Try it on one N5 sentence
Press Play, listen, and say the sentence in the pause. Listen has the same two settings for a whole level.
Start a session at your level
Every one of the 2,254 example sentences has a clip. The times below add up the real clip lengths and the pause after every play.
| Level | Points | Sentences with audio | 1×, 1 s pause | 3×, 6 s pause |
|---|---|---|---|---|
| N5Listen | 63 | 266 | 19 min | 2 h 3 min |
| N4Listen | 87 | 358 | 29 min | 2 h 57 min |
| N3Listen | 117 | 468 | 44 min | 4 h 9 min |
| N2Listen | 124 | 498 | 49 min | 4 h 32 min |
| N1Listen | 166 | 664 | 1 h 15 min | 6 h 32 min |
Each link opens that level in Listen. Playback does not start from the link: choose the whole level or some chapters, set the plays and the pause, then press Start listening.
Settings by goal
Steady learner
2× with a 3 s pause, over the chapters you have reached
Listen and read on the first play, say it with the voice on the second, and say it alone in the pause. A whole level takes longer at this setting (1 h 22 min for N4), so pick chapters. Afterward, press Practice this chapter on your level's list (N5, N4, N3, N2, N1) to build the sentences, and open Review when points are due.
Crammer
1× with a 1 s pause, the whole level
This is the most sentences per hour: one pass takes 29 min for N4 and 44 min for N3. After the pass, press Mark as hard on your level's list for any point you could not follow, play those again from the Hard points row in Listen, then build their sentences in Practice. Points you practiced come back in Review after 1, 3, 7, 14 and 30 days.
Refresher
1× with a 1 s pause, a level you have passed
N5 takes 19 min and N4 29 min. Note the points whose sentences you could not follow; after the pass, press Mark as hard on them in the level's list and build their sentences.
How to shadow one grammar point
With the sentence above, ピザが食べたいです。 (I want to eat pizza.):
- Set the player to 3× with a 6 s pause and press Play.
- First play: read along and find the grammar. 食べたい is the stem of 食べる, 食べ, with たい, "want to".
- Second play: speak with the voice, at its speed, without waiting for it to finish.
- In each pause: say the sentence once on your own, then check it against the text.
- Third play: look away from the text and shadow from the sound alone.
- Then build it: on the ~たいです page, Practice asks you to put each sentence's parts in order.
Grammar sessions in an AJATT immersion routine
AJATT (All Japanese All The Time) is a learning method built on a high volume of immersion in Japanese made for Japanese speakers: shows, podcasts, books and games. A grammar session is a different kind of input. Every sentence carries one known grammar point, the text and its meaning are on screen, and you set the pace. The sentences are written for learners and read by a synthesized voice, so a session is an addition to an immersion routine, not a replacement for it.
Two ways to fit it in: play a level's sentences in Listen before a week of shows at that level, or save the points you heard in a show and could not place, then play them from the Saved points row in Listen.
What research says about each part
- Saying it aloud. Words said aloud are remembered better than words read silently (MacLeod et al., 2010), and the same held for vocabulary in a second language (Icht and Mama, 2022).
- Reading while listening. Reading while listening built listening fluency more consistently than reading or listening alone (Chang and Millett, 2014).
- Shadowing. Shadowing improved listening comprehension in a small classroom study, mainly for learners at lower levels (Hamada, 2016).
- Building the sentence. Rebuilding a sentence is a form of self-testing, and testing yourself beats rereading (Roediger and Karpicke, 2006; meta-analysis: Adesope et al., 2017).
- Review on a schedule. Spaced review beat massed review across 48 second-language experiments (Kim and Webb, 2022). For grammar the results are mixed: one study favored spacing (Rogers, 2015), another found massed practice no worse for Japanese morphology (Suzuki and DeKeyser, 2017).
Each part has its own evidence. Whether combining them in one session adds up to more than the parts has not been studied. The shadowing studies are covered in more detail in what research says about shadowing.
Frequently asked questions
Is the audio native?
No. Every sentence is read by a synthesized voice: Google Cloud text-to-speech, a Japanese neural voice, at 0.8 times normal speed. It is clear and the same from one sentence to the next, but it is not a recording of a person. For real speech, add podcasts, shows or other Japanese made for Japanese listeners.
How long does a whole level take?
At one play per sentence with a 1-second pause, from 19 min for N5 to 1 h 15 min for N1. At three plays with a 6-second pause, from 2 h 3 min for N5 to 6 h 32 min for N1. The times add up the real clip lengths and the pauses; the table above lists every level.
Does Listen keep playing with the screen off?
In Chrome on Android, yes: it keeps playing with the screen off and shows pause and skip controls on the lock screen. On an iPhone with the screen locked it has not been tested yet. The player on this page stops when you switch tabs or turn the screen off.
Do I need an account?
No. Listen, Practice and Review are free in the browser with no account. Your progress is saved on your device.
Can I listen to only some points?
Yes. On a level in Listen, choose Chapters and tick the ones you want. The Listen page also has rows for your saved points, your hard points and the points due for review.
Related
- JLPT Grammar Dictionary: all 557 points from N5 to N1, with Listen, Practice and Review
- JLPT N5 grammar list: 63 points
- JLPT N4 grammar list: 87 points
- JLPT N3 grammar list: 117 points
- JLPT N2 grammar list: 124 points
- JLPT N1 grammar list: 166 points
- JLPT study guide: what each level asks for and how to prepare
- Japanese shadowing materials: audio to shadow beyond grammar sentences