The idea of captioning for the deaf and hard of hearing is to make comprehensible what is incomprehensible without sound.
It is not just about transcribing dialogue but also providing the maximum information about the development of the plot, the identity of the characters, their position within the image, their personality and mood, what is happening off screen, incidental sounds, music, its provenance, etc.: all the information provided by the soundtrack to a non-disabled audience, which receives it spontaneously and without realising it. It is through conventions established with the hard of hearing that we are able to communicate the information they lack, using dedicated software particularly suited to the current needs.
Live subtitling relies on high performance speech recognition, interfaced with applications dedicated to the correction, adaptation, formatting and distribution of television programme transcripts to broadcasters.