DVDRemaster and Modules

DVDRemaster uses modules to recompress video. The modules accept the MPEG2 video data (not VOB) from the standard input and write the recompressed frames to the standard output. Modules can easily be added, as long as they respect the following specifications:

• Data must be read from stdin and written on stdout.
• Each frame must be smaller than the original.
• I, P and B frame order, and GOP boundaries must be kept.
• Top field first and repeat first field flags must remain the same.
• Size and aspect ratio must also remain the same.
• VBV delay should be 0xFFFF.
• Two-field pictures can be merged into one-frame picture.

Modules will be given three arguments, the first one is a floating point value telling by how much the data size should be decreased (the ratio). The second argument is a number from 0 to 3 specifying the quality of the reencode - 0 being the worst (faster), 3 being the best (slowest). The third is the size of the M2V (the module cannot know otherwise since the data is streamed). For instance:
./M2VRequantiser 1.5 0 1024000 < original.m2v > recompressed.m2v
would recompress original.m2v, whose size is 1024000 bytes, by a factor of 1.5.

The two modules that come with DVDRemaster are based on libmpeg2 and mpeg2enc and are released under the GNU General Public License (GPL).