The main purpose of this program is to provide functionality for extract hardcoded subtitles (hardsub) from video.
It provides two main features:
1) Autodetection of frames with hardcoded text (hardsub) on video with saving info about timing positions.
2) Generation of cleared from background text images, which allows with usage of OCR programs (like FineReader, Subtitle Edit, Google Drive) to generate complete subtitles with original text and timing.
For working of this program on Windows will be required "Microsoft Visual C++ Redistributable runtime libraries 2022":
https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads
Latest versions were built and tested on: Windows 10 x64, Ubuntu 20.04.5 LTS, openSUSE Leap 15.4, Arch Linux (EndeavourOS Cassini Nova 03-2023)
For faster support in case of bug fixes please contact me in:
https://vk.com/skosnits
For donate:
https://sourceforge.net/projects/videosubfinder/donate
Features
- video text detection
- text extraction from video frames
License
GNU General Public License version 2.0 (GPLv2)Follow VideoSubFinder
User Reviews
-
The idea of the program is great. The problem comes with having to try and understand the technical language in the help file. I can't understand it. I am unable to register where the developer can be contacted. I have problems with Chinese OCR. Too much of the details in characters is removed in the filtration process. I have made some changes in: min_sum_color_diff = 300 (600 in original), between_text_distace = 0.08 (0.05 in original), text_centre_percent_offset = 0.9 (0,5 in original), and have gotten slightly better results. Does anyone have other tips on how to improve the filtration process so it doesn't remove important detail? About 80% of the characters are OCRed correctly now. The new version, 2.50, is much faster than the 2.10 version I have used (some trouble with video crashes in 2.30). The older versions had trouble ocring HD files. I see there's been a 'fix" for that. Does it fix the problem with the size of the subtitles? That's what I believe the problem was. Large files with subtitles didn't get as good recognition as a resized file. I have also had trouble opening files with avisynth avs. I'll check if it works with this version.
-
the program works fine but he choose when he want work. Crashing 9 of 10 times in "run search" and "create subs..." never works. my rig: win 10, 8gb, fx8320, gtx 500 ti, ssd 120gb. k-lite and visual studio instaled.
-
how to improve TXTimages, some characters are wrong, incomplete, missing some parts. Check attachment. What is the suitable setting to improve ? Any other easy method to create FRDimages? nissanjp@139.com