Subtitling and captioning (technically the same thing) transmission in OTT platforms for linear channels is usually handled by taking subtitles from the video feed for the channel and convert it to the various OTT transmission formats (WebVTT etc).
The video signal, typically SDI or DVB TS, contains EBU teletext (SDI) or DVB Teletext, ETSI 300 743 (TS) subtitles. These can be converted to the OTT formats as they are text based.
In some cases there is no subtitle data available, except in the format of subtitles as visible on the TV screen, referred to as DVB bitmaps. This can be solved by Cavena OCR reading the subtitles for conversion to text.
Do note, If Cavena transmission system installed on site, the subtitle data is already available, it's just another port on the Cavena STU, interfacing the OTT platform.
Cavena subtitle format
Because teletext is limited to European languages, it cannot handle Asian languages.
Cavena has developed a proprietary format used in packet 31 generic data transmission in EBU teletext to carry Asian languages as well as features not available in teletext such as more exact positioning, more than 37 characters per row, frame accurate pre-transmission etc. The Cavena subtitle format is transmitted in the P31 generic data transmission of teletext and thus works with any signals and systems handing teletext, including SD-SDI, HD-SDI (OP-47/SMPTE-2031) and DVB Teletext.
Use of this format in an OTT platform means it can now handle Asian languages and any other language in the world not supported by teletext. Support for this format is added to the OTT platform to be able to interpret the data and convert it to the transmission formats like WebVTT etc.
DVB bitmaps to OTT platforms
If the incoming video is a DVB TS containing DVB bitmap subtitles (ETSI 300 743), it is harder to convert to OTT formats as the bitmaps have to be OCR read. Cavena provides solutions to OCR and transcode DVB bitmap subtitles to all available transmission formats, including EBU Teletext, DVB teletext and the Cavena P31 format.
OCR reading is not 100 % correct at all times and requires “training” to work well with each individual font coming in. OCR reading works for any language, any font. Note. None Latin characters can only be supplied to the OTT platform with Cavena P31.