Back up the mountain today, no clouds and tropical heat, to pick oolong tea with a bunch of old grannies, who are all wearing conical straw hats. The director thinks it would be great if I could get one, too. Do they have any spares?

There is a lot of tooth-sucking and shouting in Hokkien, and then one of them says:

“You can just put a bag over your head!”

“I’ll put a bag over your head, you cheeky c—” I begin, but the director kicks me.

A straw hat is found, and a woman shows me how to pick the leaves.

“You’re doing it wrong,” she says. “You’re just grabbing the top three leaves and snapping off the stem. That might bruise the leaves before they’re ready for processing, but more importantly, if you do that all day every day, you will sprain your fingers. You do it like this.” And she levers three leaves off the stem by lifting her arm, not her wrist. while resting the stem on her index finger. It’s a deceptively small nuance, but one of the little things that we are there to capture.

I start to explain to the camera what she said, demonstrating… until she grabs my arm and says: “No, no! You’re still doing it wrong!”

“Yes,” I say, pointing at the camera lens, “I’m showing them!” It is good television and looks very natural. One of the most difficult micro elements of filming on this show has been the public’s inability to grasp that we need to shoot everything wide and again in close-up; that even impromptu moments require a second take for reactions, and that when demonstrating something I have learned, I often need to first get it wrong again. This is not a problem on a closed set with just a few people; you can explain it and they get it. It is only a problem when a crowd gathers and gets in the way, and everyone appoints themselves an expert.

The cameraman switches to his macro lens so he can zoom right in on my fingers doing it wrong, and then doing it right. This means selecting tea buds that aren’t in shadow, and making sure we both know which one we are talking about, and then slowly rehashing the events that have already been shot at a distance, repeatedly.

However, for reasons that defy understanding, we have an audience that has swelled to eighteen people, thanks to a local fixer we call Mr Jangles because of the fistful of keys that hang from his belt. He is apparently some kind of bigwig from the Iron Guanyin Appreciation Society (don’t laugh – their online feed has 30,000 subscribers), who has decided to document our documentary by taking pictures with an outmoded Canon and his BINGBONG annoying mobile phone. The director has already shouted at him three times to get out of shot or stop jangling in the background of every scene. Plus the usual drivers and wingmen, several random tea-pickers, a guy who was passing on a motorcycle, and our entire crew, which is nine more people. Oh, and someone’s hatchet-faced Chinese girlfriend. who waits until I am halfway through the shot before yelling from the trees: “YOU’RE DOING IT WRONG!”

I have nothing to do for hours on end, and then often a tiny window to perform every task planned, in the right framing, in the right light, with the right sound and background, without a passing motorcycle or granny with a hedge trimmer. The whole crew have done their level best not to cock everything up. All I have to do is say the words in the right order, without forgetting what they are, even though they are often in Chinese. If I get it wrong, then a light change (we are all hyper-conscious about the position of the sun, and the lag between takes is often enough for it to be palpable) or sound change will mean ten more minutes’ faffery. It wastes everybody’s time and concertinas our schedule later in the day, which will often mean a cancelled shot from the end. Time is money, and we will never return to this mountainside, so that idle heckle has just cost us a shot from the end of the day.

When I am talking to camera, I am trying to remember what I am supposed to say, obviously. A 20-second speech has to be carefully plotted so as not to accidentally imply that Taiwan isn’t part of China, or mix up oolong with pu’er, or forget to mention the right dynasty, or offend National Geographic’s Standards & Practices arbiters, who will make us throw a take away if they don’t like it. So the last thing I need is gesticulating, whispering, hand waving, or people dicking about with their phones (BINGBONG). It’s difficult enough to remember at all times to maintain eye-contact with the lens, rather than the director or cameraman, who are usually also in my line of sight, so the last thing I need is Mr Jangles poking his head out from under the tripod to try and sneak a photo.

Mr Jangles, in fact, has appointed himself the director’s assistant, and insists on “translating” anything she says and bellowing it up the hill in Hokkien. However, since he doesn’t actually speak English, he usually forgets the words “don’t” or “not”, and is the cause of several unwarranted mass exoduses of grannies, packing away of cameras, disappearing straw hats, and other continuity nightmares. But for some political reason I don’t comprehend, we can’t get rid of him, or any of the people he is shepherding around in his car. He then reveals that he has already been uploading his pictures of us straight to the internet, which is not his right to do, and technically contravenes several terms in our contracts.

I have been thinking a lot today about Gwyneth Paltrow, and the kerfuffle that once erupted after she supposedly demanded to be taken a mere few dozen metres from her trailer to the set of Shakespeare in Love in a golf cart. Some media outlets condemned this as prima-donnish behaviour, although the Clements contingent immediately noted that she was wearing an Elizabethan dress and facing a football pitch’s worth of muddy ground, so her decision was probably intended to save her wardrobe mistress three hours of late-night laundry.

Similarly, Tom Cruise is notorious for having banned extras from his eye-line on film sets. This has been regularly touted as evidence that he is quite mad, really, but I will observe that as the producer of his own films, it is his own money he is wasting if a take is ruined because someone tries to snatch a selfie, suddenly slaps a mosquito on their neck, or downloads the contents of their left nostril into a nearby ditch. If I had a way of napalming the grove of trees next to the tea plantation today, and could thereby rid myself of a bunch of jangly, muttering interlopers, I would have happily done so.

Up the mountain one more time, to shoot me carrying 40 kilos of tea on a shoulder balance, stumbling along the ancient pathway that winds through the hills to Quanzhou, the sea, and the world. The director wants to drone me alone on the hillside, which entails lugging my hefty load for half a mile through the terraces, pretending that I can’t see the Yuneec Q500 Typhoon as it whirrs above me. It then slowly pans up and out, leaving me a receding speck in the sunlight, stumbling through the neat green steps of tea trees, as the sun sets on the distant hills. This will probably be the closing shot of the whole tea episode.

There is no more time. We were supposed to record my closing homily in the sunset, but Mr Jangles and a bunch of other issues have chipped a minute here and a minute there, until we have lost an entire set-up. The sun has gone down, so it’s a 90-minute drive back into Quanzhou, livened up in the Buick by the sound of the director watching the drone footage and discovering that Mr Jangles turns up in it, trying to take a picture of me from the trees with his bloody phone.

Jonathan Clements is the author of The Emperor’s Feast: A History of China in Twelve Meals. These events featured in Route Awakening S02E01 (2016).