MXF parse utility?

[HaroldHallikainen]

I'm analyzing some Japanese subtitles and need to see the font file buried in the MXF files. Is there a suggested utility to extract files from the MXF?

Thanks!

Harold

[carl]

Hi Harold,
Maybe asdcp-unwrap from asdcplib - I think that might do it... it certainly will extract the MXF; not sure about the font. Let me know if it doesn't work out!
Best,
Carl

[HaroldHallikainen]

Thanks! I see you have Windows binaries at https://carlh.net/asdcplib .

It appears the DCP has three mxf files. asdcp-info says that one of them is timed text, but asdcp-unwrap gives the following error message:

Program stopped on error.
File open failure.

However, it DOES indeed extract the timed text XML with no file extension. I'm guessing the font file is in that MXF, but do not have a way to get it out thus far.

Dcpomatic Player starts to load the DCP but then complains about Hgroup (in the timed text file): Unexpected node Hgroup.

THANKS!

Harold

[Carsten]

What if you load the subtitled SMPTE DCP or MXF subtitle into DCP-o-matic and simply convert it to Interop, then look into the subtitle folder?

However, if the player fails on this DCP, it is likely that DCP-o-matic main will fail as well. But, Harold, you may still try.

But maybe the trouble is within the subtitle MXF, and if that is broken or non-standard for some reason, it could become hard to get to the ground. One reason I still prefer the InterOPEN DCP. Do you know the application that created this DCP from the headers?

Would John Hurst be on HAM?

- Carsten

[HaroldHallikainen]

Thanks Carsten!

Adding the DCP to a DCPOMATIC project results in the same error message (Unexpected node: Hgroup). It then quits the import.

Looking at the CPL, I see it was created using ClipsterDCI 6.7.1.0.

This is a test DCP we are using to look at how different subtitle renderers handle subtitles, especially Japanese subtitles as part of a SMPTE project to update ST 428-7. I've sent an email to the author to see if he can send me the compressed font file. I was hoping I could pull it from the DCP, but don't know how to do it!

Yes, John is an amateur radio operator. We've talked about ham radio on some of my visits to CineCert. I might follow up with him...

Thanks!

Harold
https://w6iwi.org

[Carsten]

So, Clipster creates a subtitle file that crashes asdcplib's unwrap? I guess John would be interested.

- Carsten

[Carsten]

BTW - Carl, there appears to be a bug when converting a subtitled SMPTE DCP to an IOP DCP (as I suggested trying to Harold):

When I load a subtitled SMPTE-DCP into DCP-o-matic and create a subtitled IOP DCP from it, it does unwrap the subtitle XML from the MXF, but misses to unwrap and apply the font file. Well, that alone is not yet a bug, but a potential missing feature. However, when creating the subtitled file/folder, the resulting XML subtitle file misses a <LoadFont Id="font" tag altogether. PKL then misses references to a font file as well. DCP-o-matic here also doesn't use the default LiberationSans. What's worse, I can not even add a specific font in the font menu - it is inactive, I can not click on 'Edit', it's greyed out. What DOES work however, is to disable the included subtitle track of the SMPTE DCP, and add just the MXF subtitle file separately. In that case, LiberationSans is used for the subtitles, and I can still point to a different font file.

It sure would be nice if imported/reused MXF subtitle files would also have their font's unwrapped and reused - especially for exotic or specific fonts (e.g. font compressed) No idea how complicated that would be.

See attached files.

Regards - Carsten

[carl]

Hi both,

It appears the DCP has three mxf files. asdcp-info says that one of them is timed text, but asdcp-unwrap gives the following error message:

Program stopped on error.
File open failure.

I think that "Program stopped on error." thing is a bug in asdcplib; like you say, it still seems to produce an output.

However, it DOES indeed extract the timed text XML with no file extension. I'm guessing the font file is in that MXF, but do not have a way to get it out thus far.

If it's just a one-off, feel free to send me the MXF file and I'll try to extract the font for you.

Dcpomatic Player starts to load the DCP but then complains about Hgroup (in the timed text file): Unexpected node Hgroup.

That's odd - can you send me the XML? carl@dcpomatic.com

When I load a subtitled SMPTE-DCP into DCP-o-matic and create a subtitled IOP DCP from it, it does unwrap the subtitle XML from the MXF, but misses to unwrap and apply the font file. Well, that alone is not yet a bug, but a potential missing feature. However, when creating the subtitled file/folder, the resulting XML subtitle file misses a <LoadFont Id="font" tag altogether. PKL then misses references to a font file as well. DCP-o-matic here also doesn't use the default LiberationSans. What's worse, I can not even add a specific font in the font menu - it is inactive, I can not click on 'Edit', it's greyed out. What DOES work however, is to disable the included subtitle track of the SMPTE DCP, and add just the MXF subtitle file separately. In that case, LiberationSans is used for the subtitles, and I can still point to a different font file.

I see what you mean - I made a bug for that.

Thanks,
Carl