves through forced choice under mild time pressure - answer before you are sure, then find out - and it does not improve much through listening to clips you already know while thinking about something else.
Stay on single syllables longer than feels necessary. The two-syllable levels are more satisfying and they let you lean on rhythm and on the sandhi rules to guess, which is exactly the crutch you are trying not to build. If a session on level 1 or 2 is still landing near chance, that is where the work is, and the method for actually hearing the tones is worth reading before doing more volume.
One more thing about headphones on a train: background noise makes third tone harder than anything else, because the low creaky part of it is the first thing a rumbling carriage eats. What the third tone actually does is the page for that, and if a run of them suddenly collapses on the Underground, it is the carriage rather than you.
Frequently asked questions
What is the best way to practise Chinese listening offline?
Download a set of short clips and drill tone identification by forced choice, because that loop needs no connection once the audio is stored. Tingyin does this directly: its paid tier stores all 637 clips on the device, which is about thirteen minutes of audio in total. Anki with a Mandarin audio deck works offline too and is free, though you grade yourself, which is where honest people quietly drift upward.
How do I download Chinese audio lessons for a flight?
Do it on wifi before you leave and check that the app reports the download as finished rather than in progress. In Tingyin the download starts on its own for a paid account shortly after the app opens, and it only marks itself ready once every clip is stored. Podcast apps like ChinesePod handle long-form listening offline in the same way, but a podcast trains comprehension rather than tone identification.
Is it safe to practise tones in a noisy place?
Yes, with one caveat worth knowing: noise costs you the third tone before it costs you anything else, because its low creaky section is the quietest thing in the set. Tingyin plays each clip on demand and lets you replay it as often as you want, so a bad moment costs a replay rather than a wrong answer. If you find your accuracy dropping only in one place, suspect the place.
What is the difference between offline downloads and progress sync?
Downloads put the audio on the device so training works with no connection; sync copies how far you have got to your other devices. They are separate things that happen to arrive together in Tingyin’s paid tier. Without sync, progress still exists - it simply stays on the device you earned it on, which is also how the free levels behave.
Can I really get anywhere doing this on the train?
For tone perception specifically, short repeated sessions are the format that works, so a commute is a good slot rather than a compromise. What it will not do is teach you vocabulary, characters or grammar, and it is not speaking practice - this is ear training and nothing else. Pair it with a tutor on italki for the parts a phone in a tunnel cannot cover.