“Hope everyone will drink a glass of milk after watching the film.”

The surreal line blinked across the screen just before the credits rolled at the end of a 4K restoration of Luc Besson’s cult-hit “Léon: The Professional,” screened years ago at the Shanghai International Film Festival.

Most viewers missed the quiet wink — to the glass of milk the hitman drinks like a ritual. But Sylvia Zhang caught it. “Someone had manually timed that line to appear at just the right moment,” the 27-year-old tells Sixth Tone.

“It was the first time I understood there was a person behind the subtitles — syncing every line in real time, matching it to the characters’ voices.”

For years, that’s how hundreds of foreign films have screened in Shanghai: Chinese subtitles cued live from plain text files, line by line, from laptops behind the curtain. The job falls to a rotating crew of volunteers — mostly students — who sit just out of view, tapping in each cue by hand.

Now in her second year as a volunteer, Zhang was back behind the laptop at the 2025 Shanghai International Film Festival (SIFF), which wrapped its 27th edition on June 22. “It’s not just pressing two buttons,” she says, comparing the work to DJing. “You’re always adjusting — speeding up, slowing down, staying alert for anything unexpected.”

Since 2017, Meng Chenhui has managed the Shanghai International Film Festival’s subtitle team. Each year, about two-thirds of the films arrive without embedded Chinese subtitles — meaning the lines must be cued manually, in real time.

Roughly 300 volunteers take on the job. “Most are students, born between 1998 and 2000,” says Meng, who once subtitled 55 films in a single festival.

“Many of the films we screen are still in early circulation,” he explains. “Especially with foreign dialogue or local dialects, there just isn’t enough time for producers to sync subtitles for every version.”

According to Yan, a staffer at an international film sales company who works regularly with SIFF, many foreign titles arrive without localized distributors — or finalized subtitles.

“Embedding subtitles takes time and money,” he says, asking to be identified only by his surname. “It’s often faster to use external display systems and refine the files with help from volunteers.”

Typically, the festival’s translation partner takes three to five days to produce a draft. Operators then spend about a week rehearsing the timing before the screening.

And with hundreds of films arriving each year, each screening is more akin to a live performance — where the smallest delay can throw off the rhythm. Most days, audiences rarely notice the effort. And for the volunteers, that’s the goal.

Lost in sync

Deng Jialin first noticed a subtitle operator two years ago at the Shanghai International Film Festival — hunched over a laptop, clicking through lines in real time.

“I didn’t understand why the job even existed,” she says now. “Actors move their mouths, and the volunteer presses buttons on a computer. What could be so hard about that?”

Curious, she joined their ranks this year. And everything fell apart.

She had spent days preparing subtitles for a film in English at this year’s festival — looping dialogue, memorizing every cue. With no embedded subtitles, she’d have to trigger each line manually, relying only on her ears.

“You have to catch the exact moment a character starts speaking,” says Deng, 28 and pursuing a Ph.D. at a university in Shanghai. “It’s like training for the IELTS listening exam.”

But when the lights dimmed, Spanish filled the theater, and the film began with an entirely different subtitle file than the English one she’d rehearsed.

“There were all these interjections — ‘¡Olé!’ ‘Aiyaya!’ — none of them were in our file,” she recalls. “The opening line sounded like a news report — something like ‘Asian News’ — but nothing matched. I actually started doubting my English. Was I really this bad?”

What followed was a blur. She and her partner leaned on muscle memory and what they could remember from the plot. With no understanding of the audio, they followed the timing of the English script they’d rehearsed — matching lines to cadence, not meaning.

By the time the credits rolled, their hands were clammy, necks stiff, and neither dared look up. “We were terrified someone would post about it on Douban,” she says, referring to China’s leading review app.

Sure enough, someone did. Her partner even messaged the user, politely asking them to delete the post — but never got a reply.

Meng Chenhui still remembers his first screenings and that same muscle memory.

In the weeks leading up to each festival, he’d loop his assigned films obsessively, memorizing every cue until it felt instinctive. “I’d start with the films screening last,” he says. “That way, the first ones were freshest in my mind.”

During prep, everything else fell away. “I’d watch with my eyes open, then listen with my eyes closed, and memorize every line. The only breaks were to eat, sleep, or go to the bathroom. By the end, only my fingers were moving. My eyelids could barely stay open.”

Subtitle files often come in two types. Some have rough timecodes embedded; others are just plain text. The first lets operators fine-tune timing. The second requires them to cue every line from scratch — watching, listening, and syncing in real time.

Shen Tianjun, a longtime volunteer in his 30s, finds the pre-timed files more stressful. The first line still has to be triggered manually, and if it’s even slightly off, the entire rhythm collapses. “We usually have to adjust the next few lines just to catch up,” he says.

That’s why the first few minutes are the hardest. Shen often replays them five times or more to memorize the rhythm.

When working with plain text, the job is often split between two volunteers, alternating roles to stay sharp. “It’s like flying a plane,” he says. “One person’s on the controls. The other’s watching everything — ready to jump in if anything goes wrong.”

But even when nothing goes wrong, it can still feel brutal.

In one theater, volunteers were stationed in the orchestra pit — a sunken space designed for symphonies. “I had to constantly shift between the small laptop screen in front of me and the massive cinema screen above,” says Deng. “It felt like doing cervical stretches the whole time.”

In another venue, the sound system proved punishing. “We were sitting directly under the speakers, and they screened ‘Saving Private Ryan,’” she says. “Oh my god — every explosion felt like it hit my eardrum. I genuinely thought I was going deaf.”

And when things do go right, no one is supposed to notice.

That’s by design. Operators are trained to disappear. “We wear black, dim our screens, and keep our keystrokes as soft as possible,” says Liu Xinyi, a postgraduate student in Shanghai. “Since everyone’s here for the movie, we try not to exist.”

Sylvia Zhang sees the role as a kind of conduit. “In front of the big screen, I feel like just another member of the audience,” she says. “But when I’m behind the curtain, cueing each line, I feel like a bridge, connecting the emotion on screen to the people in their seats.”

Yet sometimes, the audience notices — just barely.

After one screening, someone approached Deng and her partner. They braced for criticism. Instead, the individual said: “It was great — really well done.” Most people only asked where the restroom was. “That one compliment,” Deng says, “felt like it came from an angel.”

Shen once received a bottle of water from a stranger — a small gesture he shared with other volunteers with pride. Another time, a viewer asked for an autograph. “My colleague just smiled and said, ‘I’m not a star,’” Deng recalls.

Those brief moments stay with her. “Even when the plot feels a little cliché, the emotions in the theater are real,” she says. “If one person tears up, the whole row follows.”

And when it’s all over, a single line appears in the credits — small and easy to miss: “Subtitle Operation.”

“That’s enough for me,” Deng says.

Editor: Apurva.

(Header image: Visuals from VCG, reedited by Sixth Tone)