I love watching foreign TV shows and films. I am not blessed as a linguist though so I have to make do in reading subtitles.
The latest show that I am watching si Run On. It is a South Korean romantic comedy that stars Im Si-wan and Shin Se-kyung. They play as a national athlete and a subtitle translator. In one scene they are shown watching a movie and as the end credits roll she eagerly awaits the words Translated by Oh Mi-joo.
I have watched several kdramas by Shin Se-kyung like Six Flying Dragons (highly recommended) and The Bride of Habaek.
Yet this is not a review of Run On but rather was interested in the work of a subtitle translator.
In another lifetime I was in a BPO company that had this niche account that hired subtitle translators and I always saw them on the operations floor when I was going to the HR Department. They were less than 20 and had dual monitors, dictionaries, these huge headphones, and always be seeing watch a movie or a tv show as they would attempt to translate and transcribe what is being said into sentences that capture the context.
I've always thought it was cool that they were being paid to watch movies and tv shows. Yet never thought how hard it could be as well.
Sometimes I would catch them in the pantry having coffee and their topics would drift from whatever is happening in their lives to the projects they are involved in.
I would perk up whenever they talk about the latest show they are translating and hear if they would recommend watching it when it is shown or if they are struggling with getting it right.
As you can see it is not a direct translation word for word as they would need to provide context and understanding. It can't be too long as well because it can break the immersion in watching the show. If the meaning is off as well it leaves the audience scratching their heads and might even get them to stop watching.
During the early days when watching a foreign show, you downloaded it would contain the video, audio, and the translation file. You would load it all in VLC or whatever software you use and hope that it syncs.
The advent of newer technologies had it that everything is embedded in one file and all you need to do is pop in your favorite program and watch.
Nowadays on some online platforms, you have automatic subtitles that can run while watching the video.
Go to youtube and choose any video and click on the closed caption icon. Some videos may not have that option. It would usually have a default translated language or you can choose to automatically translate to whatever language you want.
Boy sometimes those auto-translate captions can just destroy your desire to watch it as there is certainly something off with the direct translation. Although there have been advances in transcription and translation software more often it still feels clunky and knows it wasn't a human that translated it. This is especially true with idioms.
We humans have beyond linguistic skills have comprehension and context. We know that there might be a deeper meaning of something. A subtle message or emotion is being invoked and a human translator would try to capture that.
They would look at the script, see the context of the environment, scene, and the emotion being conveyed and make the appropriate translation. They would read it out loud and check with colleagues if available. If not well hope for the best.
They would have several drafts before they are satisfied and all the while working on a deadline. So they need to balance between speed, context and the pressure of delivering good work.
In the end, they will submit it to editors who would go through it and make the appropriate changes or check for the quality of work. Then off to the client who hopefully will be satisfied with it.
Just like audio and video transcribers the advent of AI and machine learning algorithms is being seen as a disruptor in this shrinking industry. Having worked and trained transcribers in another lifetime a 20-minute interview could mean that one would need at least 2-4 hours to transcribe it. Longer if it is more complex.
Software that can automatically transcribe like Trint and Dragon Speech, as well as Google and Youtube, has its programs people have started using and often has about 60-85% accuracy which is compared to a human transcriber that has a goal of at least 96%.
So it was interesting for the show to showcase a translator's career and how it is not always a bed of roses. I understand her desire to see her name in the end and when you next watch a show and see a Translated by and you were happy with the translation say a silent thank you for your work to that person.