[Grant Peacock]

Good morning Brad!

The cameras were t.code sync’d in the simplest possible way – a simple BNC, coupled with menu choices in the C300’s so that one camera sends and the other receives. Although these (Canon EOS) cameras have a reputation for some drift over a long day, in this case it would be unlikely that they could get more than a frame out of sync with each other (2 hour shoot, both cameras powered and on continuously, albeit at a distance where I decided not to maintain a cable connection between them).

I doubt t.code is a basis of this problem – for most of the 2 hour event, the resultant clips are in sync to within +/- 1 frame, as evidenced by opportunities to manually check sync (eg gavel open and close). If any further investigation could be done, I would be digging into the FCPX importing process.

One odd-ball thing did happen during this shoot, with one of the cameras. As I was in the process of splitting files into aprox 30 min durations (with a rapid record stop/start again), I appeared to be right at the end of one CF card’s capacity. I think my attempt at swapping slots (available on the C300 if the camera is not recording) might have coincided with the moment that the camera itself was swapping media. That might explain why I have clip #’s increasing from Card 1 to Card 2 and back to Card 1, then reverting back to Card 2. But if the basis of a multi-cam file is t.code, I’d anticipate the clip #’s matter not. The next time I cover this event, I will be running Atomos prores recorders on each camera in order to produce a 2nd set of files…. just to be sure!

Grant.

Grant Peacock
Washington DC
GPI TV LLC

ENG and Cine-style
Cameras and Crews

http://www.gpi.tv

[Grant Peacock]

Thanks Doug and Brad for the follow-up. In the end, and relying on instincts, I looked up the latest method for trashing preferences (now pretty easy to do), and started a new Library, and did another import. This time, I directly pulled from the CF cards, generating optimized media without the intermediate step of cloning cards to the internal drive.

The resultant multicam file was mostly good. I was able to run 2 angles against a WAV audio file, using audio as the basis of sync. I was merrily cutting between angles and about ⅔ of the way along the timeline, video had again become uncoupled. I got by for the rest of the project with manually syncing up back to the relevant source clip in the Browser.

Mystery unsolved… Just hoping it doesn’t show up again anytime soon. Pondering : I am now 2 updates behind in FCPX, due to the fact that this MBP8,3 (17″ late 2011) can no longer access OS updates (topped out at 10.13.6, and FCPX 10.4.6. A long and productive companionship is about to wind down.

Grant.

Grant Peacock
Washington DC
GPI TV LLC

ENG and Cine-style
Cameras and Crews

http://www.gpi.tv

[Grant Peacock]

After many hours spent wrangling this project, and 2 calls with Apple Enterprise Support, I have run out of ideas. But I have learned that I have a very fundamental problem here – camera clips are getting mangled during the importing process. Different segments of various camera clips are being merged into a single clip, but without any interruption to the time code (free run).

Because this has happened to both camera angles, separately, I suspect that something is wrong with FCPX. As we were wrapping up the 2nd support call, the staffer wished me good luck – no answers available after an extensive screen sharing session.

How in the world are portions of different clips being merged together, and imported as a single clip? My last ditch attempt at fixing this project was to scrub thru every imported clip, looking for each non-adjacent frame of video, generating subclips, and round-tripping them out and back into FCPX, appending them with clip names such as ‘a’, ‘b’… etc. In several cases, one camera clip that had been imported originally, now had to be split into 6 subclips. After re-importing, and attempting a multicam clip based on audio sync only, I still have a scrambled mess. There may be no way to finish this edit, apart from generating a transcript of the entire shoot, and then manually editing according to dialogue. There’s already about 20 hours spent at the desk trying to solve this.

I’m 24 hours out from my next 2-cam shoot with these same cameras, and I have no way of knowing if this will happen again. Confused, and exasperated!

Grant.

Grant Peacock
Washington DC
GPI TV LLC

ENG and Cine-style
Cameras and Crews

http://www.gpi.tv

[Grant Peacock]

Thanks Joe. Yes, I’m aware of the importance of assigning something that allows clips to be recognized, and sorted into the correct angle – in this case I have used ‘Camera Name’ populating this field with ‘Side Camera’ and ‘CU Camera’. All clips checked, and all looks good via the Inspector window.

I think the next step will be to re-import to FCP, but skip over the normal cloning step – instead just doing a direct import/transcode from the CF cards to the FCPX library. I’ll post some results when I have them in. Thanks again.

Grant Peacock
Washington DC
GPI TV LLC

ENG and Cine-style
Cameras and Crews

http://www.gpi.tv

[Grant Peacock]

Jeremy, that is correct.

Isn’t that interesting?

And it’s just an issue for this one clip, surrounded as it were on the card by multiple other b-roll clips. No problems with any of them as I string everything out for b-roll on the Timeline. Repeating the import process thru the FCP X window reliably produces the muted clip.

And what’s with the silhouette of the waveform being visible?

Grant.

Grant Peacock
Washington DC
GPI TV LLC

http://www.gpi.tv

[Grant Peacock]

Thanks again Jeremy. The reimported XML file did produce a warning dialogue, but it’s basically machine code. I did also delete generated event files, but after quitting and restarting machine, and re-rendering, nothing has changed.

I’ve decided that the way to go is : new library, reimport, and rebuild the edit from scratch.

This has been an elusive problem, and it might not be one with a solution.

Grant.

Grant Peacock
Washington DC
GPI TV LLC

http://www.gpi.tv

[Grant Peacock]

Thanks again Jeremy. By only selecting the project in Browser, and not loading it to the Timeline, I was able to export the FCPXML file. (FCP apparently will not crash until I load the project file to the Timeline)

Now, would the next step be to establish a new library, and reimport media, then reimport the XML file into this?

I don’t think this is relevant, but my Finder level architecture/strategy for each new Library in FCP is to first establish a Finder-based folder inside Movies (name of folder matches Library yet to be created). Inside the folder, I create subfolders : cache, media, output, source. After opening FCP and creating a new library, I position the Library inside that top level folder (and alongside the subfolders), and modify Storage Settings so that cache and media are pointing to these relevant subfolders. I do not change the default locations for Backups or Motion Templates.

I don’t know if there’s anything in that strategy that might have contributed to an issue, as this has worked well throughout my time with FCP X in it’s various iterations.

Grant.

Grant Peacock
Washington DC
GPI TV LLC

http://www.gpi.tv

[Grant Peacock]

Thanks Jeremy,

I apologize : I neglected to mention in my post that the behaviour is identical when going back to a version that is 2 days prior to the date of the issue occurring.

This leads me to suspect that there’s damage occurring to each project as it is opened inside FCP. What makes this so confusing is that the errant behaviour has survived both OS and FCP upgrades – 10.4 -> 10.4.2 (not to mention moving from external to internal SSD drive, just to eliminate that uncertainty as well).

I wish that there was a way to save a project as a discrete file external to the FCP ecosystem, but I’m pretty sure those options don’t exist in FCP X. It may be time to call it a day and restart the project.

Grant.

Grant Peacock
Washington DC
GPI TV LLC

http://www.gpi.tv

[Grant Peacock]

Thanks Steve and Scott.

Over the weekend, I copied all media onto a new drive, and opened up a new library, importing everything over.

I’ve been able to open the prior library and keep FCP running for long enough to copy the project from the original library to the newly created library. If I then immediately close everything down, restart the computer, and open just the new library, FCP again crashes abruptly.

Based on this outcome, my commonsense guess is that there’s something associated with the project file itself that is triggering the problem. I have deleted pref’s repeatedly on startup without seeing any difference. Scott, where are the set of render files located? Does removing the single cache file from inside the library take care of that?

Grant.

Grant Peacock
Washington DC
GPI TV LLC

http://www.gpi.tv

[Grant Peacock]

To everyone : if you happen to have ‘Final Cut Pro X Advanced Editing’ from 2012 (Apple Pro Training Series) and would like the associated exercise files updated to FCP 10.3.4, let me know, and I will post them for download). It will save you some work migrating them through versions of FCP)

grant@gpi.tv

Have a great weekend, and thanks again Noah.

Grant.

Grant Peacock
Washington DC
GPI TV LLC

http://www.gpi.tv