To FCP X or to PluralEyes - What's your workflow?

I recently revisited PluralEyes (having started a thread on the COW some years ago here forums.creativecow.net/thread/344/20553 )
and here's the problem I have with using it even though it's a great product for it's specific purpose(s):

Plural Eyes replaces audio from the camera and will replace it with the separately recorded audio.
I don’t want this. I want it to add the audio to the video file (with original audio) in case I need to use the original audio. I like how you always have your original audio when syncing in FCP X.

PluralEyes assumes that your separate recorded audio is always the best, generates new files with new timestamps different than the original shoot footage. This could present problems in finding content in a DAM software. Additionally you have new masters which, if you backup your original shoot content, means you have to spend more money on drives space because your edit references the content PluralEyes generated.

Can’t sync from different cameras. Another limitation is that you can’t sync clips from one camera with clips from other cameras. This is because Export Media Files can’t handle overlapping files in the timeline, and goes back to our discussion of "standard" projects.

If there are two or more audio-only tracks in your PluralEyes project, only the uppermost audio clip that overlaps with a video clip is included.

PluralEyes will trim original audio to match the corresponding video.

Dilemma: FCP X really requires all gear timestamps to be the same or close just to get the content organized in list view to prep for syncing. It helps that devices are started and stopped at about the same time for duration sorting. If you don’t have that important clip information, then FCP X syncing is a much more manual process. PluralEyes on the other hand you can just throw content at it and it figures it out. But if replaces audio which isn’t always a good thing and it creates new timestamps on the files which aren’t true to the shoot and may be relevant at some later point.

I would prefer PluralEyes just generate an XML describing what needs to be synced/multicamed and let FCP X process it.

I am sure I am missing something her. I know that FCP X also creates new files. I do like that it can also adjust the content creation time. Just trying to find a better (less manual) solution to organize my content into synced and multicamed clips that maintains all of the pertinent info and metadata, while allowing me point back to my originally shot content during the project archiving process. Perhaps it's not really necessary.

tangierc wrote: ...Plural Eyes replaces audio from the camera and will replace it with the separately recorded audio. I don’t want this. I want it to add the audio to the video file (with original audio) in case I need to use the original audio....

You raise some good points, I intend to investigate further but I haven't had time.

I have seen PluralEyes used various ways: import into FCP X, then round-trip via XML out to PE for sync, vs initial import into PE then to FCP X. PE can produce files or go straight to FCP X, and optionally will replace audio -- or not. It can create a sync'd multicam clip for import to FCP -- or not.

I usually throw everything into PE, sync then use these options: "Import automatically into FCP X", and "Create event with multicam clip". I do *not* select the option "Create event with replaced audio".

Within FCP X, all the original audio and video is in the sync'd MC clip and I just use the angle viewer to edit the MC clip.

It seems to mostly work OK. I thought I had seen cases where I could not backtrack to the original file because the filename in FCP X was wrong. However I cannot reproduce that -- it's working fine now.

There is an "overlap" case in PE where it won't sync all clips. This happens if two clips from the same camera somehow overlap, it will not sync them both to the other timelines. Rather it picks one. This apparently happens if you do many short clips and waveform analysis thinks two clips from one camera overlap. Rather than truncate one clip and merge them, it just picks one -- apparently randomly.

If the picked clip is short, the non-picked one is long, and the upper/lower clips are long, there could be a lengthy non-sync'd region. In theory PE could detect the overlap, truncate one clip and sync them both, but it doesn't do that. An easier solution would be just picking the longest clip to sync in the overlap case. I don't know why they don't do that.

Most of the time these short clips are just junk takes, but they can interfere with the sync. I usually delete all the really short clips before I sync.

I was recently having a lot of issues going from FCP 7 to FCP X using 7toX. The media in FCP 7 had been synced by PluralEyes. I was trying to follow the chain of problems and that's how I ended up testing PluralEyes again. The specific problem with 7toX getting the XML over to FCP X was that we kept having audio with no audio channels or just plain silent., but the source had audio - playable audio.

In any case here's my findings doing some tests with FCP 7, Premiere Pro and FCP X

11/19/2015

Sync Tests
PE 3.5
FCP X 10.2 on El Capitan.
One video clip/ one audio multichannel audio for each video clip.
Multichannel audio files

PE to FCP X XML no options checked
Synced clips exist only as project , no browser clips in sync.
Audio clips in synced project dead - not audible.
Audio clips in synced project missing channels from original audio.
Audio does point back to original audio in browser - defeats the point.
Useless without synced browser (bin) clips synced
FAIL

PE to FCP XML Event option checked
This replaces camera audio with 2nd system audio
Clips show in browser
No audio configuration available in FCP X
No audio from synced clip can be heard
Audio show synced audio shows inside clip (reveal clip in timeline), but is missing channels, no waveform, can’t hear
Audio is not trimmed to match video.according to PE manual
FAIL

FCP X XML to PE then Back
This seems to work (sort of)
Synced clips are compound clips in FCP X
Camera Audio was replaced (not sure I like that)
Audio in synced clips have two mono channels, original had three
Audio in synced clips points back to compound clip, not original audio (problem)

APPCC
Batch processing creates sequences instead of clips (I prefer clips to work with them as editorial objects)
Every synched object is called multicam (which may not be the case even if there’s only one video to audio synced)

APPCC to FCP X
APC XML to FCP X (via 7to X): APCC multicam sequences become projects in FCP X. Not good for editorial. Need them to be synchronized or multicam clips.
Multichannel audio separates to tracks.

APPCC to FCP 7 via XML
Multicam sequences become projects in FCP X. Not good for editorial. Need them to be synchronized or multicam clips.

Is your footage and audio jam synced? If so, it sounds like you'd want to investigate Sync-N-Link: assistedediting.intelligentassistance.com/Sync-N-Link-X/

Export an XML from FCPX, combine audio source and footage (while retaining onboard audio) in Sync-N-Link and export back FCPX as an XML.

You might look into Davinci Resolve, which is free and can synch anything thrown at it blazingly fast. You have the option to keep or strip the original camera audio.

… Dilemma: FCP X really requires all gear timestamps to be the same or close just to get the content organized in list view to prep for syncing. It helps that devices are started and stopped at about the same time for duration sorting. …

You're aware of the 'Synch on first Marker' method?

synching here various formats, no TC just timestamps, and even those are not same (I'm lazy to check cameras)…
All I have to do is to find in all angles a synch-point (here: a refs whistle); hit M; MC one in seconds. No continous recordings, diff. fps etc etc...

That Marker doesn't have to be at the beginning; …

To add to Karsten's comment-
The Marker you add doesn't even need to be at the exact frame of the sync event. It just needs to be roughly at the same sync point. It's a great way to sync things quickly.

If you make markers, and don't specify FCPX to sync to markers, it will analyze the audio around the markers first, and extra carefully before looking at the rest of the clip. Wait, someone said that already. Never mind...

It's not jam-synced, but I am aware of that software.

I didn't know about the "loose" marker syncing, but was aware of being able to use markers.

I looked into Woowave Dreamysnc and tested it out, but it didn't fare well compared to the claims and the videos; at least not for us using content from a Canon C100. It's a very nice product that is trying to use Apple's AV Foundation rather than PluralEyes old Quicktime API. I really hope some of the kinks are worked out, but it seems FCP X is a tough one to work with from a developer standpoint; that and AV foundation.

Maybe you could give me a little insight on using DaVinci Resolve. I tried syncing as you suggested and here are my concerns:

Once clips are synced I don't see any synced clip available (like in FCP X). It seems according to the manual that you just have to trust that it is synced.

When I put the video clip that I synced with audio on the timeline, I do in fact see more audio tracks appended to the video clip than the video clip had by itself (for the camera audio).

The appended audio in the timeline is named the same as the video clip rather the audio wav file name that was appended.

Whether in the media pool or on the timeline, the supposed synced clip shows no indication of sync information in the inspector.

I went through the DaVinci manual and tried batch syncing and one video/one audio syncing and there's no indicator I can find that shows a clip was synced; sort of like FCP X creating a synchronized clip.

DaVinci did not do a good job at all of batch syncing media. It failed completely.

General performance of DaVinci seems extremely clunky and slow compared to FCP X on my same computer; not only in attempted syncing, but also, scrubbing and playback performance.


I lay all of this out because I am relatively new to DaVinci and will continue to explore it more, but I hope you can shed some light on this and perhaps some things I am missing.

Woo wave Dreamsync wasn't all I had hoped even though they're trying to use AV Foundation rather than an old Quicktime API as RedGiant PluralEyes does. Dreamsync is great product, but still needs to work some things out. My hope was to use DaVinci to sync my content, export an XML and import to FCP X.

Thanks for any insight.

Hi Tangierc: see below

tangierc wrote: Maybe you could give me a little insight on using DaVinci Resolve. I tried syncing as you suggested and here are my concerns:

Once clips are synced I don't see any synced clip available (like in FCP X). It seems according to the manual that you just have to trust that it is synced.

Resolve doesn't create a new clip with a "-synchronized" appended like FCPX, but you can see that there are more tracks or the waveforms are different. I don't think trust is the issue, but you can verify via the number of audio tracks. I always "append" the audio instead of "replacing" and turn off the camera audio in FCPX.




When I put the video clip that I synced with audio on the timeline, I do in fact see more audio tracks appended to the video clip than the video clip had by itself (for the camera audio).

The appended audio in the timeline is named the same as the video clip rather the audio wav file name that was appended.

It's pretty standard to ignore the audio WAV file name and go with the camera file name because you are synching audio to picture, not the other way around. Also, audio is generally ignored if it is beyond the video clip extents.



Whether in the media pool or on the timeline, the supposed synced clip shows no indication of sync information in the inspector.

I went through the DaVinci manual and tried batch syncing and one video/one audio syncing and there's no indicator I can find that shows a clip was synced; sort of like FCP X creating a synchronized clip.

Besides an increased number of tracks in the metadata, what else would you need?



DaVinci did not do a good job at all of batch syncing media. It failed completely.

Hard to say, but perhaps you didn't put the audio in the same bin as the picture files while you're in the Media tab. You must do this in order for this to work. I've had Resolve synch dozens of files to audio pretty quickly. Also, you need to specify whether it's synching via audio waveforms or timecode, so perhaps this was not set correctly.


General performance of DaVinci seems extremely clunky and slow compared to FCP X on my same computer; not only in attempted syncing, but also, scrubbing and playback performance.

Resolve requires a fairly beefy machine compared to FCPX, but if you're just synching audio, it should work on most modern machines. What's your setup? What kind of footage are you working with?


I lay all of this out because I am relatively new to DaVinci and will continue to explore it more, but I hope you can shed some light on this and perhaps some things I am missing.

Woo wave Dreamsync wasn't all I had hoped even though they're trying to use AV Foundation rather than an old Quicktime API as RedGiant PluralEyes does. Dreamsync is great product, but still needs to work some things out. My hope was to use DaVinci to sync my content, export an XML and import to FCP X.

Thanks for any insight.

Thanks for the reply Paul.

I too append and never replace my audio from the camera in case for some reason I do need it. I also turn it off in FCP X.

I realize very soon how very spoiled I am by the performance and feedback FCP X provides with clip generation, naming, and visual representation in both the browser and inspector. DaVinci does not provide that (at least for this particularly exercise).

Having the synced audio stay the same name after syncing (as in FCP X) is extremely helpful so I always am aware of the source file if for some reason I need to go back to it. I really don't understand why DaVinci would change that. In FCPX synced clips I'll see the video clip name and the original wave file names in the inspector and the storyline. That's a personal preference I suppose. I don't like things abstracting information more than needed. Hopefully DaVinci is able to point back to the original wav source file when trying to do that operation on the audio track that has been appended and renamed.

My media was all in the same media pool. No additional bins.

I synced using the "Auto-sync audio by waver form and append tracks" feature.

I am using an iMac 5K retina , 3.2 quad core with 32 GB RAM with a RAID 10 thunderbolt setup.

Thanks again Paul.

Hi, maybe you like it maybe not, but here's my workflow... (on this given project).

I'm a student, so example: all files were named like videoclips: projectname_0015
audioclips: projectname_0175
There was no TC no Clap or reference, no comments or logs from the shoot, audio and video were not always stopped and restarted at the same time, etc...
5 days, 11 hours of material.

1.- I leave files in place for video
- I import files to library for audio

2.- string-out of video for Day (...) + connect all audio

3.- import project in PE and let it sync

4.- send fcpxml back, only timeline, no new event, or multicam or open automatically ...

5.- open the project fcpxml in the library.

6.- now the PE project in fcpx shows me wich video goes with which audio.

7.- now I can really pick fast in the many numbered clips and sync on audio waveform inside FCPX

8.- 11 hours = took me 1 day syncing.

11. - no problems what so ever after that .... (camera audio and connected audio available)

I´m in slightly desperate hope for some optimizations to my personal workflow.

The basics:
I have a funding-gone-wrong documentary project which is quite far evolved and has so far around 100 hours of footage.
Different cameras, mainly Blackmagic Pocket and C300, and different audio configurations.
Main audio is 3 channel external audio (sound devices: Boom M/S, lavs). Additionally Shure Lenshopper with internal recording, Rode Stereo Mic X recorded to the BMPCC.
Audio files are long.
Camera(s) are started-/stopped regularly. C300 is jam-synced BMPCC obviously not.
I won´t need Multicam clips though I have two cameras running regularly.
What I need is the possibility to get (all of) the audio into audio post.
Until now I have used FCPX for shorter projects, dealing with audio sync in different ways.

My goal:
Get multichannel clips - synchronized clips or compound clips - that reference the original media and show the timecode of the included video file.
I need the reference for audio post and the timecode for sorting.


My present workflow:

1: Import everything in FCPX
2: Create projects per day and per camera with all clips in chronological order.
3: Within the daily projects create secondary storyline per audio device
4: Create and edit Roles, Channel configuration, ...
5: Export project-XML to Plural Eyes (PE)
6: Sync in PE
7: Export XML to FCPX nothing checked.
8: Import to FCPX

So far so normal. In the next steps I get rid of blank video within my yet to be finishes audio synced clips.

9: Open project from PE; create one intermediary(#1) compound clip from audio storylines (halves the mouse clicks later)
10: Cut the new audio compound clip at start and end of the video clips.
(So I will have no clips in the browser without video at start and end later.)
11: Break apart intermediary(#1) audio compound clip

And now comes the annoying part to get the timecode in the clips right:

12: For every video file within the project one by one make another intermediary(#2) compound clip from only! the audio files.
13: Within the event browser: use synchronize command to sync the resulting compound clips(#2) to the corresponding video clips.
(This takes a long time with many clips. Opening two event browsers would help)
14: Open the (many) synchronized clips in the timeline one by one and break apart the intermediary(#2) compound clips within the synchronized clips.
15: Delete all the intermediary clips.

Sounds complicated. Is complicated. Any help to optimize this workflow very much appreciated.
What you get: synchronized clips with multichannel audio that reflect the source videos timecode and the source videos name.


From searching the forums I´m not the only one with the configuration:
- many video files
- less audio files
- different audio sources
Should´t be too difficult to automate the process from a programming perspective.
I think i´l write a letter to Plural Eyes, see what they say...

nils.dettmann wrote: I´m in slightly desperate hope for some optimizations to my personal workflow.

The basics:
I have a funding-gone-wrong documentary project which is quite far evolved and has so far around 100 hours of footage.
Different cameras, mainly Blackmagic Pocket and C300, and different audio configurations.
Main audio is 3 channel external audio (sound devices: Boom M/S, lavs). Additionally Shure Lenshopper with internal recording, Rode Stereo Mic X recorded to the BMPCC.
Audio files are long.
Camera(s) are started-/stopped regularly. C300 is jam-synced BMPCC obviously not.
I won´t need Multicam clips though I have two cameras running regularly.

If your C300 is jam-synced to audio, why don't you use Sync-n-link? That's what it's designed for. If you sound person has named all your tracks, you also get it in there as roles (or subroles). It can automatically make multicams of them in addition, and then you can take your BMPCC and automatically match that to your reference audio clips. It's important to arrange your clips as an event for each day.

Thanks for your answer.
Unfortunately the C300 makes only a very small part of the footage. And even when it is present the C300 is not always recording when the BMPCC is. It´s not really a multicam setting. Sometimes the BMPCC would film different angles of the same situation but most of the the time it´s somewhere else.
I will sync the C300 by timecode.
Why is it important to create daily events?
I´m working with keywords, which is working so far.

Having an event per shoot day is just a requirement for Sync-n-link. It's how it knows how to sync many audio files to many video files. If you tried to do that will many days all in one event, it could get confused because you'd probably have overlapping timecode (especially for time-of-day timecode for example).

As for your process, have you considered just getting a hop from your audio source to your camera, so you'd just get it right onto your camera? You probably can't get the different components, but the mixed-down track might be all you need for most things, and when you need the components separately, you still have them. Otherwise, there's no "easy" way with your kind of setup other than what you're already doing with PluralEyes. It may even motivate you to replace the BMPCC to something capable of jam-sycing.

I actually recorded a mixdown track from the sound recorder to the camera.
One channel comes from the sound recorder, the other form an attached Shure Lenshopper, which is additionaly recording by itself. (Unfortunately the BMPCCs sound recording is really, really bad.)
For simple editing, the track on the BMPCC would be ok, but how would connect the original audio files later.
The main theme of the doc is a chamber music festival so sound is an important aspect.

Anyway, I would happily have gone the way of another camera. Actually the BMPCC is really a bad choice for this type of documentary or probably even documentary in general.
But as I stated in the beginning. The funding has gone wrong, which came clear only very late in the planning and I had to go with what I had.
When I (hopefully) return shooting this summer I will certainly try to film with another camera and use the BMPCC for some additional second camera work, as it was intended originally.