ccextractor/docs/HARDSUBX.txt
2021-06-08 17:57:22 +00:00

61 lines
1.8 KiB
Plaintext

Overview
========
Subtitles which are burned into the video (or hard subbed) can be extracted using the -hardsubx flag.
The system works by processing video frames and extracting only the subtitles from them, followed
by an OCR recognition using Tesseract.
Dependencies
============
Tesseract (OCR library by Google)
Leptonica (C Image processing library)
FFMpeg (Video Processing Library)
Compilation
===========
Linux
-----
Make sure Tesseract, Leptonica and FFMPeg are installed, and that their libraries can be found using pkg-config.
Refer to OCR.txt for installation details.
FFmpeg from packages (on Debian) plus a couple of other dependencies you will need:
sudo apt-get install libavcodec-dev libavformat-dev libavutil-dev libswscale-dev libxcb-shm0-dev liblzma-dev
FFmpeg from source:
To install FFmpeg (libav), follow the steps at:-
https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu - For Ubuntu, Debian and Linux Mint
https://trac.ffmpeg.org/wiki/CompilationGuide/Generic - For generic Linux compilation
To validate your FFMpeg installation, make sure you can run the following commands on your terminal:-
pkg-config --cflags libavcodec
pkg-config --cflags libavformat
pkg-config --cflags libavutil
pkg-config --cflags libswscale
pkg-config --libs libavcodec
pkg-config --libs libavformat
pkg-config --libs libavutil
pkg-config --libs libswscale
On success, you should see the correct include directory path and the linker flags.
To build the program with hardsubx support,
== from the Linux directory run:-
./configure --enable-hardsubx
make ENABLE_HARDSUBX=yes
== using cmake from root directory
mkdir build
cd build
cmake -DWITH_OCR=on -DWITH_HARDSUBX=on ../src/
make
NOTE: The build has been tested with FFMpeg version 3.1.0, and Tesseract 3.04.
Windows
-------
Coming Soon