Zoom Media’s speech to text is used for a variety of business cases one of which is automated (live) subtitling. Subtitling can help with reaching a greater audience online, but also to make the content better accessible for the deaf and hard of hearing. Since manual subtitling is a time consuming and therefore costly endeavour and because the increase in the amount of (online) content is unparalleled, broadcasters these days are looking into automated solutions. However, the automated output doesn’t always meet expectations due to the wide variety of rules that apply for different use cases. So in an effort to cater to as many people’s needs Zoom Media now supports a new set of properties for users to control their subtitles generated by our speech to text models.
Number of words
The standard Zoom Media output is set to a maximum number of 12 words per line of subtitling. However, sometimes use cases ask for less (or more) than that. For example, if a short clip should only show one word per line, customers can now customize that setting when using our API. So, start with 12 by default but tune it to any number of words you would like.
Length in time
The same goes for the length of a subtitle measured in time. This setting is set to 7 seconds by default, but now a user can change this to whatever length he or she likes.
When a silence occurs in a clip we start a new sentence when there is a gap of 1.5 seconds, but perhaps this is too short or too long in some cases. Not a problem, simply adjust the setting and you’re good to go.
Number of characters
By default, Zoom Media output contains a maximum of 42 characters. However different broadcasters prefer different settings in that regard, maybe 32 of 16, that as well is now totally configurable by the end user.