[Joe Marler]

FCPX proxies are 1/2 the linear resolution or 1/4 the pixel resolution of the original material. So for UHD 4k (3840 x 2160) the proxies are 1920 x 1080. For DCI 4k (4096 x 2160) the proxies would be 2048 x 1080. For 6K proxies are likewise 1/2 the linear resolution in each axis.

By “edited a short with HD proxies”, this could mean one of several things:

(1) You edited using FCPX-generated proxies and have access to the source material. In that case change the viewer from proxy to optimized/original and export.

(2) You edited using FCPX-generated “external” proxies and don’t have the source material. You will need to acquire this from whoever has it. Unfortunately FCPX will not reliably re-link if the proxies are externally placed, so whoever created those must tell you what drive volume name and pathname to the source, and you must configure your system likewise. In that case relink will not be necessary.

In some cases of missing proxies you can ensure the proxies and source media are in the same relation locations as the machine they were generated, switch to proxy mode, you’ll see red “missing” icons and you can select those, right-click and pick transcode to proxy, and FCPX will simply update the pointers. Always try this on one or two files first, don’t do hundreds of files else it may try to generated new proxies for all those.

(3) You edited using non-FCPX proxies, e.g, generated by the camera or some external utility. This is not currently supported by FCPX but some have reported the ability to relink these to original material in certain cases. The proxies must be some variant of ProRes.

For more details, see:

Forcing FCPX relink: https://www.geardads.com/fcpx-proxy-editing-relinking-external-proxies-and-temp-media/

Ultimate Guide to FCPX proxies: https://blog.frame.io/2017/05/11/ultimate-guide-to-fcp-x-proxies/

Uber Proxy Workflow:

https://www.youtube.com/watch?v=WAa675ImtrI

Some contents or functionalities here are not available due to your cookie preferences!

This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.

[Joe Marler]

[Tangier Clarke] “My storyline was working fine on my home computer, but bring the project to work, everything is now all grey. An attempt to relink or show in browser gives this message:

“Failure to reveal clip. The original clip could not be found or reconstructed”

The library is running from the same mobile drive between home and work.”

Tangier, you always hit the difficult problems. I’ve never seen that but Google shows a few people over the years who have. BTW the exact syntax is “Failed to reveal clip”.

Since it’s on a mobile drive it’s being frequently ejected and reconnected. I’d suggest running Disk Utility First Aid. Actually before that make an image backup of the drive.

It could be some kind of FCPX database damage. If this is a managed library the file should be there. Can you open the library in Finder and locate the file? One possibility is the file is there and the database pointer and associated metadata is not correct, the other possibility is the file itself is missing or damaged. If you can locate and play the file from Finder, then you know the file is there.

It could also be a developing bad block on the drive which exceeded retry count. That said you’d normally see a hard error or S.M.A.R.T drive warning, not wrong results on at the application layer. The only Mac HDD surface scanning tool I’ve used is Scannerz: https://scsc-online.com/Scannerz.html

There are other tools like DiskTestR but they rely on writing test data (non-destructively) to fill up the drive while monitoring errors: https://diglloydtools.com/disktester.html

I assume the mobile drive is formatted HFS+, not exFAT.

[Joe Marler]

[vidda roshinns] “if i shoot in one codec render in a different codec and export to another seperate codec. If I shoot in MP4, render in pro res 422 and export back to pro res is the actual quality of the render the same as if i had shot in pro res to begin with?”

FCPX generally renders the timeline using ProRes 422, although you can modify the project to use several other ProRes variants. So no matter what the acquisition codec or the export codec, the render codec is usually ProRes.

MP4 is a container format that frequently contains 8-bit 4:2:0 H264 material. ProRes is a codec family which supports at least 10 bits per color channel and various chroma sampling resolutions.

In many cases you don’t decide to shoot in MP4 or ProRes, your camera supports one or the other. Sometimes a camera that records MP4 internally can record ProRes to an external recorder. Some higher-end cameras can record either ProRes or MP4 internally. Some cameras which don’t record ProRes can nonetheless optionally record in higher bitrates, bit depths, chroma sampling, or can use a lower-compression or “intra frame” codec. Using these is a tradeoff of data management, workflow complications and evaluation of whether it makes a perceptible difference to the final viewer.

The recording bit depth and chroma sampling is determined by the camera, not by the codec. IOW adding a ProRes HDMI recorder to a low-end camera will not magically provide true 12-bit 4:4:4 4k recording. The camera’s sensor, imaging pipeline and output logic determines what it can export.

So in most cases a lower-to-midrange camera will record MP4 and there’s nothing you can do about it. In the edit phase FCPX renders in ProRes and your only non-ProRes option is uncompressed 10-bit 4:2:2; IMO there’s no reason to use that. So up the the point of export, there is often no codec decision to make. You normally shoot at the best resolution, codec and bit rate your camera produces, but in low-to-midrange cameras this is often only a choice of bitrate.

Your export codec will be determined by your delivery requirements. If you are uploading to Youtube or Vimeo, unless your program length is very short it may be impractical to export to ProRes. Using an H264 output codec will save a lot of space and upload time, yet produce pretty good quality if the encoding bit rate is high enough. Exporting to HEVC/H265 is also becoming and option, which can cut upload size roughly in half while retaining the same quality. However it requires hardware acceleration (and software which harnesses that) to provide decent encoding performance.

If you are handing off material to a downstream editor or color finisher, you’d generally export in ProRes not a “long GOP” codec like H264. This isn’t because ProRes makes it better but it avoids possible quality loss of re-encoding to a compressed codec. Also I think the most common long GOP versions of H264 only support 8 bit 4:2:0. There are intra-frame versions of H264 which support higher bit depths and chroma samplings but they are less common. To achieve the quality of ProRes requires an H264 variant which is encoded at nearly the same bit rate, so “turnovers” are often done using ProRes.

In general the image quality is “baked in” at acquisition time and transcoding to another codec will not improve that. Thus there’s no reason to transcode to ProRes optimized media before editing in FCPX — from an image quality standpoint. From a performance standpoint it could be beneficial.

However It is theoretically possible to transcode 4k 8-bit 4:2:0 H264 to true 1080p 10-bit 4:4:4 on output. Achieving this requires (1) Your rendering codec must be ProRes 4:4:4, (2) You don’t crop the image during editing, and (2) You export to ProRes 4:4:4. That is one case where choice of rendering and output codec could make a positive difference. Whether that difference is visible to the final viewer on typical playback devices after distribution and upload encoding is another matter. See white paper by Barry Green “The Benefits of Shooting in 4k”: ftp://ftp.panasonic.com/provideo/agdvx200pj/4kbenefits_techbrief.pdf

[Joe Marler]

[vidda roshinns] “complete and utter freezing of not just FCPx but the computer also…. forget force quit, nothing works at all. The only way to get out of it is to hold the computer start button till computer shuts down and then restart… I have had to do this at leat 12 times in the last few days up to as many as 20 times.”

In general this indicates a hardware or system software issue, not FCPX. Like Windows and other modern operating systems, each process in macOS is totally isolated from other processes and the system. It is normally impossible for an user-mode app to crash or hang the OS, whether by accident or intent.

You should first verify your backups in case the situation worsens. After that you can do the usual troubleshooting steps for cases like this:

– Check your free space on all volumes, esp. the boot drive. If the OS runs out of paging file space that can destabilize the OS and cause a crash or hang. This should be handled gracefully but is not always done.

– Run Disk First Aid on all volumes and verify no errors. Note that prior to High Sierra, you have to run it twice on each disk, once on the physical disk, then spin down the disclosure triangle and run it on the logical volume. Unfortunately First Aid only checks the file system and doesn’t do a surface check. A failing hard drive can cause the system to lock up or destabilize. This could be bad block exceeding the retry limit, or a cable or on-board controller failure. First Aid will not catch those. There are various 3rd party drive tests, Disk Warrior is often recommended but I don’t remember if it has a surface test. Disks with S.M.A.R.T. will often (but not always) report an error. Therefore another test is unplug as many hard drives as possible which will still allow problem replication. This may require copying a smaller data set to a test library.

– Run Apple Diagnostics. This is a simple test and if it finds no errors that is not a clean bill of health, but if it finds errors there’s a definite problem: https://support.apple.com/en-us/HT202731

– Re-install macOS. Unlike Windows, this can normally this can be done without wiping out all your apps, because there’s no registry to blow away. However you should have a complete verified backup beforehand, ideally both Time Machine and image backup via Carbon Copy or similar.

– If those steps don’t resolve it and diagnostics don’t produce errors (and definitely if they do find errors) I suggest making an appointment at the Apple Genius Bar to have overnight bench diagnostic run. They have access to much more rigorous hardware diagnostics which will often find a marginal or transient problem. They can also examine the macOS system log for meaningful errors.

[Joe Marler]

[Ken Pugh] “I have a managed FCPX Library and I’d like to send a copy containing only the proxy files (not the original media) so the director can edit the interviews on his laptop and send an xml back to the editor. But…. there seems no way of doing this through FCPX as the copy function automatically includes the original media (with an option to also include proxy).”

This is possible but it’s not well-supported in the current FCPX. There are two basic methods:

(1) The non-supported “proxy cheat” method described in this article: https://www.fcp.co/final-cut-pro/tutorials/1828-cheating-final-cut-pro-x-proxies-to-store-where-you-want-by-using-aliases

(2) The method described in the Ripple Training Media Management tutorial. It’s too involved to post here, and you’d need to keep referring back to the instructions anyway. The best approach is just get the tutorial that way you can rewind it and review it as needed. This covers both “external proxy” case and “editors in different locations”: https://www.rippletraining.com/products/final-cut-pro/media-management-in-final-cut-pro-10-4/

Since there’s no such thing as proxies for audio or graphics files, these must be handled separately in the proxy-only case. The Ripple Training tutorial describes that procedure.

Another complication is file relinking doesn’t work reliably in the external proxy case. So the library of whatever system you build the proxies on will have a disk volume name and pathname to the proxy files. That must match between source and destination machine, otherwise you have to relink and that’s not reliable. The easiest way build them on an external drive which you can name, have source/destination folder trees and volume names the same.

This entire workflow needs to be fleshed out and made more robust. IMO things like this are needed a lot more than superficial user-facing items like color-coded video lanes.

[Joe Marler]

[Andrew Kimery] “An indie self-distributing their own movie online and not offering a rental option at the start is really odd to me”

Frederick Wiseman has long self-distributed his independently-produced documentaries on line, not in streaming but DVD form. You could only buy, not rent. I just checked and his famous 1968 documentary “High School” was put on Vimeo last year; don’t know if that’s legit or not.

Wiseman’s recent documentary Ex Libris also cannot be rented — and it was only released a few months ago: https://www.youtube.com/watch?v=Ddjnbbjl0kM

Wiseman plans to put his entire back catalog freely on line, but only for library patrons.

So in the independent documentary world, I’m not sure you can say not offering a rental option at release time is unusual. Whether they release it on DVD, stream it, rent it, sell it or give it away is up to the filmmaker.

Some contents or functionalities here are not available due to your cookie preferences!

This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.

[Joe Marler]

[Gabriel Spaulding] “At no point does it feel like marketing. It’s storytelling… The film was not paid for or sponsored by Apple and the filmmaker does not work for Apple. I continue to find it interesting (and baffling) that anyone who shines a positive light on FCP X is seen as suspicious…”

I saw it and the extended interviews. I was worried it would feel like a blatant commercial for the product, but it’s really not. Some will interpret it that way. Oftentimes a detailed look at the development of a product can feel a bit like an advocacy piece. This was true for the Pulitzer Prize-winning book Soul of a New Machine, about the missteps and development of a new computer by Data General: https://a.co/a9EWol2

Off the Tracks definitely presents things in a positive way and is not acerbic-toned investigative journalism. But so what — that is not a requirement for storytelling.

It’s interesting that Off the Tracks speaks several times about the democratization of video, yet this documentary itself is a product of that. It doesn’t require “a purpose” to justify its existence. It wasn’t made by a huge team. It was largely a story that a few people wanted to tell, and it was directed, produced and shot by one person.

Before the current era, such narrow topics could not be covered at this level of quality and detail. E.g, decades ago there was talk of making a movie about Soul of a New Machine. It never happened, partially because in that era such a narrow topic was too expensive and risky to tackle. Today that could be done. In fact anyone motivated could go interview the surviving members from that book and make a documentary right now. That is what democratization of video enables. It’s true that FCPX didn’t create this trend but it embraced it.

[Joe Marler]

[Andrew Kimery] “What I don’t like about PPro is the work I need to do to clean things up after the sync. I was hoping X would save me time on this front, but if I have to do a comparable amount of prep prior to sync in X then it’s a bit of a wash time wise. “

With FCPX you normally have to tag all clips with a camera name or angle name before doing the sync. If a camera emits metadata that FCPX understands, this isn’t required, but usually it must be done. It’s a quick procedure if each batch of clips has a unique filename convention or you tagged them on import. In the Event Browser using List View, just group select all clips from a given camera and enter a unique camera name or angle name. IOW if you have three cameras you can batch-tag all clips for a multi-day shoot in only three steps.

If they are properly labeled, usually FCPX does fairly well. Like Premiere, FCPX does not compensate for sync drift. Premiere notifies you of a clip sync failure, whereas FCPX just takes its best shot and places it on the timeline, intermixed with all the other successful syncs.

With any of these audio-sync products they work better if you have longer clips or ideally a continuous audio channel the other cameras and recorders can sync against. So one solution is just leave an audio recorder running.

We shoot a lot of three-camera interviews and if multiple takes are required we might stop the multichannel audio recorder in between. Some cameras may have a false start as the operators adjust them, but even with all this it’s uncommon to have more than about 15 clips per interview.

Some cameras will segment logically contiguous clips into separate files, which can be harder to manage. Where possible we configure the cameras to not segment files.

In a succession of multi-cam interviews I always sync each one separately. If we have back-to-back interviews with the same setup, I don’t try to sync multiple interviews in one pass.

100 to 200 clips is a lot to sync in one operation. I would normally use Plural Eyes for that. It requires no prep whatsoever, you can just drag in a huge pile of clips and it will try to figure out how many cameras and recorders. However if you have multiple similar cameras, e.g, GH4 and G7 or A6500 and A7RII, the files internally look alike — bitrate, bit depth, chroma sampling, codec id, etc. are identical and there’s no differentiating data in the video header. Even Plural Eyes cannot always put them in the right group. So some form of manual organization before import is often needed.

[Joe Marler]

[Andrew Kimery] “I’ve given everything an angle and selected the option to tell X to use the angle info (as opposed to leaving it on “automatic”)”

Are you sure you did this for *all* clips? Each audio recorder must have all its clips uniquely labeled in Inspector under Camera Name or Camera Angle. Obviously you can select the entire batch of clips in the Event Browser from a single camera or a single audio recorder and label them.

Provided they are labeled correctly I’ve had pretty good results with FCPX syncing multicams via audio — using fully automatic. However I usually have an unbroken channel from an audio recorder.

You are correct you can use *both* “First Marker On Angle” *and* audio sync. The markers will fine-tune the area FCPX uses to examine for audio sync. These markers need not be placed exactly: https://support.apple.com/kb/PH12585?locale=en_US

If you have several short “junk” clips, try omitting those from the sync. Those can throw off the sync attempt.

You can open the MC clip and check the sync and attempt a selective re-sync inside the clip. To do this:

– Inside MC clip select monitoring angle by clicking the little TV icon at the left of its timeline. This can be audio or video and need not be the same for each re-sync attempt. You’re basically telling it “within this MC, sync this clip to that one”.

– Click a different clip you want to re-sync. Then click the clip label at left and select “Sync Selection to Monitoring Angle”. It will re-sync only the selected clip to the monitoring angle.

If you didn’t have an unbroken audio channel, you may need to repeat this to sync each group of clips that were running at a given time.

I previously tested FCPX’s audio sync vs Premiere on some difficult cases and they seemed roughly equivalent. However Premiere would tell me if a clip failed to sync, whereas FCPX would silently guess and place it somewhere on the timeline.

Even if audio *levels* are OK but one audio source contains lots of echo and other noise, sometimes the sync will fail. E.g, at a concert or large wedding reception where two cameras both have good audio levels but are widely separated. If one is close to a PA speaker and the other highly echoed slightly time-delayed sound, the audio sync may fail.

PluralEyes is better at this but it can be a hassle to use since it doesn’t support syncing from the FCPX Event Browser, which is the normal workflow. You have to start with your media in Plural Eyes, sync and import to FCPX, or start in FCPX with your media on the timeline and round trip to Plural Eyes. It is designed for the typical track oriented NLE workflow.

That said, Plural Eyes is very powerful and can match hundreds of clips in a single batch. If your audio recorders were not set to the correct date/time and the clips were not placed in correct folders, Plural Eyes can search across every A/V clip from an entire project and sync them. It also does continuous sync drift correction.

The Plural Eyes audio search is many-to-many, and can be pretty slow if matching hundreds of clips. Normally you don’t need that but sometimes you have a lost audio clip without correct date/time and need to find what matches it. I have requested they add a feature to search for a waveform match between one selected clip vs a large pool of candidate clips. This should be possible at much higher performance.

[Joe Marler]

[Ryan Neuman] “been editing professionally for over 2 years now with Final Cut and I feel like I’ve reached my limit with it….- MacBook Pro (Retina, 15-inch, Mid 2015)…I shoot primarily with a Canon XA10 24FPS 1/48 Shutter Speed. I’ve posted a few other times on this forum about glitchy/liney footage and I’m still having the same problems…The solution I’ve been using is opening the AVCHD footage and saving to my desktop (I think by doing this it saves as a MOV file?) and then I drag that into my timeline. Is there a way to avoid this because it takes a ton of time….I had to go back and fix A LOT of the footage I took. I remember before the latest Final Cut update (10.4) it never did this. Is there a converter I can use that doesn’t take long?…Lately, Final Cut has been SUPER slow too (almost to the point of not even being able to use it). I operate off the external hard drive…Do I need a firewire, is that why it’s been so slow?…

I have a lot of 1080p 29.97i XA10 AVCHD material and it works fine on FCPX 10.4 and earlier — provided you do the right thing. But this can be a confusing area, as is typical with interlaced material. Also, AVCHD (and how FCPX handles that) adds another layer to manage.

Whether it’s classified as PsF or not, the video header in those files is flagged “interlaced”. That means FCPX should treat it as interlaced, so the first clip dragged to a default project will create an interlaced project. When exported this in turn will produce an interlaced file. That file will exhibit interlaced artifacts if the player does not deinterlace.

You can play the original .MTS file in VLC with deinterlacing off and it will show comb-type interlacing artifacts. So the file itself has those characteristics, FCPX isn’t necessarily doing anything to it.

On 10.4 you can set the deinterlace check box in Inspector, if you pull down the bottom left selection to “Settings”. That will hard deinterlace the output. Or you can upload the interlaced file to Youtube and it will deinterlace it.

Re import and editing performance on AVCHD and space consumption, you should never copy .MTS files out of the AVCHD bundle and import those with “leave files in place”. This will cause significant I/O performance problems. You should (1) Import from the AVCHD bundle. That automatically re-wraps and copies the content to the library. Or (2) Re-wrap with EditReady, which enables proper importing “in place”. The re-wrap is very fast and then importing with “leave files in place” will be lightning fast.

Re “glitchy/liney footage” I don’t see that on XA10 footage that’s been properly processed in 10.4. I think prior to 10.4 there was a behavior if you imported interlaced footage to a progressive project without setting the deinterlace check box for that clip, it would discard every other line, producing poor-quality “liney” footage. It doesn’t appear 10.4 does that. This may have been isolated to certain Canon PsF content; I can’t remember.

If you upload to Youtube an improperly deinterlaced file, their encoder cannot tell the file is interlaced so it leaves it that way. If every other scan line has previously been discarded, it will look poor. Or if an interlaced file was uploaded but the video header is marked progressive, it won’t try to deinterlace it.