Introduction

If you want to learn Chinese by watching shows and movies you actually enjoy, the hardest part is the gap between learner material and native media.

There are two ways to cross it with comprehensible input. You can pick a show you really want to watch and pre-study its vocabulary until it becomes understandable (which is what I did with Qi Hun in my last post). Or you can keep things easy from the start: only ever watch things you already mostly understand, and let each podcast or movie set you up for the next one.

How far can that second approach actually take you?

Take Nirvana in Fire (琅琊榜), a 54-episode imperial court drama full of palace titles and political scheming. Suppose you start with the 1,500 most common spoken Chinese words (if you’re not sure where you sit, you can check with Hanzimatic’s vocab test) and follow two strict rules:

  1. Never start a podcast, movie, or show unless you already recognize at least 90% of the words in it.
  2. Learn each one’s most common unknown words up to 98% comprehension before moving on.

Can you get from 1,500 words all the way to Nirvana in Fire without ever dipping below 90%? It turns out you can – in roughly 10,000 total words, spread across 41 podcasts and movies.

Crossing the 90% wall

I collected transcripts and subtitles from a couple hundred Mandarin movies, TV pilots, and learner podcasts (filtering out Cantonese and non-Mandarin dialect films) and wrote a script to search for a path.

At 1,500 words, Nirvana in Fire is rough: you recognize 65.4% of the words across the season (61.4% on Episode 1). And not a single native Mandarin movie or drama hits 90% at 1,500 words – the closest is An Elephant Sitting Still (大象席地而坐) at 88.3%, followed by Au revoir Taipei (一页台北) at 87.1%.

That small gap is where learner podcasts come in. Six episodes of Lazy Chinese and Xiaogua Chinese take you from 1,500 to 1,837 words. On Step 7, An Elephant Sitting Still crosses 90.0%, followed by Au revoir Taipei and two more TeaTime Chinese episodes to bridge one last gap. From Step 11 on, it is native Mandarin cinema all the way. (Starting earlier works too: with enough beginner podcasts, the same unbroken 90% chain connects down from 1,000 words.)

Does a media path beat plain frequency flashcards?

Here is how your comprehension of Nirvana in Fire climbs across the 41-step path, compared to just memorizing the exact same number of most-common spoken words from a frequency list:

Through the modern movies, the media path actually trails a plain frequency list slightly. Around Step 19-20 (~5,500 words), the lines cross right as Big Fish & Begonia and Painted Skin start bringing in formal and historical vocabulary. By Step 41, you hit 90.0% across Nirvana in Fire (88.4% on Episode 1) at 10,211 total words – nearly 6 points ahead of studying the same number of general frequency words (84.3%).

The full 41-step path to Nirvana in Fire

At each step, the script looks at everything currently at 90%+ comprehension and picks the title whose new words help most with Nirvana in Fire. In total, it covers 41 titles and 8,711 new words (bringing you to just over 10,000 total).

Stage 1: Bridging into native movies (Steps 1-10)

# Title Arrive +New (Total) Nirvana Sample words
1 Lazy Chinese: Mother-daughter relationships 91.0% +129 (1,629) 66.0% 两个, 早就, 毕竟
2 Lazy Chinese: An illiterate Chinese woman wrote a book 92.7% +48 (1,677) 66.3% 不敢, 是不是, 心里
3 Xiaogua Chinese: You met me at a very Chinese time 90.5% +94 (1,771) 66.6% 心中, 委屈, 放在
4 Xiaogua Chinese: Finding My Past in Beijing 93.7% +20 (1,791) 66.7% 一条, 才能, 心情
5 Xiaogua Chinese: We Were Not Treated the Same as Men 95.4% +17 (1,808) 66.8% 问问, 付钱
6 Lazy Chinese: The most wanted killer in 2004 91.0% +29 (1,837) 66.9% 云南, 普通, 欺负
7 An Elephant Sitting Still (大象席地而坐, 2018) 90.0% +312 (2,149) 67.9% 放心, 没事, 倒是
8 Au revoir Taipei (一页台北, 2010) 90.1% +138 (2,287) 68.5% 竟然, 几个, 交给
9 TeaTime Chinese: Concluding 2024 90.0% +28 (2,315) 68.7% 今年, 一步, 加上
10 TeaTime Chinese: Reading The Stranger by Chiung Yao 90.9% +67 (2,382) 68.8% 出门, 再也, 感兴趣

Stage 2: Modern Mandarin cinema (Steps 11-18)

# Title Arrive +New (Total) Nirvana Sample words
11 Waiting Alone (独自等待, 2004) 90.1% +427 (2,809) 70.3% 当年, 就算, 多年
12 Us and Them (后来的我们, 2018) 90.2% +412 (3,221) 71.9% 大哥, 亲自, 明日
13 A Sun (阳光普照, 2019) 90.9% +303 (3,524) 72.8% 这位, 至于, 手段
14 What Women Want (我知女人心, 2011) 90.3% +344 (3,868) 73.5% 高手, 吩咐, 大夫
15 Police Story: Lockdown (警察故事2013) 90.4% +383 (4,251) 74.5% 不知, 如今, 一向
16 Badges of Fury (不二神探, 2013) 90.0% +337 (4,588) 75.8% 今日, 江湖, 十三
17 Lost in the Stars (消失的她, 2022) 91.2% +328 (4,916) 76.8% 处置, 出手, 性命
18 Mr. Six (老炮儿, 2015) 90.7% +415 (5,331) 77.7% 有所, 按照, 原本

Stage 3: Period, fantasy, and wuxia (Steps 19-41)

# Title Arrive +New (Total) Nirvana Sample words
19 Big Fish & Begonia (大鱼海棠, 2016) 90.1% +173 (5,504) 78.5% 心思, 要紧, 不妨
20 Painted Skin (画皮, 2008) 90.1% +143 (5,647) 79.2% 师父, 天下, 战场
21 Raise the Red Lantern (大红灯笼高高挂, 1991) 90.1% +185 (5,832) 79.8% 府上, 请安, 告辞
22 Gone with the Bullets (一步之遥, 2014) 90.1% +615 (6,447) 80.8% 公子, 日后, 便是
23 Monk Comes Down the Mountain (道士下山, 2015) 90.4% +185 (6,632) 81.4% 娘娘, 皇上, 说到底
24 A Brighter Summer Day (牯岭街少年杀人事件, 1991) 91.1% +271 (6,903) 81.9% 有意, 差事, 偏偏
25 Little Big Soldier (大兵小将, 2010) 91.3% +130 (7,033) 82.5% 太子, 兄长, 终究
26 A Writer’s Odyssey (刺杀小说家, 2021) 90.1% +226 (7,259) 83.1% 并非, 前来, 旨意
27 The Dream Factory (甲方乙方, 1997) 90.0% +418 (7,677) 84.0% 殿下, 公主, 亲王
28 The Four (四大名捕, 2012) 90.2% +209 (7,886) 84.9% 京城, 统领, 王爷
29 The Four 3 (四大名捕大结局, 2014) 90.6% +222 (8,108) 85.7% 此事, 宫中, 此案
30 Legend of the Demon Cat (妖猫传, 2017) 90.9% +199 (8,307) 86.5% 陛下, 皇后, 未必
31 Crouching Tiger, Hidden Dragon (卧虎藏龙, 2000) 90.8% +271 (8,578) 87.2% 想必, 乃是, 之下
32 Rise of the Legend (黄飞鸿之英雄有梦, 2014) 91.4% +147 (8,725) 87.5% 金陵, 暗中, 早已
33 Brotherhood of Blades (绣春刀, 2014) 90.3% +184 (8,909) 88.2% 刑部, 必定, 二位
34 The Thousand Faces of Dunjia (奇门遁甲, 2017) 91.9% +206 (9,115) 88.4% 这般, 此人, 听从
35 An Empress and the Warriors (江山美人, 2008) 90.2% +172 (9,287) 88.7% 只怕, 两位, 朝政
36 The Lost Bladesman (关云长, 2011) 90.1% +206 (9,493) 88.9% 今夜, 有劳, 天下人
37 A Touch of Zen (侠女, 1971) 90.1% +231 (9,724) 89.3% 罪名, 各地, 岂能
38 Double World (征途, 2020) 91.2% +122 (9,846) 89.5% 老臣, 恩准, 株连
39 The Last Supper (王的盛宴, 2012) 90.0% +108 (9,954) 89.7% 转告, 信不过, 郎中
40 The Untamed: The Living Dead (陈情令之生魂, 2019) 91.5% +128 (10,082) 89.9% 罢了, 地牢, 旧事
41 The Yin-Yang Master: Dream of Eternity (晴雅集, 2020) 90.3% +129 (10,211) 90.0% 宫里, 可知, 令人

My favorite quirk in the list is Step 27: The Dream Factory (a 1990s Beijing comedy about actors staging fantasies for paying clients) introduces 殿下 (“Your Highness”), 公主 (“princess”), and 亲王 (“prince”) because the actors stage a dinner with a fake princess to cheer up a lovelorn client. Step 26 (A Writer’s Odyssey) pulls a similar trick, jumping between a modern thriller and a novelist’s ancient fantasy world. And Gone with the Bullets at Step 22 is a 615-word marathon hiding behind a 90.1% arrival – pushing from 90% to 98% on a fast, talky two-hour film is still a lot of words.

Final thoughts

Recognizing 90% of the words in a subtitle file isn’t the same as effortless listening – speaking speed, accents, and idioms still make real speech harder than reading, and learning ~8,700 new words across 41 titles is still a year-plus project (at 15 new words a day, that’s over a year and a half). Still, it is encouraging how quickly native movies open up once you pass ~1,800 words, and that you can pick up almost all of that vocabulary inside movies you mostly understand instead of grinding through isolated frequency lists.

Starting at 1,500 words skips the hardest part for beginners, of course – getting from zero to 1,500 (or at least to that 1,000-word podcast threshold) is a different puzzle, which I’m planning to look at in a future post.

If you want to see where your own vocabulary lands on different shows, you can try Hanzimatic’s vocabulary test, browse the public decks, or upload subtitles from whatever you’re watching. Contact me at [email protected] if you have thoughts!