Closed Captions: to automate or not to automate?

When it comes to adding captions to your online marketing videos, there are generally two routes to take. Either you can use speech recognition software to create what are known as automated captions or you can hire somebody to manually create subtitles, or even do them yourself if you have the appropriate skills and training. Automated captions are often found on video content on social media sites such as Facebook and YouTube and viewers can activate or deactivate them as they see fit. However, are they really all they’re cracked up to be and are they really better than nothing?

Let’s start with automated captions. There are two main advantages to using speech-to-text software to create captions for your videos. Time and money. Automated captions are almost instant, and in some instances, like videos on YouTube or Facebook, you may not have to do or pay very much at all.

Another great perk of automated captions is to make live non-scripted events such as meetings or webinars more accessible. Many platforms such as Zoom have an automated captions function that uses speech-to-text software to display captions almost in real time for those who are d/Deaf or hard of hearing.

However, one of the main downsides of automated captions is a lack of accuracy. Fast talkers, strong accents and specific terminology are just some issues that can lead to automated captions having errors in them, meaning that the output can be unclear and confusing, and in some cases, completely useless.

Whilst they may take longer and be more expensive, captions produced by humans have many advantages over automated captions. Humans are more capable of understanding spoken dialogue than a machine, so can create more accurate captions with relation to what is being said.

However, good captions are not only measured by how accurately they depict the dialogue in a video. For closed captions aimed at people who are d/Deaf or hard of hearing, labelling sounds is also important for context and accessibility. The sound has to be both identified and labelled in an appropriate way for it to be useful to viewers who are unable to hear the video, something which is very difficult for machines to achieve.

Captions also need to be split in an appropriate way over lines, and from caption to caption. Different style guides have different ideas about this, but the effect is the same. They should be split when natural pauses occur, for example at sentence or clause level.

Finally, the way that automated and pre-prepared captions appear on screen can differ. Some automated softwares mean that subtitles appear one word at a time rather than a sentence at a time. Again, this is difficult for accessibility because viewers can get more meaning from reading a whole sentence rather than a word at a time, especially if some of those words have been transcribed incorrectly. Humans can also change the positioning of captions to avoid covering up important on-screen text where possible.

Whilst automated captions are very useful in contexts such as live events and as a placeholder whilst more accurate, effective and accessible captions are being created, they shouldn’t be used simply as a box-ticking exercise. For captions that make your videos more accessible and engaging, you should always involve a professional captioner in the process, whether that’s creating captions from scratch, from a script or from automated transcription. Your viewers will thank you.

To find out more about a professional captioner or subtitler can help you, get in touch.