Creating a DCP from Bluray

[Bill Seipel]

Had some strangeness with our DVD/Bluray player tonight.
It played for 115 min without fail then.... quit.
Projectionist was able to then get a picture back up...but no sound.

I suspect HDMI skullduggery. I was not there.

Which brings me to:
Creating DCP's from blurays.
How long does it take for a 90 min film?

Obviously this will be hardware dependent but thought I'd get some advice.

(Unsure why this date thing is attached to post)

[Agnius Acus]

With a modern CPU, it shouldn't take more than 3-4 hours, assuming 2K / 24 FPS / 150 Mb/s. Try using DCP-o-matic.

[Dennis Benjamin]

It's a process as you need to rip the Blu-Ray first (extract the video and audio file from the Blu-Ray). They do have software for this. Once you have a .MKV file or .MP4 file, you can make one of those files into a DCP.

[Bill Seipel]

The good news is I have a lot of experience with DCP o matic. I've just never made anything longer than an ad.

[Ryan Gallagher]

If you get better than real-time transcoding with DCP-o-Matic you are doing pretty good. I'd avoid laptops if it is going to be something you do on the regular.

DCP-o-matic will definitely utilize all threads available, so the more cores the better. My Dual CPU 12Core/24Thread vintage Mid-2012 Mac Pro outperforms *most* modern laptops. GPU quality is going to be less important for DCP creation tasks, it is a CPU workload from my experience, though it may have some benefit in the blu-ray ripping stage (unsure).

Edit to correct the model of my Mac

[Leo Enticknap]

Related to which, has anyone managed successfully to make a 3-D DCP from a 3-D BD? A customer recently asked me if that is possible (for a legitimately licensed screening for which a DCP version of the content is not available but a BD is, using a Series 1 projector that cannot accept 3-D HDMI) and I started to do some experimentation, but didn't get very far.

I ripped the BD, resulting in the conventional .m2ts file in the STREAM folder, plus a folder called SSIF, presumably containing the 3-D versions. From what I can find out online, SSIF stands for stereoscopic interleaved file, with each file containing both the left and right eye frames. I have three DCP authoring solutions available: the Easy DCP Creator plugin for Resolve, an old-ish version of CineAsset, and DCP-o-Matic (all running under Ubuntu). The first two won't allow me to import the SSIF file at all. DCP-o-Matic will, but when I change the file type to 3-D, it reports that it isn't a 3-D video file. When I select 3-D alternate, it shows the running time of the SSIF file as being half that of the actual movie, and DCP-o-Matic crashes if I try to play any of it in the preview window (with 3-D alternate selected as the file type).

[Ryan Gallagher]

No experience, but i believe mkv blu ray rips are AVC/MCV packaged, and perhaps DCP itools expect SideBySide (SBS) frame packed instead?

This thread mentions a tool can be used to convert from AVC/MCV to SBS, which may play nicer in the DCP tools?

https://forum.makemkv.com/forum/viewtopic.php?t=23475

Good Luck!

[Leo Enticknap]

Ripping is not the problem: the end result of that process is an exact duplicate of the file structure on the BD, only unencrypted. The problem is getting from the SSIF file to a 3-D DCP. I'm wondering if there is an intermediate conversion process that could be done to turn it into a file format that one of the DCP mastering apps I have can use, but I can't find anything online about software that produces or otherwise works with SSIF files.

EDIT - this app claims to be able to transcode a SIFF file into a frame sequential MKV file, which Resolve and DCP-o-Matic should be able to work with. I'll experiment with it when I get a moment.

[Ed Gordon]

Re: Can MakeMKV properly rip 3D Blu-rays?

Post by preserve » Fri Oct 23, 2020 4:07 am
Short answer, yes, MakeMKV works properly with 3D Blu-rays. It copies the data exactly as it is on the disc into the MKV container.

What you do with the MKV afterwards for playback will depend on the requirements and capabilities of your playback software and hardware.

3D Blu-rays are not encoded SBS or Top Bottom, they are encoded as AVC/MVC. This is output by hardware Blu-ray players as frame-packed. (You can use BD3D2MK3D to convert to SBS if you wish.)

Some related threads:

https://www.makemkv.com/forum/viewtopic.php?t=9702

https://www.makemkv.com/forum/viewtopic.php?t=14456
Using: ASUS BW-16D1HT 3.00

[Marcel Birgelen]

So-called frame-packed 3D BluRays are a ton of hurt, because this specific playback format isn't readily supported. The upside is that you get the full resolution out of 3D, instead just half like with the SBS stuff. The SBS stuff essentially relies on normal playback infrastructure, but pulls the 3D magic in the rendering part, sometimes even in the projector itself. I remember shit like subtitles being rendered across the entire frame and then chopped into left and right by the projection system... The problem is that almost anything 3D died about 10 years ago...

Check this guide that uses some of the same tools to remux a "frame packed" 3D BluRay into an ISO, the lossless way and with functioning subtitles. If you don't have to deal with subtitles, having those two separate m2ts streams should be sufficient. I'm not sure if DCP-o-Matic supports two separate m2ts streams as input for stereo 3D, but DaVinci Resolve and Avid Media Composer simply allow you to slap both on the timeline, one for the left and the other for the right eye. The DCP-o-Matic route would be the quickest one, if it works.

I've used this particular info once in the past, for a private screening of Gravity on a non-DCI, two-projector setup. The playback was something I haphazardly rigged together with MPlayer and a Black Magic card with dual 3G-SDI outputs...

[Leo Enticknap]

Thanks Marcel. The tutorial you link to involves a nontrivial learning process, so I think I'm going to try the app I found, and see if it'll take my SIFF file and give me separate left eye and right eye MKVs. If so, then, as you point out, I should just be able to dump them into Resolve, and mission accomplished.

3-D is like a character from a 1970s zombie pic: it plagues us forever, and is impossible to kill. Just when we think we've seen the back of it, someone tries to revive it, either on the mainstream or arthouse side. Just after Christmas I upgraded a site from a DSS200/cat745 to an IMS3000 in an NC3240 projector (they very sensibly decided to do that as a planned upgrade, rather than wait until the cat745 borked and be forced to do it as an emergency response, losing shows in the process). After completing everything, I suddenly remembered that there was the Dolby DFC100 and color wheel 3-D system to take care of. What I thought would be a 20-minute process of setting up a few configurations turned into a six-hour nightmare, involving changes to the projector settings that were totally unmentioned in Dolby's cheat sheet, that SharpNEC had no idea about, and that I had to figure out by trial and repeated, frustrating error. And of course the colors had to be re-shot with a pair of glasses gaffer taped over the lens of the meter.

Audiences hate having to pay extra and wear the uncomfortable glasses, theater staff and managers hate having to wrangle and clean hundreds of pairs of them, but yet attempts to revive 3-D seem to persist.

[Ajala Opeyemi Noah]

My work with DCP o matic 2 is excellent thus far and have been working well for me with DP2K-10S

[Mike Blakesley]

Creating DCP's from blurays.
How long does it take for a 90 min film?

Circling back to the original question:

I use MakeMKV to rip the original movie file to an MKV. That takes maybe around an hour or a bit less.

Then, I use DCP-O-Matic to create the final film. I usually start that process on my computer when I leave for the evening and it'll usually be done by the next morning. I've had some films take just a few hours, and others still be running into the next morning. I'm not sure why the difference, but of course the length of your source movie will play a big part in processing time. Yes, hardware plays a huge part in it too, but more recent versions of DCP-O-Matic seem to work faster, so get the latest version.

Hint: Some blu-rays are mastered loud, and can create clipping distortion in your sound system on playback, which will ruin your audio presentation. To get around this I always reduce the gain by 1.5 DB when creating the DCP. This seems to result in about the same volume as a standard theatrical DCP from the studios.

The last bit of the process also takes a bit of time: TEST THE RESULTS. We do a lot of DCPs for private shows, and I always insist that we have the source material at least a week before the show time to allow for testing and re-converting, if necessary.

[Steve Guttag]

I've generally take the time to do the audio analysis to see if there will be any clipping and reduce the level, appropriately. Watch out for subtitles. There are some movies where subtitles may only be there for some scenes. You may have an english speaking movie but a scene or two where an alternate language is used between characters or due to a particular location...if you don't have the subtitles from the Blu-ray rip, you won't have them to put back into the DCP (and then you have to find the right subtitle file that just has them for the few scenes that need them but not playing throughout).

[Carsten Kurz]

The DCP encoding time highly depends on the number of HW cores/ computing threads available. Some of the more economic AMD Ryzen CPUs get you well above real-time encoding. So, an encoding time of around one hour is now possible for 24fps/2k features on single CPU systems.



Here are benchmarks which give you an idea about CPU/systems choices: https://dcpomatic.com/benchmarks/input.php?id=1


Now, 3D MKVs in AVC/MVC - DCP-o-matic currently can not read that 3D format, you need to convert it to separate streams, SBS, frame interleave, whatever.

https://www.videohelp.com/software/BD3D2MK3D

The full audio analysis in recent versions took very long to finish - that seems to be fixed now. In addition to common peak and RMS values, it also yields LUFS/R128, and since recently also LEQ(m).