I've translated a couple scenes for myself using Davinci Resolve's AI transcription tool and various JP to EN text translators. And there is a lot of manual intervention, re-processing, and a sprinkle of filling-in-the-blanks by context. it results in serviceable subtitles but probably with a lot of inaccuracies. I would also love AI subtitles on SLR but it wouldn't be very good i think. But in the case of this sort of content, maybe low quality and inaccurate but somewhat coherent and plausible could be acceptable?
