Since Adina doesn’t shy away from anything related to text, we have taken on subtitling videos. We have transcribed interviews, and translated videos from Finnish to English and from English to Finnish. Easy as peas, right? Well, actually no.
If you’re lucky, the interviewer has planned the questions beforehand and thus the questions are easy to understand and transcribe, but occasionally the interviewer loses the train of thought which makes both the interviewee and the transcriber ponder what the interviewer really means with the question. However, the transcriber must be able to interpret what the interviewer wants to say to avoid distorting the contents of the interview.
The interviewee on the other hand just expresses their stream of consciousness. The interviewee does not always end sentences, the contents of the message or subject changes in the middle of speech, and the interviewee may repeat the same thing several times but with slightly different phrases. Occasionally, the interviewee says something that doesn’t really have any meaning. The transcriber must still respect the speaker, try to create some sense in the subtitles, and not transcribe slips of the tongue, even though the video viewer might get a few laughs out of them.
Speakers who do not use their native language often cause problems. Even if the accent is comprehensible, the speaker may use words that are the combination of two languages, or words that vaguely resemble a real word. The transcriber must use the context of these words to find out the meaning, but there’s still a great risk of misinterpretation.
The transcriber must have good general knowledge, as the subjects of videos vary considerably. Sometimes you need to do a lot of background research to find the right terms and the names of people and organisations. Companies have their own terminology, and if you deviate from it, viewers may not understand your subtitles. That is why knowing the audience is important: is the video targeted towards customers, shareholders, or internal employees? When you know the audience, you’re able to deduct how familiar they are with the subject and then decide on the correct terms.
Because the viewer must be able to read the subtitles approximately concurrently with the speaker’s speech, the transcriber must consider how much text the viewer is able to read on the screen, as well as how much text fits on the screen. A good practice is to play the video in the background and try to read the subtitles while listening to the speech. If the transcriber, who by now knows the video inside out, does not have time to read the subtitles, neither will the viewers who watch the video for the first time.
How to make subtitles fit? You can leave out fillers, such as “actually” or “basically”, because they don’t have any effect on the understandability of the subtitles. The same goes for repetition: speakers tend to repeat their message, but you don’t have to do the same on the screen. You can also combine two or three sentences into one coherent sentence, especially if the speaker tends to leave sentences unfinished. Leave unnecessary adjectives out, and try to come up with shorter synonyms for words.
The most challenging aspect in subtitling is that the text has to work as text. The viewer may not understand the language of the speaker, or the viewer doesn’t have the opportunity to listen to the audio in the video. The text must be easy to read and form coherent entities.
If you just listen to the video once, the subtitles won’t be good enough. You need to test how the subtitles and the video work together a couple of times, and after all the editing make sure that you haven’t changed the contents of the subtitles too much compared to the speech.
It is a time-consuming and challenging process, but very interesting.