Question

It would be useful to be able to dump mkv file to some text (like in mkvinfo -v), edit it and reconstruct back, leaving binary data as hex dumps, for example.

Something like:

$ mkv2xml < test.mkv > test.xml
$ head test.xml
<EBML>
    <EBMLVersion>1</EBMLVersion>
    <MaxIdLen>4</<MaxIdLen>
    <MaxSizeLen>8</MaxSizeLen>
    <DocType>matroska</DocType>
    <DocTypeVersion>2</DocTypeVersion>
    <DocTypeReadVersion>2</DocTypeReadVersion>
    <Segment>
        <Tracks>
        <Track>
            <TrackNumber>1</TrackNumber>
            <CodecID>V_MJPEG</CodecID>
            <Video><PixelWidth>320</PixelWidth><PixelHeight>240</PixelHeight></Video>
        </Track>
        </Tracks>
        <Cluster>
            <Timecode>5.023</TimeCode>
            <SimpleBlock track="1" time="5.045">
                ffd8fffe001a80......
$ xml2mkv < test.xml | mplayer -cache 100 -

This can be useful to simplify writing [debugging] tools work with video without actually messing with muxing/demuxing details, just by looking at output.

Answer 1

Implemented myself in Python: https://github.com/vi/mkvparse

Can convert to xml, change something (with xml2/2xml, for example), convert back and play in one pipeline.

Example XML file: http://vi-server.org/pub/xml2mkv_test.xml

Trimmed (even more) snippet here:

<mkv2xml>
<EBML>
  <EBMLVersion>1</EBMLVersion>
  <DocType>matroska</DocType>
</EBML>
<Segment>
<Info>
  <TimecodeScale>1000000</TimecodeScale>
  <MuxingApp>libebml v1.0.0 + libmatroska v1.0.0</MuxingApp>
  <Duration>46665.0</Duration>
  <DateUTC>1282406803.0</DateUTC>
  <SegmentUID>9d516a0f927a12d286e1502d23d0fdb0</SegmentUID>
</Info>
<Tracks>
  <TrackEntry>
    <TrackNumber>1</TrackNumber>
    <TrackType>1</TrackType>
    <CodecID>V_MPEG4/ISO/AVC</CodecID>
    <CodecPrivate>
      014d401fffe10014274d401fa918080093600d418041adb0ad7bdf0101000428
      ce09c8
    </CodecPrivate>
    <Video>
      <PixelWidth>1024</PixelWidth>
      <PixelHeight>576</PixelHeight>
    </Video>
  </TrackEntry>
  <TrackEntry>
    <TrackNumber>2</TrackNumber>
    <TrackType>2</TrackType>
    <CodecID>A_AAC</CodecID>
    <Audio>
      <SamplingFrequency>48000.0</SamplingFrequency>
      <Channels>2</Channels>
    </Audio>
  </TrackEntry>
  <TrackEntry>
    <TrackNumber>3</TrackNumber>
    <CodecID>S_TEXT/UTF8</CodecID>
  </TrackEntry>
  ...
</Tracks>
<Cluster>
  <Timecode>0</Timecode>
  <SimpleBlock>
    <track>1</track>
    <timecode>0.0</timecode>
    <keyframe/>
    <data>
      00004e1925b82001ebd34d359f84a010001019b0f94a107609ec4dc0b6341175
      98899cbb3a0a9de562532111ed77f7876006014200856a5fc3c2ccdf6e10a385
      ...
    </data>
  </SimpleBlock>
  <SimpleBlock>
    <track>10</track>
    <timecode>0.009</timecode>
    <keyframe/>
    <data>
      01449ffe0b246a488f09c715768924b01aa2e9b2415aaa29e4696becea270d36
      a0d565bd24d5e6f09f3d4e000a8f192aa6aaa071f2ba14bd2fa14d633ad665ee
      72e92eadce23415fd6bb871941941b1bfd5b0e75ad6826755e61ef5483f93cd3
      82d61c779fa74b4acd62572b399ac3d9039900e94a5c542e4110dd29b65e2000
      220011a10c00000000b0a8540100880380
    </data>
    <data>
      0144d7902c180b04c28161a8482a160a048aa1209848423208cdeed6de7c6b7a
      .....
    </data>
    ...
  </SimpleBlock>
  <BlockGroup>
    <Block>
      <track>4</track>
      <timecode>0.042</timecode>
      <duration>3.375</duration>
      <data encoding="text"><![CDATA[A jobb oldaladon láthatod...
...tudod mit...]]></data>
    </Block>
    <BlockDuration>3375</BlockDuration>
  </BlockGroup>
  <BlockGroup>
    <Block>
      <track>5</track>
      <timecode>0.042</timecode>
      <duration>3.375</duration>
      <data encoding="text"><![CDATA[Auf der rechten Seite sieht man...|...rate mal...]]></data>
    </Block>
    <BlockDuration>3375</BlockDuration>
  </BlockGroup>
</Cluster>
</Segment>
</mkv2xml>