Hi. I use this program when watching subbed videos or playing choose your own adventure style FMV games. One thing I noticed is that often with subbed videos is that a sentence in the script is often broken up into multiple different subs with a single subtitled line staying on screen for less than half a second. It's pretty tough for translation services to keep up. One thing I noticed that due to the way these subs are broken up, common phrases are translated multiple times that could probably be cached somewhere on disk and retrieved avoiding the translation service round trip.
Also kind of niche but it would be nice if this feature supported multiple instances of UGTLive running so it doesnt hit stuff like file locking issues. I usually do this because I want to OCR multiple areas of the screen but for performance reasons and because its much easier for "settling" to happen so I dont want to OCR the entire screen and try to translate it. Usually the most common thing I do is have OCR running for persistent UI elements in one instance, while the other instance solely focuses on subtitles at the bottom.
Hi. I use this program when watching subbed videos or playing choose your own adventure style FMV games. One thing I noticed is that often with subbed videos is that a sentence in the script is often broken up into multiple different subs with a single subtitled line staying on screen for less than half a second. It's pretty tough for translation services to keep up. One thing I noticed that due to the way these subs are broken up, common phrases are translated multiple times that could probably be cached somewhere on disk and retrieved avoiding the translation service round trip.
Also kind of niche but it would be nice if this feature supported multiple instances of UGTLive running so it doesnt hit stuff like file locking issues. I usually do this because I want to OCR multiple areas of the screen but for performance reasons and because its much easier for "settling" to happen so I dont want to OCR the entire screen and try to translate it. Usually the most common thing I do is have OCR running for persistent UI elements in one instance, while the other instance solely focuses on subtitles at the bottom.