Comment to Keep it real!
-
Thank you Nathe. Yes, we are well aware of the need for subtitles. This is not a trivial task though. Automated subtitles require a solid AI engine to process. There are a few available, but none will allow easy integration or an API for external voice recognition. If you know of one, please share.
We started working with the Google’s API.AI now to develop bots support in JOT Messenger, but even this API is still limited to text input only. Google has the technology to process voice, but they won’t share or sell it yet.
Perhaps one option we should go for in the interim is uploading a copy of such videos to YouTube to allow for captioning. Alternatively the author has to write the trbascript and post it as a Post with embedded video from the album. What do you think?-
What about sign language to text, nearly impossible... at least in consumers side. Which is why I believe that text input is more universal. Not everyone likes to do it because using time-frames and fill in the captions... naturally, it's time consuming for most.
Anything that's Google, I stay away. Even their "open-source" that requires Google account which requires running nonfree software (Javascript sent by the site).
I think that's doable. Look up "WebVTT" (Web Video Text Tracks), timed text in connection with the HTML5 element.
-
I'm pretty sure that we'll have some kind of service or API to convert voice to text soon, as AI engines progress. WebVTT spec as I understand is just for showing captions in the standard HTML5 video element, which is handy for standardized playback but doesn't help with recognition in any way. Still need something to "understand" all the talking.
Perhaps that's an idea for a community? Say, people share most interesting/popular videos and some members write up transcripts for the rest of the community to read. You can monetize by setting up a marketplace for quick transctipt service, or tipping, or subscription based access with revenue sharing.
-
