FCPX BMCC RAW & a7s h.264 Roundtrip Issue (Timecode Extents)

FCPX BMCC RAW & a7s h.264 Roundtrip Issue (Timecode Extents Don't Match)

I've shot a feature film using the BMCC (at 24p) and the Sony a7s (at 23.976) and I loaded all the files into Resolve so that I can roundtrip through FCPX. I put each scene's footage into it's own timeline (with either only BMCC files or only a7s files) and then I delivered each scene into separate folders as ProRes 222 Proxy (about 61 hours of footage). I loaded those files into FCPX, did a test edit, and then exported the XML (1.5). Once I import the XML into Resolve I deselect Automatically import source clips into media pool and I get the error "failed to link because the timecode extents do not match any clip in the media pool."

I look at the loaded timeline and it's all red (not connected). Do you know why this is? There was a7s and BMCC footage in the timeline; neither of them are linked correctly (BMCC isn't linked at all and a7s just linked to another, incorrect clip).

Resolve Settings:
Timeline Frame Rate: 24 frames per second
Use Timecode: Embedded in source clip
Assist using reel names from the: Embedding in source clip file
Mixed Frame Rate Format: Resolve
Sort timeline using reel name and timecode is checked

FCPX Settings:
Timeline Frame Rate: 24p

Some additional notes:

1. The BMCC footage timecodes are the same in FCPX as they are in Resolve. However, the a7s footage timecodes are not (is this because of the frame rate in FCPX might set a7s footage to 23.98p and not 23.976?).

2. I tested this using the XML from Export to FCP and also without it.

Potential Problems:

1. I forgot to add the reel name before I rendered out the proxies so I used Adobe Bridge to add "Alternate Tape Name" into the metadata of each of the clips (Alternate Tape Name is what shows up as Reel Name in FCP and sends Reel Name back to Resolve in the XML).

<metadata>
<md key="com.apple.proapps.studio.reel" value="Scene 57"/>
</metadata>

2. I also added the reel name into Clip Attributes of all the Resolve clips manually (User Defined). I also made sure the reel name in FCPX was the same too. Shouldn't have an effect on the above outcome though.

3. I started this project in Resolve 11, went through Resolve 12b1 and b2 and now am on 3 with this same project. Shouldn't have an effect on the above outcome though.

4. a7s files have duplicate names albeit in different folders (and have different reel names set in Resolve).

5. The file structure of my files isn't always consistent, but this isn't present in any of the affected clips because it would be a bad use case. For instance, I have some files under Scene 57/Video/ and some files under Scene 58/Video/a7s and some files under Scene 59/Part 1/Video/. Shouldn't have an effect on the above outcome though.

6. Some a7s files were shot at 120p, but didn't use any of these files in this test. So shouldn't have an effect on the above outcome.

------

The only possible solution I can think of right now is to start fresh with a new project in Resolve b3 and try changing the Mixed Frame Rate to Final Cut Pro X. Then to re-render the 61 hours of footage all over again (took 7 days) and ensure that the reel names are in there and that "Export to FCP" is always set... But I don't think that's going to work.

Just in case anyone needs it, I got a message from Dwaine on this one before it posted here. It said:

"Hi Devin,

The first thing to do to try to figure out what's going on in a time code extents error situation, is to look at the Source In/Out timecodes in the Edit Index list on the Edit page. And see if those have any relationship at all to the Media Pool Start/End timecodes of the clips. Likely there won't be any relationship, or they would link. Then figure out why the 2 don't match.

Regards,
Dwaine

Dwaine Maggart
DaVinci Product Support
Blackmagic Design USA"

I posted a couple of responses to this. He hasn't gotten back yet, but wanted to share them with you.

Thanks Dwaine. Yes, I see a problem here. Two different ones for the BMCC and a7s footage. BMCC first:

Timecode matches between the original media pool with the proxy in FCP. However, the Source In and Source Out are different from those in the Edit Index. Source In is correct, but Source Out is wrong. Source Out, instead of grabbing the original Media Duration in FCP, looks like it's grabbing the Duration of the edited clip in the FCPX timeline. That means if I make any edits to the footage in FCPX it won't have the correct Source In and Source Out. However, just to do a test, I extended the clip all the way in FCPX and re-imported. The Source In and Source Out are now correct, but it still gives me the timecode extents error. That's strange. How can that be?

I see also that the Name of the clip in the Edit Index has a "- v1" at the end of it, which I never put there. Not sure if that's anything...

--------

Then, with the a7s, I'm having the same Source In and Source Out issue as the BMCC, however, I see something else (wondering if this is because I converted the a7s 23.976 footage into 24p during the ProRes Proxy delivery? ... but I had the a7s in a 24p timeline the entire time in Resolve, so it should see that when it roundtrips?)

In Davinci the timecode in the Media Pool for this a7s footage is:

START T/C — 00:00:00:00
END T/C — 00:03:32:00
DURATION — 00:03:32:00

but in FCP it's

MEDIA START — 00:00:00:00
MEDIA END — 00:03:34:02
DURATION — 00:03:34:02

Duration and media end are longer? I guess it converted it to 24p and added length to fit. But when I bring it back into Resolve it doesn't recognize that. I have mixed frame rates set to Resolve and not FCPX. Is that the problem here?

What doesn’t make sense is that when I go into the Davinci timeline and I extend and contract the clips it changes the Source In and Source Out. Is Source Out supposed to always be Source In + the Media Duration, or no? Is Source In supposed to be the T/C start for that clip? I guess the Record In and Record Out are holding the difference?

-----

No, it doesn't look like that's holding the difference. But yeah, there's something going on with this important. I'm looking at the XML into FCP and the XML back from FCP and here's what it looks like:

ORIGINAL XML TO FCP:

<asset id="r21" src="file:///Volumes/BBE%20Editing%20Drive/BBE%20Feature%20Film/Editing%20Files/G-RAID%201/Scene%2057/Video/Blackmagic%20Cinema%20Camera_1_2015-06-04_0249_C0020.mov" duration="1841/12s" audioChannels="2" format="r1" name="Blackmagic Cinema Camera_1_2015-06-04_0249_C0020_[000000-003681].dng" hasVideo="1" audioSources="1" start="10180/1s" hasAudio="1"/>


<clip duration="1841/12s" offset="118463/24s" format="r1" enabled="1" name="Blackmagic Cinema Camera_1_2015-06-04_0249_C0020_[000000-003681].dng" start="10180/1s" tcFormat="NDF">
<video ref="r21" duration="1841/12s" offset="10180/1s" start="10180/1s">
<audio ref="r21" duration="1841/12s" offset="10180/1s" srcCh="1, 2" lane="-1" start="10180/1s"/>
</video>
</clip>

XML BACK FROM FCP:

<asset id="r4" name="Blackmagic Cinema Camera_1_2015-06-04_0249_C0020" uid="519DB76944DF5E9738B71C4A4C8D64D7" src="file:///Volumes/BBE%20Editing%20Drive/BBE%20Feature%20Film/Editing%20Files/G-RAID%201/Scene%2057/Video/Blackmagic%20Cinema%20Camera_1_2015-06-04_0249_C0020.mov" start="10180s" duration="3682/24s" hasVideo="1" format="r1" hasAudio="1" audioSources="1" audioChannels="2" audioRate="48000">


<clip name="Blackmagic Cinema Camera_1_2015-06-04_0249_C0020_[000000-003681].dng" offset="5600/2400s" duration="151000/2400s" start="10180s" tcFormat="NDF">
<adjust-volume amount="-96dB"/>
<video name="Blackmagic Cinema Camera_1_2015-06-04_0249_C0020 - v1" offset="10180s" ref="r4" duration="1841/12s" start="10180s">
<audio name="Blackmagic Cinema Camera_1_2015-06-04_0249_C0020 - reverse stereo" lane="-1" offset="10180s" ref="r4" duration="1841/12s" start="10180s" role="dialogue" srcCh="1, 2"/>
</video>
<metadata>
<md key="com.apple.proapps.studio.reel" value="Scene 57"/>
</metadata>
</clip>

I'm trying to look for any discrepancies. It looks like the duration going in is divided by 12s and going out is 24s. Not sure why this would be. Either way, it comes out to the same duration, so that should be fine. Offset is divided by 1s when it goes in and doesn't have that when it comes back, but that would still be the same value.

I went back to Davinci 11.3.1 just to check and the functionality is the same, so it's not a problem with the Resolve 12 beta 3. I'm wondering what I'm doing incorrectly as I know many people are able to get this to work. When I turn off "Assist using reels" I don't get the timecode extents problem, but instead it just links to incorrect files. So is there a problem with how I'm doing the "Assist using reels" then? Maybe I haven't got it set up right?

Interesting though, that when it links to the incorrect files the timecode for the two files are different, but close:

Actual Clip is Blackmagic Cinema Camera_1_2015-06-04_0249_C0020:
START T/C — 2:49:40:00
END T/C — 2:52:13:10

Instead Links to Blackmagic Cinema Camera_1_2015-05-24_0248_C0089:
START T/C — 2:48:41:00
END T/C — 2:59:07:16

So I think having the filename assist is going to be of a big help. But when I select Assist using reels with filename I get the timecode extents error.

With the a7s footage we have this:

Actual Clip is C0009.MP4
START T/C — 00:00:00:00
END T/C — 00:03:34:02

Instead links to C0009.MP4
START T/C — 00:00:00:00
END T/C — 00:04:29:23

So it would be ideal if it would link it exactly to the exactly Start and End timecodes. Again, with Assist Using Reels, even when I have one file in Scene 57 and another in Scene 99, still gives me that Timecode extents error.

Hi Devin,

This looks like a 23.976 / 24 issue. When rendering out "Individual Source Clips" from resolve, it will render them at the frame rate they are set to in the Media Pool regardless of the timeline frame rate. I wonder if FCPX is re-interpreting the files to 24 and outputting the relative timecode. Try setting your A7S clips to 24p under "Clip Settings" in Resolve, re-import your XML, and see if that links up.