12.09.2021, 19:20
Yes, if it needs additional manual correction it don't fits well to Hybrid concept. It is strange that image to text engines still produce mistakes. Robots are so learned that can recognize any complicated text on captchas, but still can't read DVD subtitles without mistakes 
Tested few more DVDs and can't see any mistakes in converted subtitles yet. Sure i didn't test full length movies, but at first look to random places there are no any missed or incorrect text at all. So i guess it all may depend of OCR engine.
Recognizing speed also may be different and depends of program design and optimization.
Probably soon or late i will transcode all those crappy DVD subtitles from my collection to text format.
Anyway even if it won't happened in Hybrid, currently i can recommend Subler with OCR Tesseract language data modules as free and stable tool to convert to IDX/SUB DVD subtitles to SRT.
Tested few more DVDs and can't see any mistakes in converted subtitles yet. Sure i didn't test full length movies, but at first look to random places there are no any missed or incorrect text at all. So i guess it all may depend of OCR engine.
Recognizing speed also may be different and depends of program design and optimization.
Probably soon or late i will transcode all those crappy DVD subtitles from my collection to text format.
Anyway even if it won't happened in Hybrid, currently i can recommend Subler with OCR Tesseract language data modules as free and stable tool to convert to IDX/SUB DVD subtitles to SRT.