DCP-o-matic Player reliability?

[BoltonFilmFest]

How confident can you be on the playback in DCP-o-matic?

Last year at our film festival we QC'd all our DCP's in person at our cinema, some were provided to us directly by distributors, others from DCP houses, some we made ourselves using DCP-o-matic, some were made by filmmakers. Other than a few minor issues with volumes (which we just simply adjusted by adding some notes to the projectionist) we didn't have any playback failures. If it played in DCP-o-matic Player it always seemed to play at the cinema, that was our experience anyway. But more widely to the forum community can we be confident if a DCP plays in DCP-o-matic it is going to play in the cinema? Our DCP's are ingested into a TMS server, which manages the playlist, so this question is purely asking from a home based office perspective. We don't intend to use the DCP-o-matic player for anything other than to check the playback.

With thanks

Adrian

[Carsten]

No, successful playout in DCP-o-matic player will not guarantee that the DCP will play okay on a cinema server.

However, if you perform the full DCP verification in player, your chances are very high that the DCP will not cause issues.

Verfication in player and playout in player are two completely different things.

This could become an interesting thread though as to what can be done with DCP-o-matic and DCP-o-matic player in a festival environment.

[macuser]

I agree with Carsten, SW-Player's like Dcp-o-matic may play DCP's that even are not ingestable because of hash mismatch.

The festival I work for is playing DCP's since 2010. In my opinion the process of producing a DCP should end with a screening in a real cinema to confirm a 100% working DCP. It is in the responsibility of the manufacturer to deliver a working DCP. The festival is not a test house and doesn't have time for that.

When I started with DCP's there weren't too much information (and software) around. With the time I started to use "dcp_inspect" for basic testing. That line command tool does at least what the ingest process on a DCP-Server would do. A real good feature is the audio analysis that shows in a small graphics all the channels. It checks many other parameters that are very useful for a festival. Alongside I also used a license of easyDCP Player+ for some years.

Since 2020 I use DCP-o-matic Player and dcp_inspect for all DCP's inspection. The player is very useful to check subtitles as not all seem to apply the DCI naming convention. With the time I learnt how and what to check to find DCP's that could cause problems in a cinema. I am open to share my knowledge with other human beeings.

Kind regards
Max

[Carsten]

I would also like to have audio analysis/LEQ(m) in player. It would even be a useful feature in standalone verifier, especially for a festival. It would be helpful being able to drop a folder of DCPs onto the verifier, then get a nice detailed report of issues what may be worth looking at, including LEQ(m).

[jamiegau]

The website, https://admin.d-cine.net allows you to upload a DCP to do a full analysis based on clairmeta (Unencrypted of course)
Over clarimeta it also does:
Produces a graph of video bitrate over time
More importantly, a full LEQ(M) result and also over time graph.

I am in the middle of upgrading the interface to a new framework. Slow-going as I am so busy these days.

I was also considering adding this ability to create this analysis output in the cinema-catcher-app, though you would need a resonably powerful machine to do the processing.

Would people be interested in this?

Mainly because, I see a lot of poorly curated preshow volume control between content.
If, when a DCP comes into a cinema, it auto grabs them, download and produces a report with Leq(m) and other reports. Possibly fire a warning of audio is too loud. But also converts the DCPs to mp4s so you can simply log into the web interface and check the volume in the browser (Play new TLRs ADVs back and relative to other known content volumes you can pick up what may be too loud.

I thought these good ideas and was working towards them, but no one appeared interested. Such poor care to showmanship in the world.

James

[IoannisSyrogiannis]

It might be that people who would care for the service are not familiar with that, James. While your services for and your wisdom shared in the community is appreciated, many people arrive here, for instance, not having yet introduced themselves to the basics of cinema exhibition.
It's a good sign, when people turn up here or at film-tech, or projectionniste, etc. but that indicates as well rather few clues on where to go and get the info one needs for making deliverables for cinemas.

I suppose that by "Over clarimeta it also does:" you mean besides that, or on top of that. I haven't found any such features on clairmeta itself. It may indicate the overall bitrate of a file, but that's about it. (Useful, though.) No question about loudness measurements.

[jamiegau]

@ IoannisSyrogiannis
I am not talking bout people on here or Film-tech. I have spoken to some larger chains who have voiced concern on preshow issues.
They talk about it but then never take any meaningful steps to try to rectify it.
There was a working group at ISDCF to iscuss how extended meta-data could be used to help with issues like this. Wend nowhere, and tools like this were discussed. But in the end, even with the pain points indicated. No action was taken from my standpoint by all involved. I feel they thought adding features like a Leq(m) level into the extended meta-data was not useful enough and could cause more issues then it fixes. To a degree I agree as audio measurement is not a black and white decision. Context is important, and computer QC tools cannot effectively QC those issues.

It's why I focused on tools that allowed such decisions to be made, still by humans, but in a web browser and not a QC cinema. (Cheaper/faster)

Here is a link to a video that goes over the items discussed.
https://youtu.be/I76LPOpkvt0?si=VlC5alCILKyKhS8h&t=803

[Carsten]

I know many cinema operators who adjust preshow volumes, yet not based on LEQ analysis, but they listen through the preshow when they program it and then adjust it. It's a different approach, but certainly easier to accomplish and understand for all operators. Larger chains would certainly prefer an automated system. Reality is often more complex. Who decides which trailers and ads to get, who decides which and when to use them. Who is able to access their trailers on a standard computer (as opposed to ingesting them from 'inaccessible' ext2/ext3 or NAS media).

Extended metadata is useless since server manufacturer are too slow to build such functions into their server software. Also, if we could rely on companies to embed solid LUFS data, we could just as well expect them to master their clips properly loudnesswise - which they simply don't.

That's why those operators interested in preshow presentation sit through it and adjust to their personal liking, instead of trusting numbers.

[macuser]

I think we should make a difference here,

a regular cinema with regular DCPs does not have to worry about the content they get. If the cinema server ingest the content properly the DCP should just work, maybe a short audio listening in the cinema helps the operater to get the correct volume right from the beginning.

It is different for a festival I think. You will get a lot of DCPs from many different sources, from well known manufacturers as well as from many inexperienced 'first users' with whatever software they use. A pre-check with an adequate tool can reveal and find problems that might lead to troubles during a projection. Proper naming a DCP helps the operater to make a good SPL (Show Play List) and helps the festival to show the
correct DCP with the correct subtitles.

During the time I learned a lot about all these problems that can occur. The tools I used were first of all the command line tool 'dcp_inspect' which runs best on a Linux Ubuntu system. It does all the checks a DCP-server would do but will give you many more useful information like average video bitrate, subtitel information (how they are implemented), audio analysis with a small graph showing also the used channel configuration and many more hints and information. I think we could discuss of what makes sense or is just 'nice to know', but the visual check with a software player will give you even more confidence but can never be a tool that gives 100% certainty.

Kind regards
Max

[jamiegau]

Many cinemas are completely run, sessions and playlist wise, remotely. From office setups full of key staff who know how to support the equipment. They do not have a cinema in the adjacent room to check out the volume levels of all the trailers and Ads coming in on a daily basis.

That's the now of cinema, not the future. In larger chains, its unrealistic to think that all these preshow files are tested in a cinema before they go to screen. There is a growing need for cinema operations to find a way to achieve a operational procedure to do that type of QC without having exclusive access to a cinema screen. It's uneconomical and will not be done. Preshow quality has been significantly imparted in my limited experience in my region. As mentioned, I have spoken to EU larger chains voicing a similar concern.

The cost and economics of preshow "requires" we find a workable way to deal with it. I free tool that does a Leq(m), LUFS measurement and waveform visualisation, plus allows the audio file to be played on any PC via webpage app to allow relative to relative audio comparison, would get us 90% there, and is most likely "good enough". Obviously, going to the trouble and time and cost of stepping into a real cinema to judge would be best. But so would returning to the 30% more attendance levels we use to have. The economics of cinema are under pressure.