Compliance of MKV files generated by MakeMKV

[xaviour]

The Matroska Foundation provides a couples of command line tools to check compliance of MKV files ("mkvalidator") and optimize them ("mkclean"). When used on files
created by makemkv, "mkvalidator" outputs a lot of warnings and error. For example:

Code: Select all

ERR202: Unique element Position in Cluster at 1323 found more than once
ERR202: Unique element Position in Cluster at 1323 found more than once
ERR202: Unique element Position in Cluster at 53826 found more than once
[...]
WRN103: Unnecessary secondary SeekHead was found at 137152604
[...]
ERR0A2: The Cluster position 1299 at 1335 should be 1287
ERR0A1: The Cluster PrevSize 52494 at 53843 should be 52503

Fortunately most of the times using "mkclean" with the "--remux" option fixes it so I did not see fit to report the problem.

The bad news is that for a few files generated from DVD, "mkclean" goes into an infinite loop during the remuxing phase before the reclustering phase. Obviously, an infinite loop is a bug and "mkclean" should be fixed. The problem is that "mkclean" is meant to optimize compliant MKV files and not to repair broken MKV files. Dealing with all possible type of corrupted files would have a prohibitive cost so I doubt that the Matroska people will bother with it. At best the tool will bail out leaving me with a broken source file.

Usually such files are too big to be uploaded but I now have a 130 MB file coming directly from makemkv and I am willing to make it available to the makemkv developers in the hope that they will fix the problem.

[Romansh]

http://www.makemkv.com/forum2/viewtopic.php?f=6&t=3371

[mike admin]

Believe it or not, that slipped without being noticed. Thanks for pointing out - should be quite easy to fix.

[xaviour]

Believe it or not, that slipped without being noticed.

Yes I do believe it. Quality control is hard. Fortunately a command line program like mkvalidator is easy
to integrate in automated regression testing. A flow using an ISO file plus makemkvcon plus mkvalidator is easy to script.

Thanks for pointing out - should be quite easy to fix.

I guess this means you have been able to reproduce it yourself so I do not need to upload any file?

[xaviour]

Obviously the problem is still present in 1.6.6.

[xaviour]

The problem is still present with 1.6.7.

[xaviour]

The problem is still present with 1.6.8.

[xaviour]

The problem seams to be solved in 1.6.9.

[mike admin]

Should be in 1.6.10 . mkvalidator would still print bogus errors on certain audio and subtitle tracks - but that's a bug in mkvalidator.

[xaviour]

mkvalidator would still print bogus errors on certain audio and subtitle tracks - but that's a bug in mkvalidator.

Care to elaborate about the bug in mkvalidator? I could not find a bug report either in the sourceforge instance of Matroska nor in the various mailling lists.

[xaviour]

The problem seams to be solved in 1.6.10.

[mike admin]

mkvalidator would still print bogus errors on certain audio and subtitle tracks - but that's a bug in mkvalidator.

Care to elaborate about the bug in mkvalidator? I could not find a bug report either in the sourceforge instance of Matroska nor in the various mailling lists.

mkvalidator would print an error that certain audo and subtitle tracks have non-keyframes. For PGS and TureHD that's the way it is there are both key and non-key frames in the stream.

[xaviour]

For PGS and TureHD that's the way it is there are both key and non-key frames in the stream.

And are the author of mkvalidator aware of this problem?

[vidsquibber]

That's a good question, you should think with the amount of feathers it has ruffled over here somebody would have mentioned it. how do we get in touch?

[xaviour]

how do we get in touch?

The best way is to use the development mailling list linked on the contact page.
Writing a technically sound message linking to a small sample video that exhibit the issue is the best way to proceed.