View Full Version : MKVToolNix v99.0 released
leon
15th April 2020, 08:10
@Mosu
OK, thanks. So you'll update to the next version of Qt as soon as it's available, right?
If you look at the link you posted you can see the issue was closed with the release of 5.14.2.
I don't know how that's relevant to the version of Qt used in MKVToolNix, but whatever.
ryrynz
15th April 2020, 08:12
@Mosu
OK, thanks. So you'll update to the next version of Qt as soon as it's available, right?
Mosu just said it uses the latest.. There nothing to fix...
It's relevant if you paid attention to a) what I said
b) the link you provided and lastly c) what Mosu said!
leon
15th April 2020, 08:22
You certainly don't know what you're talking about. I suggest you read the previous posts (2 pages before this). I perfectly understand what Mosu says, thank you.
Mosu
15th April 2020, 19:06
Actually, the bug says that it was too late for 5.14.2 and that it was fixed in 5.14.3. 5.14.3 hasn't been released yet.
I usually update my mingw cross-compilation environment once a month, normally within the first weekend of the month. You can guesstimate from that information which MKVToolNix release will use 5.14.3 once that'll have been released.
mkver
16th April 2020, 00:45
That's definitely a good point. Announcing such a change in advance is indeed a good idea, something I've done in the past with other changes as well (e.g. the deprecation & later removal of options such as the verbose identification mode). It isn't something I'm willing to work on right now, though I'll probably revisit the topic early next year.
Any update on that?
Mosu
16th April 2020, 17:54
No, nothing new on that front. I haven't thought about the topic at all the past couple of months, to be honest. It doesn't have a high priority for me.
AYColumbia
19th April 2020, 19:59
Hi Mosu,
I noticed that mkvmerge.exe does NOT exit with an error code when it fails that would set the errorlevel used in batch files. Would you please consider returning an exit code other than 0 on any failure please just so I know the process failed? For now, I just work around it by checking if my output file exists. Thank you.
stax76
19th April 2020, 20:06
Post command line, output and media info (https://github.com/stax76/MediaInfo.NET).
https://mkvtoolnix.download/doc/mkvmerge.html#mkvmerge.exit_codes
Mosu
20th April 2020, 15:10
Hi Mosu,
I noticed that mkvmerge.exe does NOT exit with an error code when it fails that would set the errorlevel used in batch files. Would you please consider returning an exit code other than 0 on any failure please just so I know the process failed? For now, I just work around it by checking if my output file exists. Thank you.
mkvmerge should always exit with 2 in case of errors and 1 in case of warnings. As stax76 said, please post the command line used as well as the full output generated by mkvmerge.
Perenista
21st April 2020, 21:19
Is there a way to force MKVToolnix to stop saving a modified file this way?
Let's say original MKV is MOSU.mkv
I change a few things and save a new copy. MKVToolnix says it will save as MOSU (1).mkv
Actually I don't want that. I want it to replace MOSU.mkv and the new file to have that same original name.
If not possible during saving, deleting the original MOSU.mkv file after the completion and saving as I said above.
*****
Also, is there a way to force the file (using MKVToolnix) to select a certain subtitle track if a certain audio track is also selected? Important: not the other way around, so selecting the subtitle first would not change to the desired audio track. Or this is always manually chosen by the user, during playback?
Mosu
22nd April 2020, 08:53
Is there a way to force MKVToolnix to stop saving a modified file this way?
In the preferences disable the checkbox that says to make the destination file name unique.
Also, is there a way to force the file (using MKVToolnix) to select a certain subtitle track if a certain audio track is also selected?
No.
@Mosu
Would you please help me understand how to decode CodecPrivate data field (codec_private_data) reported by mkvmerge (I'm comparing it to the headers traced by ffmpeg and there seems to be some extra chunks of information that ffmpeg headers don't have)? A more detailed description of the issue is in this post (https://forum.doom9.org/showthread.php?t=181316)
Thank you!
Mosu
23rd April 2020, 18:48
I cannot help you with your original goal (encoding something so that the codec private data matches). The content of CodecPrivate for H.265/HEVC uses the same structure is the same as the hvcC atom in corresponding MP4 files. Its structure is described in the usual ISO/IEC documents.
Stereodude
25th April 2020, 20:39
Using the "Remove dialog normalization gain" option on an E-AC-3 audio track that contains Atmos during muxing results in an audio stream that is no longer decoded as Atmos, just DD+.
MediaInfo reports basically the same information for the audio stream before and after, but my Denon doesn't see the after version as Atmos. The volume offset is gone though.
before:Audio
ID : 2
Format : E-AC-3 JOC
Format/Info : Enhanced AC-3 with Joint Object Coding
Commercial name : Dolby Digital Plus with Dolby Atmos
Codec ID : A_EAC3
Duration : 1 h 57 min
Bit rate mode : Constant
Bit rate : 768 kb/s
Channel(s) : 6 channels
Channel layout : L R C LFE Ls Rs
Sampling rate : 48.0 kHz
Frame rate : 31.250 FPS (1536 SPF)
Compression mode : Lossy
Stream size : 645 MiB (14%)
Language : English
Service kind : Complete Main
Default : Yes
Forced : No
Complexity index : 16
Number of dynamic objects : 15
Bed channel count : 1 channel
Bed channel configuration : LFE
Duration_Source : General_Duration
After:ID : 2
Format : E-AC-3 JOC
Format/Info : Enhanced AC-3 with Joint Object Coding
Commercial name : Dolby Digital Plus with Dolby Atmos
Codec ID : A_EAC3
Duration : 1 h 57 min
Bit rate mode : Constant
Bit rate : 768 kb/s
Channel(s) : 6 channels
Channel layout : L R C LFE Ls Rs
Sampling rate : 48.0 kHz
Frame rate : 31.250 FPS (1536 SPF)
Compression mode : Lossy
Stream size : 645 MiB (14%)
Language : English
Service kind : Complete Main
Default : Yes
Forced : No
Complexity index : 16
Number of dynamic objects : 15
Bed channel count : 1 channel
Bed channel configuration : LFE
hubblec4
26th April 2020, 12:22
Hi Sterodude
Could you open an issue on GitLab (https://gitlab.com/mbunkus/mkvtoolnix/issues?scope=all&utf8=%E2%9C%93&state=opened) with a sample file please?
If you have no GitLab acc I can do it for you if you provide a sample of your Dolby Atmos file.
Mosu
26th April 2020, 12:43
You can open an issue on GitLab, that would probably be best, but this isn't something I'm going to work on any time soon if ever. Just FYI.
Perenista
26th April 2020, 15:58
What is the difference between "Determining automatically" and YES for DEFAULT TRACK FLAG?
Example: let's say I have a movie with 2 audio tracks: english and portuguese.
I always want this file to use as default track the english. Note this is not FORCED, it is my preferred track, to be played whenever I open this file (except, of course, if we instruct Media Player Classic to ignore this and select other language instead).
>>>>>>>>>>>>>>>>>
OK, I can configure YES for english and NO for portuguese, while editing the MKV.
This is what I mean:
https://i.imgur.com/0SXIL69.png https://i.imgur.com/Fw67Qny.png https://i.imgur.com/IEdSxhc.png
But if I come back later and want to edit some other stuff in this MKV, I always find it like this:
https://i.imgur.com/rvU5emI.png
Didn't I say NO to that setting after editing? Why MKVToolnix keeps changing to "determine automatically" whenever it opens again? (and this happens with subtitles, too)
sneaker_ger
26th April 2020, 16:45
https://gitlab.com/mbunkus/mkvtoolnix/-/wikis/Default-and-forced-flags-and-default-yes-no-in-the-GUI
markfilipak
29th April 2020, 14:00
Hey Mosu, I hope you're well.
I have a problem. I have a soft telecined movie. There are 6 VOBs. I need to concatenate them. I have tried these 3 command lines:
#1: "C:\Program Files\MKVToolNix\mkvmerge.exe" -o "OUT.VOB" H:\VIDEO_TS\VTS_01_1.VOB
#2: "C:\Program Files\MKVToolNix\mkvmerge.exe" --default-duration 0:24000/1001fps -o "OUT.VOB" H:\VIDEO_TS\VTS_01_1.VOB
#3: "C:\Program Files\MKVToolNix\mkvmerge.exe" --default-duration 0:30000/1001fps -o "OUT.VOB" H:\VIDEO_TS\VTS_01_1.VOB
In every case, each OUT.VOB appears combed when played back -- actually, they look like they were combed but then slightly filtered [Note].
[Note} mkvmerge is pretty quick, too quick to be transcoding (though OUT.VOB looks transcoded).
It's important that OUT.VOB be simply concatenated because I'm following the concatenation with ffmpeg, 24-to-60fps trancoding via 55-telecine and to do that I need clean 24 FPS.
Is there a mkvmerge command line that will solely concatenate the source VOBs without changing them?
Regards,
Mark Filipak.
More Info:
I use MPV to check the video at each step. MPV, Shift-I displays
FPS: xx.xxx (specified) xx.xxx (estimated)
where estimated is calculated on-the-fly from the PTSs of the preceeding 10 frames.
Source VOB: FPS: 29.970 (specified) 23.976 (estimated)
OUT.VOB #1: FPS: 59.940 (specified) 23.976 (estimated)
OUT.VOB #2: FPS: 23.976 (specified) 23.976 (estimated)
OUT.VOB #3: FPS: 29.970 (specified) 29.970 (estimated)
ffmpeg input VOB needs to be: FPS: 29.970 (specified) 23.976 (estimated)
videoh
29th April 2020, 14:08
Use copy /b in a cmd window. You don't need mkvmerge for this.
copy /b a.vob + b.vob + c.vob combined.vob
https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/copy
markfilipak
29th April 2020, 14:27
Use copy /b in a cmd window. You don't need mkvmerge for this.
copy /b a.vob + b.vob + c.vob combined.vob
https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/copy
The result has PTS errors and the ffmpeg transcode malfunctions.
Mosu
29th April 2020, 19:27
mkvmerge can only create Matroska files, not MPEG program streams (that's what VOBs are).
videoh
29th April 2020, 20:41
The result has PTS errors and the ffmpeg transcode malfunctions. Then you didn't do it right. :p
Maybe try knitting? Or miniature golf?
markfilipak
29th April 2020, 21:10
mkvmerge can only create Matroska files, not MPEG program streams (that's what VOBs are).
An MKV is fine. Actually, MKV is preferred. I'm sorry if I implied otherwise.
Do you have any guidance regarding the combing? I don't understand how combing can happen if mkvmerge is solely performing remuxing, but then, I don't understand a lot of things. ;-)
PS: I checked things and ...you know what? I did name the files '.VOB'. But you know what? I could have named them '.EXE'. It doesn't matter to MKV. It plays anything that it finds contains a video stream.
markfilipak
29th April 2020, 21:28
Then you didn't do it right. :p
You don't understand. ffmpeg has a bug that mishandles gaps in PTSs and/or DTSs that aren't monotonic. 'copy /b' doesn't fix that, of course. mkvmerge (and HandBrake, too) do fix the problem somehow. Remuxing seems to work magic sometimes. Unfortunately, mkvmerge appears to create a different problem: Combing. The combing is what I seek to avoid. I would use HandBrake, but HandBrake insists on transcoding and I don't want transcoding.
I hope it's clear now.
Maybe try knitting? Or miniature golf?
What? Do you want me to play with myself?
Mosu
29th April 2020, 22:17
Do you have any guidance regarding the combing? I don't understand how combing can happen if mkvmerge is solely performing remuxing, but then, I don't understand a lot of things. ;-)
I don't support MPEG program stream/MPEG elementary stream handling in mkvmerge anymore. That code was contributed by someone else who hasn't been around for more than a decade now, and I have no interest in fixing issues in that code.
As I've never dealt with telecined content I cannot give you any advice whatsoever.
PS: I checked things and ...you know what? I did name the files '.VOB'. But you know what? I could have named them '.EXE'. It doesn't matter to MKV. It plays anything that it finds contains a video stream.
mkvmerge generally determines the file type by the content. There are certain cases where it uses the file's extensions in situations where content-based detection is ambiguous (e.g. not detecting an MPEG transport stream with H.264 inside as an H.264 elementary stream if the file's extension is ".ts" or ".m2ts").
videoh
29th April 2020, 22:41
I hope it's clear now. Nah. You asked how to just concatenate VOB files and I told you. You wrote:
It's important that OUT.VOB be simply concatenated
I never claimed it would fix any ffmpeg problems. Best bet for you is to post a sample source stream(s) and tell us how to duplicate your issue. Don't make all kinds of assumptions and then start blaming people if they don't necessarily accept them.
What? Do you want me to play with myself? Divine energy will flow through you, if you allow it.
markfilipak
30th April 2020, 07:57
I never claimed it would fix any ffmpeg problems.
There is no blame to assign, my friend. There is no blame at all.
What I have is VOBs that provoke time stamp mishandling by ffmpeg. I have found that remuxing via mkvmerge fixes those problems and that ffmpeg then transcodes the remuxed videos correctly. So, of course, I want to use mkvmerge as a preprocessor in scripts.
The fix should logically be limited to fixing the time stamps in elemental streams, but mkvmerge's output appears to be more than remuxed, it appears to be transcoded because the frames are no longer pristine, they're combed. I don't know what to make of that and I don't have the tools that would reveal what's actually happening in raw frames.
So, I come here to find out becasue you folks are informed. If I'm making assumptions, I appologize. It's because I'm somewhat ignorant. If you have any suggestions, I'd be very grateful.
Stay safe, stay healthy,
Mark.
videoh
30th April 2020, 09:06
If you have any suggestions, I'd be very grateful. As I said, best bet for you is to post a sample source stream(s) and tell us exactly how to duplicate your issue. We don't have magic wands or clairvoyance.
mkvmerge is not transcoding.
sneaker_ger
30th April 2020, 09:13
As you say mkvmerge is not re-encoding. If it was combed after using mkvmerge then if was also combed before. Probably your player activated de-interlacing/IVTC for the VOB file without you knowing thus appearing as if it was progressive. After remuxing to mkv your player isn't detecting this anymore.
videoh
30th April 2020, 11:06
Quite right, sneaker_ger. A small sample cut from the VOB would allow us to verify that. OP should cut a section that comes out combed in the MKV. Set range in DGIndex and then Save Project and Demux Video. Post a link to the resulting M2V file.
AYColumbia
30th April 2020, 15:12
Post command line, output and media info (https://github.com/stax76/MediaInfo.NET).
https://mkvtoolnix.download/doc/mkvmerge.html#mkvmerge.exit_codes
mkvmerge should always exit with 2 in case of errors and 1 in case of warnings. As stax76 said, please post the command line used as well as the full output generated by mkvmerge.
Thanks I'll provide that info the next time it happens. I had not saved the prior info and I should have provided it. My bad.
Hey everyone!
It's May. It's time. Time for another MKVToolNix release.
In v45 I accidentally introduced two bugs that users stumbled across almost immediately, and I nearly released a bug fix release only a few days after the v45 release, but ultimately decided against it. Therefore I suggest you upgrade to v46 now, especially if you're dealing with H.265/HEVC video in MP4 files.
Nothing's changed for package managers.
Here are the usual links: the MKVToolNix home page (https://mkvtoolnix.download/), the Windows installer/portable version & macOS DMG & Linux AppImage (https://www.fosshub.com/MKVToolNix.html) and the source code (https://mkvtoolnix.download/source.html).
The Windows and macOS binaries as well as the Linux AppImage are available already. The other Linux binaries are still being built and will be available over the course of the next couple of hours.
Here are the NEWS (https://mkvtoolnix.download/doc/NEWS.md) since the previous release:
Version 46.0.0 "No Deeper Escape" 2020-05-01
New features and enhancements
mkvmerge: when splitting, the placeholder "%c" can be used in the destination file name. It'll be replaced by the name of the first chapter in the file. Implements #2791 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2791).
mkvmerge: when reading MPLS playlists mkvmerge will include a tag named "SOURCE_ID" in the track's statistics tags that conveys the fact that the source was a Blu-ray and what the track's ID was in the source container. When reading Matroska file existing "SOURCE_ID" tags will be kept. The format used is the same format MakeMKV uses. Implements #2774 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2774).
mkvmerge: for audio-only files mkvmerge will now write a cue entry every 500ms instead of every 2s.
MKVToolNix GUI: job actions: split up the option "execute when the job finishes successfully or with warnings" into two separate options: "…when finishes successfully" and "…when exits with warnings". Implements #2798 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2798).
Bug fixes
mkvmerge: Matroska reader: when regenerating UIDs for chapters mkvmerge will now fix referencing tag chapter UID targets to have the same values. Fixes #2804 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2804).
mkvmerge: MP4 reader: fixed a crash that could potentially happen while trying to identify H.265/HEVC tracks.
mkvmerge: MP4 reader: when reading H.265/HEVC tracks without a frame offset table ("ctts" atom) present, mkvmerge did not parse the "hevcC" structure correctly in certain cases, causing no video frames to be output at all or the output not being playable. This fixes #2782 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2782), a bug introduced in v45.
mkvmerge: MPLS handling/MPEG TS reader: timestamp restrictions as given in an MPLS file were only honored for the first play item, not for subsequent play items. For those play items the whole referenced file was always copied. Part of the fix of #2789 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2789).
mkvmerge: MPLS handling/MPEG TS reader: under certain circumstances mkvmerge copied one frame too many before switching to the next play item. Part of the fix of #2789 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2789).
mkvmerge: for audio-only files cues will only be created for I frames (important for e.g. TrueHD where decoding can only start on a sync frame). Fixes #2790 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2790).
MKVToolNix GUI: fixed a crash that happened when closing the preferences after having closed at least one multiplexer tab. Fixes #2785 (https://gitlab.com/mbunkus/mkvtoolnix/issues/2785).
MKVToolNix GUI: macOS: changed the default of the "elide tab header labels" feature introduced in v45 from "no" to "yes" to mirror how versions before v45 behaved on macOS.
Have fun!
MrVideo
1st May 2020, 23:01
Tbut mkvmerge's output appears to be more than remuxed, it appears to be transcoded because the frames are no longer pristine, they're combed.
I'm jumping in here a little late, so excuse me this has been discussed with you in previous posts. But, do you understand what a combed video frame is?
As noted, MKVmerge does not do any recoding/transcoding, it only copies source material.
A combed video frame is a result of two different fields in a 480i, 576i or 1080i video file. This can be because the video was shot with a 480i/576i/1080i video camera, or a 23.976 progressive video source was telecined to 29.97, where two video frames, out of five, can have combed video frames.
For a true video source, you have to live with the combing, or recode using a deinterlacing function, which combines the two fields into a single progressive frame. That ultimately results in a loss of time space. By this I mean if the two fields have rapid motion, that motion appears to move at a pace of 59.94 images a second. But, when you deinterlace, the time is now has a motion of 29.97 images a second, or time/space is now 1/2 of what it originally was. I personally prefer to keep the full motion intact. The process of deinterlacing can also reduce the vertical resolution.
If the 29.97 interlaced video is a result of the telecine process to convert 23.976 source material to 29.97, the IVTC (inverse telecine) process can restore the interlaced 29.97 video back to 23.976 progressive video, removing the combing effect.
If you already knew all this... then NEVERMIND!
markfilipak
3rd May 2020, 02:00
I'm jumping in here a little late, so excuse me this has been discussed with you in previous posts. But, do you understand what a combed video frame is?!
"progressive video source [that] was telecined"
A combed frame from telecine mixes 2 half-pictures that are separated in time by 1/24 second, and as you cited, a 23-telecine has 2 of them in every 5 frames (40%), and they are back-to-back: the 3rd & 4th frames.
"shot with a 480i/576i/1080i video camera"
Each combed frame in that case results in fields separated in time by 1/60 second -- less than half the combing -- but every frame is combed (100%).
Both are awful.
As for my project, the only thing I can figure is that I must have somhow grabbed a hard telecined video to use for testing. That seems a year ago, though it was just yesterday or the day before, and I've since used many test videos and discarded some. I've progressed past that. I'm still using mkvmerge and in fact am using it for 2 of the 4 processing steps, but I'm using it in a slightly different way. In other words, I got no problems with mkvmerge. Not any more.
MrVideo
3rd May 2020, 12:26
""shot with a 480i/576i/1080i video camera"
Each combed frame in that case results in fields separated in time by 1/60 second -- less than half the combing -- but every frame is combed (100%).
Yes and no. If the two fields have an imperceptible time difference, it will appear as though the frame was a progressive frame. There will be no combing effect visible.
Just think, before digital TV came along, no one really complained about interlaced video, no matter the source. We've become spoiled.
darksen
4th May 2020, 04:57
Is there any way to make MKVToolNix use a separate path to create the files and then move the finished files to the destination? I mean, a temp working folder?
This to prevent fragmentation as much as possible as it is impossible for MKVToolNix to pre-allocate the space beforehand.
Also, what does --engage-cow do?
MKVToolNix GUI: job actions: split up the option "execute when the job finishes successfully or with warnings" into two separate options: "…when finishes successfully" and "…when exits with warnings". Implements #2798.
Thank you! I was going to request this.
Is there any way to make MKVToolNix use a separate path to create the files and then move the finished files to the destination? I mean, a temp working folder?
Nope. No plans for it either.
Also, what does --engage-cow do?
It's "--engage cow", and just give it a try and you'll see.
Seeing that with exhale (https://gitlab.com/ecodis/exhale/) there's a xHE-AAC encoder out there I was wondering if there are plans to support xHE-AAC in mkvtoolnix?
There are no plans so far. I don't know what work (if any) would be required.
Nejiro
5th May 2020, 13:17
Greetings to all, I am new to this forum and I apologize if I use a translator, I use MKVToolNix on Ubuntu 20.04 lts, I have some 1: 1 bluray in which I removed the various languages that I do not need by transforming the m2ts file into mkv, all it works well but I noticed with mediainfo that the bitrate of the video on the mkv file is lower than the m2ts, so I lose quality on the mkv file compared to the m2ts? Or does mkv remove things that are not needed without me losing quality?
Thanks
videoh
5th May 2020, 13:38
Overhead for an M2TS is higher than for MKV. The actual video is not affected in any way.
Nejiro
5th May 2020, 14:17
Perfect, thank you very much for your reply, so with an mkv file you won't lose anything, that's what I wanted to hear ......
Nejiro
5th May 2020, 15:32
Excuse me another question, when I create mkv files with Handbrake I have the possibility to insert in appropriate spaces plot of the film, director, year etc ... with MKVToolNix is it possible to do it? I can't find where to insert this data .....
Thanks
videoh
5th May 2020, 15:46
One possibility:
https://forum.videohelp.com/threads/384585-Matroska-text-file-attachment
MKV gurus here may know more about it.
And, oh, welcome to the forum!
hubblec4
5th May 2020, 17:36
Excuse me another question, when I create mkv files with Handbrake I have the possibility to insert in appropriate spaces plot of the film, director, year etc ... with MKVToolNix is it possible to do it? I can't find where to insert this data .....
Thanks
Such things could be preserved with Matroska Tags.
MKVToolNix has no Tags editor you could try chapterEditor for this task.
Nejiro
5th May 2020, 18:54
I imagined there wasn't a tag editor like on Handbrake, creating an html file I wouldn't know how ...
Thanks
Nejiro
6th May 2020, 11:38
Excuse me another question, I have an mkv file in which an audio track starts about 5 seconds behind the video, how can I synchronize them? I select the audio track and in the audio delay field I enter 5,000? I don't know how it should be inserted exactly the value 5000 or 5.000? If instead it is the opposite and that is that the audio is in advance I have to put the minus sign in front, that is -5000?
Thanks
The "delay/sync" functionality simply adds the number of milliseconds given in the box to all timestamps. So if audio is 5s behind your video you have to subtract 5s, which means: add "-5000" ms. So you enter "-5000".
Liisachan
6th May 2020, 21:11
insert in appropriate spaces plot of the film, director, year etc ... with MKVToolNix is it possible to do it? Like others say, it's easy once you know how to tag MKV files. At first, you may feel XML files are difficult, but once you create your "template", you can just reuse the same file again and again, with just editing the tag values.
There is, however, one subtle, tricky aspect...
MPC and MPC-HC parse DESCRIPTION and show it as Description in Information (Ctrl+3).
MPC-BE parses the last COMMENT and shows it as Description.
Because of this, if you'd like to get a consistent result (for both HC and BE), set the last COMMENT identical to the global DESCRIPTION tag (as a workaround). If you only use a global tag file, this simply means COMMENT and DESCRIPTION have the same value in it. If you use par-track tag file(s) too, the COMMENT of the last tag file matters: e.g. if you have not only global.xml but also video.xml, audio.xml, and sub.xml and the track order is video audio sub, then COMMENT in sub.xml matters.
If you don't mind such minor detail (HC/BE compatibility) you can safely ignore this :) It's just a (weird) workaround for specific players, not at all required in the matroska specs.
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.