Extracting hardcoded subtitles using OCR

[1v0live]

Since i am having problems posting a software review, i am giving here my opinion.

Surprised to see VideoSubFinder so highly rated. I was recommended this by a friend to extract hardcoded subtitles together with Subtitle Edit doing the OCR of the extracted images.
Not only the UI is super ugly and confusing, it fails too much at recognizing the subtitles, sometimes only grabbing garbage. Even the corrected captured subtitles were so poor (image quality) that Subtitle Edit could not OCR them. Also Subtitle Edit to blame cause it is using the old Tesseract engine.

This is all old tech now, for this purpose.

I then discovered 2 programs: VidSubX and VideOCR. By the way, surprised that they have not yet a VideoHelp page. (I just submitted them, let's see what happens).
They are very similar and super easy to use. Just open the video, select the region where the subtitles appear and start. It automatically creates a srt file with no need for a second tool. They have 2 versions: CPU and GPU. I used only the GPU version since it is more powerful. They both use PaddleOCR.

VidSubX took 20 minutes for a 95m movie. It completely nailed it. Even captured some credits that appeared in the subtitles region.
VideOCR took 11 minutes for the same movie. Same result with even some captures of scenery words that happened to appear in the subtitles region.

[Subtitles]

Dealing with hardcoded subtitles has always been a problem because of the lack of a reliable and accurate OCR applications.
The new technologies of PaddleOCR and VideOCR opens the door to deal with hardcoded subtitles in a way that nobody could have dreamed it is possible.
I have tested over the past few weeks many videos in different languages English, French, German, Polish, etc. and the results are simply amazing.

[Asdfgf]

Thanks! I will give them a chance. I've made many subtitles using videosubfinder and abby fine reader, but if you can do the same job using only one tool... that sounds nice!