Whether it be TikTok, Instagram Reels or YouTube Shorts, most of us are familiar with vertical video content these days. In fact, everything from feature films to PR videos is being increasingly created with the vertical format in mind, even if the primary medium is an old-fashioned cinema or TV screen.
Much more focus is being placed on that middle third of the shot, in some cases to the detriment of the overall product, as marketers seek to ensure their product looks good on the social media mentioned previously.
Unfortunately, it’s not only the cinematography that suffers from being so centre-focused. Plenty of conventional videos now feature subtitles clearly designed to be squeezed into a 9:16 aspect ratio. The logic here is sound, I suppose. If the people creating the video can simply crop the video down to size without having to reshoot or resubtitle it, that saves them plenty of time whilst ensuring optimal appearance across social media. However, in the case of subtitles, that still only works if they are done well, following core principles like readability and sensible segmentation. In many cases, these guidelines are simply being ignored.
So, what works and, crucially, what really doesn’t? As a brief case study, I’d like to consider a video recently uploaded by the UK’s Foreign Secretary, David Cameron, not out of any particular political affiliation or agreement with the message, but because it has been widely praised as an excellent example of communication. Now, it does win marks for how concisely and clearly it gets the messages across, but the one thing that stood out to me was how poorly the subtitles let it down, and although I saw it in a classic aspect ratio over on Twitter, it was obvious it had been framed and subtitled with vertical media in mind.
There are two key aspects of good subtitling that I think really let this video down. The first regards sensible segmentation, essentially making sure each line starts and stops at a logical point in the sentence. The generally accepted guidelines on splitting your subtitle lines up go into a lot of detail, with rules such as avoiding ending a line on a preposition and trying to ensure each subtitle contains a “meaningful unit.” There’s a good reason for this, too: it helps viewers/readers understand what the subtitles say more easily and allows them to split their focus between the text and the visuals. In the case of Cameron’s video, you can see there are plenty of examples where individual lines cannot be called meaningful units at all. Here are just a couple of screenshots that I took:
Creating subtitles like
this would be akin to
writing a
sentence like
this for a
blog post.
It’s clunky and harder to read, so it should be avoided. Solutions can be found whilst ensuring the subtitles are suitable for vertical media, but in many of the examples found, nothing more complicated than simply following the guidelines is required. In more complex situations, you might rephrase/condense the speech whilst maintaining the full meaning, and there’s almost always the option to use two lines rather than one, even for vertical videos.
The next issue with subtitles like this stems from the poor segmentation, and that is poor readability. The “for what” subtitle in the screenshot above basically flashes in and out of view, almost certainly spending less than 1s on screen. Like splitting your subtitles up badly, this hampers viewer/reader understanding and makes everything somewhat unsettlingly staccato. Giving viewers a very short time to read a subtitle is like only giving readers half the letters in a legible font. It just makes it harder to understand the message you’re trying to get across.
Again, there’s no need to make things so awkward for your viewers just to fit into a vertical aspect ratio. For a start, by simply splitting your lines more sensibly, you wouldn’t be left with super-short, two-word lines briefly flashing up every second. Instead, you could have slightly longer subtitles, containing meaningful units, that give your viewers time to read and understand the message. Of course, you may not be able to fit quite so many characters per line into a vertical video, although industry standards suggest you can still get much closer to a nice 37 characters than the measly 10 characters in the word “supporting.”
Now, you may think it doesn’t matter too much because people can still hear what’s being said and the video in question is not being translated from another language. However, that would be to miss two key points. The first point is accessibility - adding subtitles enables more people to understand your message, particularly Deaf and hard of hearing people and non-native speakers. The second point centres on people’s use of social media, with various studies showing most people do not turn the sound on when watching videos on social media. Ultimately, to reach the widest audience possible, (good) subtitles are necessary.
Finally, I’ll turn to the elephant in the room. There is a reasonable chance the subtitles were created for Cameron’s video using AI, perhaps with a little tidying-up on spellings or words that the machine ‘misheard,’ but certainly without any trained human intervention on the segmentation or timing. While, like machine translation, there may be advantages to using such software to start the subtitling process, as we’ve seen from the examples here, subtitling is a skilled craft and, even with excellent speech-recognition, any AI output will need careful human intervention to ultimately deliver high-quality subtitles. That being said, my recommendation would always be to start with a human and get the job done right first time!
I hope this has been an interesting insight into best (and worst) practice when it comes to subtitling for vertical media. It’s something I’ve been doing a lot more of in my work over the last year or so, and I know it can sometimes be even harder than conventional subtitling. It would be great to hear from others doing similar work, and feel free to get in touch with me if you have more questions about the subtitling process, be that for vertical or traditional media.
Oll an gwella,
Chris
Comments