I have been asked to explain how to create a 5.1 OV and 7.1 VF from an MKV file that contains both a 5.1 and a 7.1 audio stream. The same would be helpful in creating e.g. an english OV plus a dubbed VF when both the english audio stream as well as the dubbed version would be in the same MKV.
Both is very typical if you e.g. rip a DVD or Bluray with the plan to create OV and VF from the MKV and thus put multiple audio streams into an MKV.
At first, this appears as a simple task if you are familiar with the audio matrix, DCP audio channel patterns, and the OV/VF creation scheme. However, it turns out it is a bit unnecessary complicated due to the fact that the MKV contains both audio streams WITH the video.
The trouble is - it is at first easy to create a 5.1 (or eng) OV DCP. Now, when the VF needs to be created, the OV DCP needs to be loaded into DCP-o-matic, then the 7.1 (or dubbed) audio needs to be added. Here it becomes unnecessary complicated, because the second stream of the MKV file can not be added to the project as audio only, the video always comes with it, and it blocks the OV DCP video to be referenced. It is possible in DCP-o-matic (many people don't know this) to delete just the audio portion of an interleaved video/audio file (click on the audio track in timeline view, and select 'remove' - the video file will stay). Deleting the audio is actually the same as deselecting all audio channels of that file in the audio matrix, as you can see when you visit the matrix after you have deleted the audio portion in timeline view - by again selecting audio channels, you can revert the deletion and the audio will reappear in timeline view. Nice - but, however, it is not possible to detach the video and leave just the audio in the project. Maybe Carl can think about wether there is a way to accomplish this in a future version.
Because of that, we have no trouble to load the 5.1 OV into the VF project and refer to it - but we can then not access the 7.1 audio stream that is still contained in the original MKV source footage. There is no direct way to add this audio only to the project to create the 7.1 VF.
Two workarounds are possible - start with the 7.1 or dubbed audio already extracted as a separate audio only file instead of having all streams in a single MKV. That may be possible by using the RIP-software, if it allows to extract audio only to an MKV or other file format. Or use an MKV stream splitter/demultiplexer to extract the 7.1 audio from the multistream MKV. Or, load the MKV into Audacity (possible when the FFMPEG plugin is installed), and extract/export the audio from there.
If the audio then is available as a separate file, it is no problem to load the 5.1 OV, tick 'refer to existing DCP' for video, add the 7.1 audio file, delete/unselect the original 5.1 audio from the 5.1 OV, and create the 7.1 VF (adjust audio channel pattern under DCP->audio first and check audio matrix for additional channel mappings for 7.1).
If you don't have access to an MKV demultiplexer or Audacity, you can achieve the same all just with DCP-o-matic, if you first create an intermediate 7.1 OV from the original MKV file (selecting just the 7.1 or dubbed stream in the audio matrix). This will go very fast, since the video from the previous 5.1 OV pass will be reused. After the 7.1/dubbed OV has been created, the 7.1 or dubbed audio is contained in the audio MXF file of that 7.1 OV DCP. This 7.1/dubbed audio MXF can then be added to the 7.1/dubbed VF project as audio only, while the video will again be referenced from the 5.1 OV.
Again, this works without any additional software, but may be a bit confusing to follow from the 5.1 OV through the 7.1 OV to the 7.1/dubbed VF.
After the 7.1 or dubbed VF has been created, the intermediate 7.1/dubbed OV may be deleted (keep the audio MXF to be on the safe side), or kept in case a standalone 7.1 or dubbed OV could be necessary. It will eat the same space on disc as the 5.1 OV, of course, but it doesn't need to be distributed to cinemas, only the 5.1 OV and the 7.1 VF or dubbed VF need to be shipped.
Does anyone have a better idea to accomplish this from a multistream source? It is quite easy to create a 5.1 OV and 7.1 OV by simply changing the audio matrix and choosing another DCP channel pattern under DCP->audio after the 5.1 OV has been created. I wish there would be a similar simple way for the VF, but I guess the only way to keep it simple is to have the VF audio as a separate file.
I don't know wether that has a drawback in different scenarios, but maybe we could have a 'use' checkbox for video and audio, just as we have one for subtitles? That way we could simply get rid of a video track in order to only use the audio of an interleaved file.
- Carsten
specific workflow for OV/VF from multistream MKV content
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
specific workflow for OV/VF from multistream MKV content
You do not have the required permissions to view the files attached to this post.
-
- Posts: 81
- Joined: Tue Apr 15, 2014 1:06 am
Re: specific workflow for OV/VF from multistream MKV content
Another problem that you forgot to mention.
Most of the time the blu ray frame rate is 23.976, that is not problem. DCP o Matic will speed the OV up to 24fps.
However when it then comes to create the 7.1 version file from the MKV then that is still in the frame rate of 23.976 that does not match the new 24fps OV anyway.
So then some frame rate conversion of audio only would be needed. I dont know if there is some software that you can tell the program that this audio is now 23.976 but needs to be speeded up to 24fps.
Most of the time the blu ray frame rate is 23.976, that is not problem. DCP o Matic will speed the OV up to 24fps.
However when it then comes to create the 7.1 version file from the MKV then that is still in the frame rate of 23.976 that does not match the new 24fps OV anyway.
So then some frame rate conversion of audio only would be needed. I dont know if there is some software that you can tell the program that this audio is now 23.976 but needs to be speeded up to 24fps.
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
Re: specific workflow for OV/VF from multistream MKV content
Hi Gunnar,
good point, however, in general, DCP-o-matic will cater for that:
- in the case of creating a 7.1 (or dubbed) 'intermediate OV', the 23,976 referenced 48KHz audio will be resampled accordingly immediately (see screenshot above). So when you later import the audio 7.1/dubbed audio MXF into the VF project, it will already be in 48kHz.
- if you import 23.976fps referenced 48KHz audio and attach it to video content with a different frame rate, DCP-o-matic will also resample (downsample) in order to align runtime for 24fps/48KHz. You can trigger this manually as well (e.g. if you only have audio loaded, without video hinting at the timebase) - select the audio file under content, go to 'Timing', and adjust 'Video frame rate' at the bottom to 23.976. You will then notice under 'Audio' or in content properties that DCP-o-matic will resample the audio to 47952Hz. See attached screenshots. Actually in the german translation I have renamed 'Video frame rate' to 'Video frame rate reference', so it may be a bit easier to understand and to separate it from the 'video frame rate' parameter under the DCP tab.
If necessary, the free Audacity can be used to resample audio (available on all platforms). But in general, it is better to let DOM do the work, as it will automatically choose the right resampling factor.
BTW - I just notice that DCP-o-matic reports sample rates badly under the content tabs as '48000kHz' or '47952kHz'. It's correct as '48000 Hz' under content properties. Well I guess everyone will get it anyway...
- Carsten
good point, however, in general, DCP-o-matic will cater for that:
- in the case of creating a 7.1 (or dubbed) 'intermediate OV', the 23,976 referenced 48KHz audio will be resampled accordingly immediately (see screenshot above). So when you later import the audio 7.1/dubbed audio MXF into the VF project, it will already be in 48kHz.
- if you import 23.976fps referenced 48KHz audio and attach it to video content with a different frame rate, DCP-o-matic will also resample (downsample) in order to align runtime for 24fps/48KHz. You can trigger this manually as well (e.g. if you only have audio loaded, without video hinting at the timebase) - select the audio file under content, go to 'Timing', and adjust 'Video frame rate' at the bottom to 23.976. You will then notice under 'Audio' or in content properties that DCP-o-matic will resample the audio to 47952Hz. See attached screenshots. Actually in the german translation I have renamed 'Video frame rate' to 'Video frame rate reference', so it may be a bit easier to understand and to separate it from the 'video frame rate' parameter under the DCP tab.
If necessary, the free Audacity can be used to resample audio (available on all platforms). But in general, it is better to let DOM do the work, as it will automatically choose the right resampling factor.
BTW - I just notice that DCP-o-matic reports sample rates badly under the content tabs as '48000kHz' or '47952kHz'. It's correct as '48000 Hz' under content properties. Well I guess everyone will get it anyway...
- Carsten
You do not have the required permissions to view the files attached to this post.
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
Re: specific workflow for OV/VF from multistream MKV content
This is what happens when I add an 48kHz/23.976fps referenced separate audio file (AIFF) to a previously loaded 23.976fps MKV:
You do not have the required permissions to view the files attached to this post.
-
- Site Admin
- Posts: 2548
- Joined: Thu Nov 14, 2013 2:53 pm
Re: specific workflow for OV/VF from multistream MKV content
There has been some discussion on a better GUI for this kind of thing on this bug. Way back when I had intended to add a fairly elaborate GUI for OV/VF (probably multiple Content tabs that you could use to set things up). I added the "refer to" checkboxes pretty much as a quick hackm and it has been surprisingly useful, but I think it's a long way from the Right Way to do VF/OV in the situations you are talking about.
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
Re: specific workflow for OV/VF from multistream MKV content
What do you think, would it be possible to simply split/disable arbitrary streams of multistream content? That way, you could easily use e.g. just the 7.1 audio from a video+5.1+7.1... mkv?
- Carsten
- Carsten
-
- Site Admin
- Posts: 2548
- Joined: Thu Nov 14, 2013 2:53 pm
Re: specific workflow for OV/VF from multistream MKV content
I can't think of a reason why not... there's a somewhat related bugs here and there.
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
Re: specific workflow for OV/VF from multistream MKV content
Yup - quite often I modify existing trailers, trimming either a few frames of video (e.g. 'In 3D') or audio. Usually I load the MXFs in order to trim them individually. Works nicely with separate files, but would be handy if possible with e.g. MKV
- Carsten
- Carsten
-
- Site Admin
- Posts: 2548
- Joined: Thu Nov 14, 2013 2:53 pm
Re: specific workflow for OV/VF from multistream MKV content
Makes sense. I've marked one of the bugs against 2.13.x so I'll try to get on that when I can shove 2.12.0 out of the door
-
- Posts: 2804
- Joined: Tue Apr 15, 2014 9:11 pm
- Location: Germany
Re: specific workflow for OV/VF from multistream MKV content
I also checked into MakeMKV - it won't allow to save only audio MKVs, there must be a video stream.
- Carsten
- Carsten