[Joe Marler]

[Greg Ball] “90-minute presentations with 3 cameras. 2 angles are the presenter wide and tight shots, and the third is an Atmos recording of the PowerPoint slides. All video is 1920 X 1080P 29.97 frames per second.

I’m trying to synchronize the angles, and it’s taking forever. Is this normal? Any way to speed this up?”

Besides the previous comments you can also use PluralEyes which has better and faster sync algorithms than FCPX: https://www.redgiant.com/products/shooter-pluraleyes/

Another option is use Tentacle sync devices. My doc team uses these frequently and they work very well. I just synced a 3-hr-long six-camera 4k shoot using Tentacle and it took a few seconds, despite most of the cameras shooting stop/start: https://tentaclesync.com

[Joe Marler]

[Greg Ball] “…editing a series of 90-minute presentations with 3 cameras….I’m trying to synchronize the angles, and it’s taking forever. Is this normal? Any way to speed this up?”

From Apple KB article PH12659 “Final Cut Pro X: Create multicam clips in the browser”:

https://support.apple.com/kb/PH12659?locale=en_US&viewlocale=en_US

“Some audio recordings are not suited for use with this feature. Selecting this option may result in long processing times”

To expedite this the article suggests:

“First Marker on the Angle: Final Cut Pro uses the first marker in each angle as the sync point….you can use the first marker to define a region that can be fine-tuned with the “Use audio for synchronization” option (described above). In other words, you don’t need to place the marker exactly—just close enough so that the automatic audio sync feature can sync the angles the rest of the way.”

[Joe Marler]

[Oliver Peters] “The screen shot does show mask qualifier controls on the color wheels page.”

Interesting. However it doesn’t show a major needed improvement, which is the ability to crop from within a color effect. This forces a workflow where you must hop in and out of the color effect many times to crop/uncrop while checking the vector scope. Maybe they will eventually enable this.

[Joe Marler]

[Vincent Cabview Holland] “large file 4K clip I can scroll through it amazingly fluently from within the browser by hovering the mouse over it. But when I place it into the timeline I cannot scroll as fluently anymore all of a sudden. This has always been the case for me on several iMacs and different versions of FCP, and I am wondering why this is, as it should be playing from the same file within the library. Currently on the iMac Pro, SSD or HDD makes no difference. Project settings at automatic, but have tried all different settings. Different cameras too. When I use optimized media, it is OK, but this is severely time consuming for hours of film each day. “

This is an excellent question. Here are three factors which affect this:

(1) The execution code path used for Browser skimming/playback is apparently different from timeline skimming/playback. Even if Browser waveforms are enabled and timeline waveforms disabled, and the Viewer is in Better Performance, the difference exists.

I don’t know why FCPX cannot use the same low latency code path for timeline skimming. Maybe it must monitor for various user editing events unique to the timeline.

(2) Certain codecs magnify this effect. Your material is maybe Sony XAVC-S 4k which is H264. By contrast similar 4k H264 material from a Panasonic GH5 at the same bit rate, bit depth and chroma sampling behaves much smoother in FCPX. The internal composition of each H264 codec can vary. E.g, GOP size (fixed or variable), number of frames between I or P frames, etc.

(3) You are currently using the iMac Pro, which does not have Quick Sync but uses hardware acceleration in the T2 chip for H264 decoding. It will be a little more sluggish than the Kaby Lake or later version of Quick Sync in the 2017 or later iMac. The same Browser vs timeline difference exists with Quick Sync, but it is lessened.

If you are not on the latest version of macOS, FCPX and Apple Pro Video Formats, it is good to upgrade. Late last year there were some significant improvements in H264 handling unique to the iMac Pro. However it appears this was mostly encode-related, not decode.

It appears the Panasonic material uses a fixed GOP size whereas the Sony XAVC-S material uses a variable GOP size. My guess is the variable GOP encoding is more difficult to smoothly decode. This is not unique to FCPX; Premiere is also much slower on 4k XAVC-S.

I tested a 4k XAVC-S clip, then exported it from FCPX using Master File>Computer>H264>Best Quality, then re-imported it. The timeline performance was smoother. Inspection of the header showed it was fixed GOP, size=30, vs the Panasonic material of size=8.

I then encoded the same file using Compressor and selected GOP size 8. When imported to FCPX the file was yet faster and smoother.

I also tested re-wrapping the XAVC-S with EditReady before import and importing the clips using “copy to library” vs the bare files extracted from the camera’s folder tree. I didn’t observe a major difference.

There is nothing you can do about the difference in skimmer latency between the Browser and timeline. That exists with various codecs but is more noticeable on “difficult” variants of H264 such as Sony XAVC-S. Your only other options are transcode the XAVC-S to proxy or change to cameras which emit better-performing H264.

[Joe Marler]

[Tony Sarafoski] “there seems to be no way I can switch the original media over to the proxy media that KeyFlow has generated inside KeyFlow.

I would like to take my edits on the road from time to time, and so far I haven’t been find a way where I can switch between Proxy or Original Media files, similar to what is possible within FCPX.”

I don’t use KeyFlow Pro, but it’s not possible to relink proxies in FCPX, nor to attach externally-generated proxies as proxies within FCPX.

Examination of the internal SQL tables within the library indicates they are not storing full metadata for each proxy file, but adding a few rows for each proxy file which “piggyback” via FK<->PK relational links to the corresponding orig. file metadata. FCPX proxies and the underlying database methods and schema were originally designed for in-library use only. It’s possible that when user-specified external storage locations were added in 10.1.2 (which enabled external proxies), the database methods and schema were not adequately updated.

However it is possible to manually relink to the proxies as original media. If the proxy files exist on any disk volume, FCPX can often relink to these as original media — not proxies — despite lower resolution or differing codec. Then when final export is required, relink to the full-res originals is possible. In this scenario proxies are used but the FCPX viewer is set to “original media”. The proxies can be FCPX-generated or camera-generated.

I don’t know if that method might be applicable to KeyFlow Pro.

Relink constraints:

– Files must have the same audio config (same number of channels), but sample rate can differ.
– Pixel aspect ratio should be the same, but it may relink even if different.
– Clip duration must be the same.
– File suffix must be the same.
– Codec need not be the same, e.g. you can create 720p H264 proxies and relink to those as original media. However for normal FCPX-generated proxies of 4k material, the codec will be ProRes 422 at 1080p.
– Camera-generated proxies are all different cases and relink to those and back to full res media should be tested for each situation.
– After relink to lower-res material, the viewer may show a windowboxed frame. This is usually a cache issue. To fix this shut down FCPX, manually delete cache and start it again. Cache will be located in a folder or the library, as shown in the library inspector. The cache is a single file bundle named LibraryName.fcpcache. It seems harmless to delete, but do so at your own risk.

[Joe Marler]

[Greg Ball] “I’m not looking to select all clips. I just want to select ONE clip to see the length of that individual clip CTRL+D has ALWAYS worked for me since the beginning of FCPX. Suddenly it’s not working.”

Sorry, I mis-read that. But on 10.4.6 CTRL+D to display/change clip length still works for me.

However I usually don’t use CTRL+D to get clip length, since the length of any selected clip is automatically shown in yellow text below the viewer and to the right of the project name. You can jump the selected clip forward/backward with CMD+right arrow or left arrow.

The displayed length units (HH:MM:SS:FF or frames or fractional seconds) is determined by FCPX>Preferences>Time Display.

[Joe Marler]

It works OK on 10.4.6. You have to select all clips in the timeline via CTRL+A, then do CTRL+D. If they are not all selected it won’t work.

[Joe Marler]

When proxy creation “fails without warning”, does it just stop, or is there any kind of error message? Does FCPX require restarting or is it still up?

It could be a bug, but I’m suspicious it might be a space consumption issue.

Is the 900GB free space on an external drive? What is the free space and % of free space on the system drive?

What FS7 resolution, frame rate, bit rate and codec are you using?

Is the FS7 material being successfully imported with “leave files in place”? Or is it being copied to within the library? If copied within the library, then proxies are created, this will take more space than RED files imported using “leave files in place” plus proxies. OTOH if FS7 proxy creation succeeds when done in small batches, I don’t understand that.

You could try running Disk Utility First Aid on all drive volumes to make sure space reporting is updated and there are no file system anomalies.

Unfortunately I don’t know of a continuous free space monitoring tool for Mac. Unfortunately this is an area where Windows is far superior, since the built-in PerfMon tool can continuously graph and monitor disk consumption and even send alerts based on user-defined thresholds.

However on macOS it’s easy to write a simple shell script in terminal to continuously record space c, which could be redirected to an output file for later perusal. Below uses just 10 iterations; to do more increase the count.

See various references for using terminal commands. There are lots of free ones but this book is pretty good:

https://www.amazon.com/dp/B00XV0ZZBA

#!/bin/bash
# Display space consumption at 1 sec intervals for $count iterations

count=1
while [ $count -le 10 ]
do
sleep 1
date
printf “Space on MacHD: ”
df -H /volumes/Macintosh\ HD | sed 1d
printf “Space on RAID array: ”
df -H /volumes/TBay32TB/ | sed 1d
printf “\n”
((count++))
done

[Joe Marler]

[Eric Santiago] “The RED never fails at proxy after its been imported as “link”.
However, the FS7 files need to be manual.
Cant select a huge list and create proxy.”

I’ve created proxies on over 8,000 4k XAVC-S files in one batch; it worked OK.

Proxies must go somewhere, either in the library or in a storage location specified by the library inspector under storage locations “modify settings”.

The most common cause of proxy creation failing is running out of space. When you select a large group of files, FCPX will estimate total space requirement and may abort. If short on disk space, creating proxies for a few files might work.

Check (1) Where your proxies are being created and (2) Free space available on that disk volume.

To create proxies in a user-specified location, after importing all media with “leave files in place”, change the library storage location to a different folder. Proxies created after that will be placed in that location. For more details see Ripple Training’s 10.4 media management class:

https://www.rippletraining.com/products/final-cut-pro/media-management-in-final-cut-pro-10-4/

[Joe Marler]

[Oliver Peters] “I would presume that if you wanted to add your own raw drive(s), you would have to purchase the MPX chassis from Apple. Potentially there will be third-party MPX modules available. And, would you even be able to add a 3.5″ spinning hard drive or will you be limited to 2.5” SSDs or only NVMe drives?
“

The Pegasus J2i appears to be a simple sled for two 3.5″ SATA HDDs. The announcement said it will use a “custom cable assembly”, getting data from the Mac Pro’s SATA header. While not mentioned, there are 6-pin PCIe power headers on the motherboard. It might get power from a PCIe 6-pin to SATA female cable, maybe something like this: https://images-na.ssl-images-amazon.com/images/I/61iKLfXQl4L._SX425_.jpg

The J2i appears to be just a piece of sheet metal, so that would be the “old school” way of doing it – power and data via cables. It would also occupy the vertical space of an MPX module while holding only two 3.5″ drives.

The R4i MPX module is apparently “cable free”, getting all power and data from the PCIe slot and MPX extension slot. It appears to use 4 x 8TB SATA 3.5″ HDDs. The MPX extension slot also has Thunderbolt access, so I don’t know if Promise is using a PCIe host controller (more likely) or Thunderbolt. Also unknown is how many internal Thunderbolt busses the new Mac Pro has.

In theory you could swap out the R4i’s 8TB SATA hard drives with 4TB SATA SSDs in 2.5″ adapters. There are even 8TB SATA SSDs now available but they are expensive. Or if not compatible, some other mfg. could make an MPX module designed for higher density 2.5″ SATA SSDs. However SATA isn’t the best interface for SSDs, just saying it’s theoretically possible. But conceptually two MPX modules each with 4 x 8TB SATA SSDs would give 64TB SSD internal capacity, plus the 4TB on the motherboard.

PCIe-based NVMe cards are also available. The ASUS Hyper M.2 x16 Card V2 supports up to four PCIe NVMe daughter cards, for 8TB — per board. I don’t know if it’s compatible with the new Mac Pro but it illustrates the form factor and capacity exist for substantial PCIe-based SSD storage:
https://photos.smugmug.com/photos/i-7CT8Knh/0/6af36959/L/i-7CT8Knh-L.jpg

Beyond this, Samsung has 8TB PCIe SSDs in a M.2-size but new NF1 form factor. If a compatible PCIe card with 4 x NF1 sockets were made, each PCIe slot could hold 32 TB SSD: https://www.samsung.com/semiconductor/insights/news-events/samsung-introduces-8tb-ssd-in-nf1-form-factor/

A maxed-out Mac Pro with 1.5TB RAM and stuffed full of 8TB PCIe SSD cards would be expensive. However the NF1 form factor SSDs will be widely used in data centers: https://www.samsung.com/semiconductor/insights/tech-leadership/server-scalability-enhanced-with-nf1-ssd/

If an MPX module was designed to accept the above 8TB NF1 cards for perpendicular insertion (see above photo), at 5mm width per card each MPX module might hold 30 8TB cards for a total of about 240TB. This of course assumes adequate cooling, power and bus bandwidth.