Yeah, it is good for accessibility, that’s why I merely dislike it over absolutely hating it. The implementation of an LLM will help those who need it, anyone that doesn’t want to use this can handily avoid it.
Good question, I don’t know if the feature has officially been deployed yet. You should keep an eye out for any feature reviews, as I’m not interested to find out.
I also wonder how well it would work at translation. I have a lot of old DVDs that are in Mandarin that do not have subtitles and finding subtitles would be very difficult.
I mean, if I had to first translate them and export the subtitle file and then put that into a translator and then it reimport it I can make do with that but the lower the bar is the more likely I’ll actually consume the media
Depends, I think. Phonetic-based character systems like hiragana are done very well in portable translators because the characters are distinct and easily reproduced, even without ‘ai’. Glyph based ones and word-based will be harder, e.g. YouTube subtitles tend to be janky from time to time.
Oh, and I think the plan is to transcribe in real time, not via subtitle files.
Hrm, I don’t like it but thankfully VLC isn’t going to force it on users like most scumbag companies do.
At least this is a legitimately good use case for AI that improves accessibility.
Yeah, it is good for accessibility, that’s why I merely dislike it over absolutely hating it. The implementation of an LLM will help those who need it, anyone that doesn’t want to use this can handily avoid it.
Is it accurate?
Good question, I don’t know if the feature has officially been deployed yet. You should keep an eye out for any feature reviews, as I’m not interested to find out.
I also wonder how well it would work at translation. I have a lot of old DVDs that are in Mandarin that do not have subtitles and finding subtitles would be very difficult.
I mean, if I had to first translate them and export the subtitle file and then put that into a translator and then it reimport it I can make do with that but the lower the bar is the more likely I’ll actually consume the media
Depends, I think. Phonetic-based character systems like hiragana are done very well in portable translators because the characters are distinct and easily reproduced, even without ‘ai’. Glyph based ones and word-based will be harder, e.g. YouTube subtitles tend to be janky from time to time.
Oh, and I think the plan is to transcribe in real time, not via subtitle files.