Recording live subtitles - can it be done?

[Mevan]

Hey guys,

I am not spam. I am however asking for help. I figured if anyone would know it would you be you guys.

I work at a charity called Full Fact, we're an independent factchecking organisation based in the UK (Read more here: https://fullfact.org/about/). Every Wednesday we live factcheck Prime Minister's Questions. But to properly live factcheck and do it more efficiently we need to have a transcript. So for the past few weeks we've been trying to figure out if you can record subtitles as they are released, live.

Does anyone know how to do this? Could anyone make it happen? Would be grateful for any help or advice.

- Mevan

tl;dr cool charity needs help recording subtitles

[wesnewell]

Set up a manual record for Wednesdays from whenever to whenever on the channel(s) you want to record, or daily, weekend, weekday, etc. Whatever you want. Or a custom record schedule.

[stuartm]

The subtitles can be captured, MythTV does this by default alongside the video and audio, however your problem will be that DVB subtitles are image based and I'd imagine you'd want to convert that to text. To do that, without having someone manually transcribe them you'd need to feed the subtitle images through OCR (Optical Character Recognition). In short, what you are wanting can be done, but to be done efficiently and with the minimum of difficulty for your staff you'd probably want a simple dedicated application which takes the subtitle stream, parses out the image fragments then pipes them through an OCR algorithm. None of what I'm describing is very difficult and there are almost certainly even existing solutions out there somewhere, both free and commercial. Knowing where to look is the tricky part. If you can't find anything suitable, the option of hiring someone to write it for you remains, one of our US developers did something very similar as a paid side project for US (ATSC) closed captioning.

If you just want to be able to rewind the video and rewatch the subtitles then you can do this with MythTV, but that doesn't sound like what you're describing. In the interests of transparency I should also point out that if that's all that you want to do, MythTV would be overkill. Even a standard off the shelf freeview or freesat DVR should be able to do that much, and it's a lot less work to setup than MythTV.

P.S. Live subtitles aren't the most reliable, since they are being typed as quickly as possible errors are often made and some things are abbreviated, paraphrased or omitted entirely, however you probably already know this.