DCP-o-matic

Hi folks, new here. Been making dcps since 2011. I never saw this problem:
A dcp encoded with resolve/kakadu and packed with dcp o matic will load fine on a server but will be skipped completely on playback.
Here are the videos the projectionist sent me. I think its a dolby server but could ask him specific questions.
gui:
https://drive.google.com/file/d/17Aksge ... sp=sharing
cinema screen:
https://drive.google.com/file/d/1QcUFAe ... sp=sharing

From what can be seen on the cinema screen, it is most likely a j2k encoding issue, but still I though I´d post it here where people know their stuff
Otherwise the only thing that comes to mind is a frame rate mismatch. It is a 25cps dcp. Could they be trying to play it back at 24 and this be the problem?
Thanks a lot in advance!
Pedro

After unwraping, it looks that the files are flagged as 4k but are 1998x1080. I found this info loading the extracted j2k files into easy dcp creator.
More coming soon.

I never thought about this before - but as DCP-o-matic is able to pass through J2K - what if DCP-o-matic is set to 4k, while the source footage from Resolve is actually 2k? Will DCP-o-matic recognize this issue and recompress automatically?

Maybe this also an issue with the free Resolve version - it may not allow 4 export?


Can you try to load the falsely flagged MXF into DCP-o-matic and right-click properties?

After further investigation, it was not resolve, but the fnord after effects plug-in what was really used for encoding. In its settings, you can set the j2k profile to 4k but output a 2k sequence. This will create the faulty files. It is a pity that both dcp o matic and open dcp will allow this sequence to be used, while easy dcp will reject it. Maybe there is something that can be done to detect this error files and show a warning.
Thanks!

Can you try to load such a faulty DCP into DCP-o-matic player 2.15.x (current test version) and perform the DCP integrity check?

Yes, maybe when source content is meant to be passed trough, DCP-o-matic should perform some basic tests and give some hints.

However, full checking is not always possible. In the case of e.g. a VF added to an OV, one simply has to assume that the OV is technically correct.


Maybe DCP-o-matic should issue a warning always when pre-encoded content is about to be passed through. We had loads of problems with people assuming that any J2K was DCI compliant. Many times DCPs created with J2Ks from Resolve failed. For pass-through, DCP-o-matic should warn the user that it is his responsibility to make sure that the content to be passed through is fully DCI compliant.

- Carsten

Carsten wrote: ↑Mon Jan 31, 2022 10:25 pm Maybe this also an issue with the free Resolve version - it may not allow 4 export?

The free version tops out at UHD resolution, so it won't do 4k DCPs.

One catch with Resolve is that it has a regular JPEG2000 image sequence export in addition to the full 'DCP' export. However, I've never had any issues with DOM using that output sequence.

It seems replies are coming out of order, it was not resolve, but an fnord encode with the wrong settings (4k profile in a 2k sequence):
After further investigation, it was not resolve, but the fnord after effects plug-in what was really used for encoding. In its settings, you can set the j2k profile to 4k but output a 2k sequence. This will create the faulty files. It is a pity that both dcp o matic and open dcp will allow this sequence to be used, while easy dcp will reject it. Maybe there is something that can be done to detect this error files and show a warning.
Thanks!

If you could make a short file which demonstrates this problem it would be really helpful - then I could add a check to DoM.

Ok, here is a 2 second secuence, let me know if you need anything else or more frames:
https://drive.google.com/file/d/1a7oDSj ... sp=sharing

Thanks. Making a DCP with those, and verifying the DCP, does give an error. But I'll see what can be done about checking the files on the way in.