TOPIC:

I typically create multicams with Sync-N-Link from 2 or more FX6 cameras and a sound recorder that are sync'd via timecode. Occasionally I see a small number (3 to 12) of 1 frame duration multicams in addition to the full duration multicam. The odd thing is if I delete either of these multicam clip pairs, the other is also deleted.

The timeline pic is of the full multicam but the 1 frame looks exactly the same. When opening the 1 frame multicam, the playhead is in the same spot as in the full multicam. Move it in the full multicam, then open the 1 frame and it's at that new position. Like they are clones but one has an incorrect duration.

I sent the XML that Sync-n-link generates to Intelligent Assistance but they said they don't see anything odd about it and importing it didn't result in any 1 frame multicams.

Media and library are on an external spinning 2TB OWC USB disk, formatted at HFS+. Media and cache are external to the library on this disk. The FX6 media is dragged into the FCP event so it doesn't make copies. I've reset FCP prefs but no difference. Anyone run into something like this? It's not a show-stopper but I'd like to understand why it happens so I can avoid it.

One possibility is if you mark favorites or rejected ranges on the multicam clip, and make a slight keystroke error you can create 1-frame ranges. You may not notice those until you filter on "show favorites", "hide rejects", etc. Then it looks like a duplicate but it's one frame from the unfiltered multicam.

Another possibility is some kind of timecode glitch. If you enable clip skimming and skim the 1-frame multicam, what does the timecode show? Open the multicam, set a marker at that entry point and skim all lanes of the multicam constituent clips and note the clip timecode values at the marker. Note you can't skim clip timecode on the monitoring angle, so you have to temporarily click the little TV icon on another lane to skim the former monitoring angle.

Do that for both normal multicam and the 1-frame multicam, and see if the exterior multicam clip timecode and entry point clip timecodes are the same. Make sure all Browser filtering is off.

For each one of those multicams, when positioned inside the multicam on the entry point, for each lane do a SHIFT+F to find the corresponding location in the Browser. Check the clip timecode on that point to verify it matches. Is the same frame inside both multicams refering to the same timecode value in the Browser when you do SHIFT+F?

Is all of this MXF or is any of it ProRes recorded on an Atomos? There is a possible issue if doing simultaneous internal/external recording whereby the two files may not be frame synchronized, depending on the below FX6 settings:

Project>HDMI Rec Control must be set to either "SDI/HDMI Remote I/F" or "Parallel Rec." "SDI/HDMI Remote I/F" is used if no SD card media is in the camera (you want to only record on the Ninja V)."Parallel Rec" is used if you are recording internal/external in parallel. That said, "SDI/HDMI Remote IF" will *also* record in parallel, but the internal/external media is not synchronized to frame accuracy. The Parallel Rec option will record internal/external media with frame-accurate sync, but it requires media in both.

There is yet another issue which can cause a 1-2 frame sync discrepancy if both FX6s were jam-synced from certain sources. I don't know what causes that, and since we switch to continuously-connected timecode we haven't seen it. In this case skimming the clips from both cameras in FCP shows the multicam was properly synced on the same timecode value but the two frames are time-skewed about 1-2 frames with respect to each other or the external audio.

Thank you very much for the quick response Joe.

I notice this immediately after the Sync-N-Link xml is imported and select the imported event. No favorites, rejects or marks are done by me. I believe I've tried it both ways with "Trim clips to common video and audio duration" or "No Trim (plus Rejected ranges)" in the Sync-N-Link settings.

I skimmed the multicams and checked multicam and source timecodes. Everything appears correct. The audio was manually offset by 1-2 frames via Sync to monitoring angle to account for a slight audio delay between the recorder and the cameras, but the "incorrect" multicams are there before these changes are made.

Adding markers in the "correct" multicam's angle editor also adds them in the "incorrect" multicam. To clarify, the "incorrect" multicam in the angle editor is identical to the "correct" one, it's just that in the browser it lists its Media Duration as 1 frame and any changes made to one are made to the other.

EDIT: I just noticed that one of the 3 "incorrect" multicams has a Media Duration of 2 frames. In all three of the "incorrect" multicams, the audio extends beyond the video lanes, though nearly all the multicams have their audio lane extending longer than the video. If an "incorrect" mulitcam is selected in the browser, the available frame (or two frames) is the last frame or two frames of the audio. I attached a screen shot in the angle editor of the "incorrect" multicam with a 2 frame duration and parked the playhead at the end of the audio. You can see that this is the End in the browser. I seem to recall that I've run into "incorrect" multicams where a video angle was longest and its End frame was the single frame, but I can't be positive.

Only MXF and multichannel wave files in use.

I haven't seen anything exactly like that. Are all of your media filenames unique? Are there any media files with duplicate on-disk filenames which are in separate folders? There was a problem with loading XMLs if filenames were not globally unique which could create duplicate clips, but I haven't seen that in a long time. It did not produce 1-frame or 2-frame clips.

If Intelligent Assistance can load your XML without problem, that implies you could load it on another machine without the media and it would also load OK. Maybe that would be an interesting test.

It might be interesting to export your multicam, project or library XML (whatever is equivalent to the Sync-N-Link XML), then compare the two -- especially the multicam definitions. For things like that I use BBEdit which is good at editing XML files, also I use Beyond Compare for doing "diff" comparisons on XMLs: www.scootersoftware.com/

On the outside chance it's related to something on your disk, you could run Disk Utility First Aid (ideally in Recovery Mode). On APFS it should be run three times from bottom to top, starting with the Data container: support.apple.com/en-us/HT210898

I don't think FCP uses Spotlight for file lookups, but if you've never rebuilt your Spotlight indexes it wouldn't hurt to do that (assuming you have it enabled): support.apple.com/en-us/HT201716

Thanks for your help Joe. Yes, all our filenames are globally unique. In addition, only the media for this specific show is on the portable drive, so not possible to collide with other shows. The drive was freshly formatted HFS+ before ShotputPro'ing the media onto it and creating the library. I handed off the drive to the producer so I can't test with it anymore.

I did keep a copy of the XML so I tried your test and imported the Sync-N-Link XML without the media present and there are no "incorrect" multicams. Interesting. I wonder if this could be related to dragging the FX6 media into the events to avoid FCP making copies/rewrapping versus importing.

Are you saying on one machine the media files were imported with "leave files in place" and on another machine imported with "copy to library", and the XML from one was loaded on the other one? Or was it just one or two files that were imported differently? If so are those files used in the "problem" multicams? Just casting about for any possible relationship.

No, sorry if I wasn't clear. FCP prefs are always set to "Leave files in place." All media is dragged into an event. I usually drag in one day at a time so I can assign camera names and angles. Once all days are in and assigned, I export the event as XML using Intelligent Assistance's Sync-N-Link Metadata View, then open the XML in Sync-N-Link and tell it I want multicams using the audio file's name as the multicam clip name. The resulting XML is automatically imported back into FCP and links to the media (which is all external in the original Sony folder hierarchy).