[Joe Marler]

[Grant Peacock] “then importing to Final Cut 10.4.6, every effort at generating a multicam clip causes bizarre results. (using audio or timecode as the basis of the generator doesn’t improve things)”

Before creating the multicam clip, make sure all material from each camera is labeled with a camera name or angle name in the inspector “info” tab. Select all clips from one camera then enter a name; repeat for the other. Then try sync by audio.

If you aren’t sure which clips came from each camera, during import they can be automatically keyworded if you named the offload folders something unlike. E.g, C300HouseLeft, C300HouseRight, etc. That streamlines selecting them and labeling material from multiple cameras with similar codecs or filename conventions. Use FCPX Preferences>Import>”Keywords from folders”.

[Joe Marler]

[Brett Sherman] “the question is whether there is an advantage for the Mac Pro versus iMac. I calculate that you can get 3 iMacs during the time you use the Mac Pro (depends on if you sell or how you calculate the value of passing the iMac down to another use). I figure this is around an 8-10 year time frame. So over that time, how will the capability of the iMac improve? “

This is an excellent question, also with regard to the iMac Pro vs Mac Pro. You must consider the resell value of your iMac or iMac Pro when every 3 yrs you upgrade to a new one, and compare that to the upgrade cost of the Mac Pro.

There are cases where the available GPU horsepower of the new Mac Pro is very useful. I just spent several weeks finishing a doc with lots of contributed low-light GoPro footage. It had to be used, as there was no other alternative. This required a vast amount of Neat Video processing.

The very latest version of Neat Video is about 2x faster, and the internal sizing tool picks an optimized mix of CPU, CPU+GPU or all GPU. For my 10-core Vega 64 iMac Pro it was all GPU. Despite all that it was extremely slow. In my workflow that is very rare but if I did that a lot I’d consider a Mac Pro.

The iMac Pro has lots of ports – 4 Thunderbolt/USB-C, 4 USB-A, and 10-gig E. I need that frequently when I’m working collaboratively and must connect to the NAS and offload a stack of SSD cartridges.

The iMac Pro is actually quite portable. I use this Gator case which works well, although I wouldn’t check it as luggage on a plane: https://www.amazon.com/dp/B07QVMRGJ3

The iMac Pro is due for an upgrade and there are theoretically chips available, but as usual we don’t know anything. If an iMac Pro was upgraded with the new Xeon and a significantly upgraded GPU I might get one.

https://www.macrumors.com/2019/10/07/intel-xeon-chips-imac-pro/

[Joe Marler]

[Eric Santiago] “I have to do a quick case study on the advantages between the 2013 Mac Pro D700s vs 2019.”

Any new Mac will be vastly faster at H.264/HEVC codecs, since the 2013 trash can doesn’t have hardware acceleration. Those aren’t editing codecs but you often must handle those for ingest.

The GPU horsepower of the new Mac Pro isn’t just for VFX work. If your workflow involves GPU-intensive plugins like Neat Video (or similar Resolve tasks), it can make a big difference.

[Joe Marler]

[Giuseppe Mangione] “I thought to change the speed of the clip. But I didn’t expect that inserting a 25fps clip in a 24fps timeline the clip speed change automatically. I notice that in the bottom of the video inspector, it says “25 converted to 24”

Anytime a differing frame rate is added to an existing timeline, something must be done to “rate conform” the added footage. There are various methods depending on the frame rates involved. E.g, when adding 30 fps footage to a 24 fps timeline, frames are usually discarded. Even in that case the rate-conformed clip length may be slightly altered, because the “pulldown” or other conforming method is done in small groups of frames. I think this happens if the clip isn’t an exact multiple of that group size. That would create border cases where they’d have to discard an end frame, which should be avoided.

In the 30>24 case picking “automatic speed” on the 30 fps clip will cause a 20% slowdown.

For 25>24 fps, the rates are so close that the main method to rate conform them is slow down the material by 4% and correct audio pitch. While this is internally achieved via a speed change I suppose it’s not classified as such, as other algorithms are theoretically possible but rarely used. E.g, you could hypothetically analyze and reconstruct every frame using optical flow methods which discards some frames but blends the difference across the entire clip to produce a frame count matching the 24 fps. There might be highly specialized rate conforming software or hardware which does this for major film companies, but to my knowledge no NLE does it.

[Giuseppe Mangione] …And the speed of the clip is changed (from 10 second to 10:10) In the retime editor it should be 96%, not 100%! This is what surprises me. And If you make “automatic speed” it reduce the speed of another 96%. The clip of 10 second became 10:20!”

The FCPX UI seems to separately classify retiming and rate conforming – even if sometimes the underlying technique is similar. E.g, if you add a 25 fps clip to a 24 fps timeline, then in the video inspector under Rate Conform pick “optical flow”, the Modify>Retime>Video Quality>Optical Flow shows a greyed out option. Then if you slow the clip by 50%, then optical flow retiming becomes available.

You could probably argue maybe the UI should more clearly indicate what is happening under the covers but it’s widely known that 25/24 rate conforming almost always uses retiming with most NLEs. If the UI clearly showed a colorful retiming bar on the clip this would probably confuse people.

[Joe Marler]

Make sure you are not running the Chrome browser. It may saturate the MacOS VideoToolBox framework, locking out FCPX and causing very sluggish and unpredictable behavior.

[Joe Marler]

[Giuseppe Mangione] “Can someone explain me why if I insert a 25pfs clip long 10 sec. in a 24pfs timeline the clip cover 10 second and 10 frames of the timeline?

And If I insert a 120pfs, or a 29.97fps, clip long 10 sec in the same timeline it still cover 10 second?”

The best way to rate conform 25 to 24 fps is via a speed change, IOW to slow the 25 fps material by 4 %. That is what FCPX does and I think it automatically does audio pitch correction. The other option is discard a frame occasionally which would cause a motion cadence issue.

120 and 29.97 are essentially evenly-divisible multiples (within 0.1%) so discarding 75% of the 120 fps frames produces a perfect frame-rate match to 29.97 (essentially 30 fps).

Larry Jordan article on frame rate conversion: https://larryjordan.com/articles/frame-rates-are-tricky-beasts/

[Joe Marler]

[Andrew Johnstone] “Good to hear that FCPX does manage very large project (library files) – all be it on a much speedier/hefitier machine…”

I edited the original 20TB library using a top-spec 2017 iMac 27, however an iMac Pro is really better for tasks of that size.

Basic guidelines:

– If 4k H264, it will generally require transcoding to proxy for optimal editing performance.

– If 4k ProRes it can probably be edited without proxies, if the I/O is fast enough and if on a 2017 or later iMac 27.

– Use media imported with “leave files in place”. If the camera media is tree-oriented and doesn’t allow this type of import, consider re-wrapping (not transcoding) with EditReady before import: https://www.divergentmedia.com/editready

– Do not ever import bare .mts files from an AVCHD package using “leave files in place”. This can create a large I/O burden due to how FCPX handles that internally.

– Use FCPX library inspector Storage Locations>Modify Settings to place cache and storage of generated media external to the library itself. This includes proxies and optimized media. Doing so keeps the library small and facilitates easy file-level backup by duplicating the library with Finder.

– Use only HFS+ or APFS external drives.

– Do not copy clips or projects between libraries unless they are inside a “transfer event” created for that purpose. Otherwise spurious duplicates can occur.

– Efficient post production starts in production. Make sure all cameras and recorders have the correct time of day, and re-verify this each production day. If shooting multi-cam try to use a slate. If doing stop/start multicam, consider using timecode devices like Tentacle: https://tentaclesync.com

– Similarly, try to ensure all cameras are using the same frame rate, and this matches the expected frame rate for editing and final distribution. Even with modern optical flow rate conforming, mixing non-multiple frame rates like 23.98 and 29.97 is not perfect. In a perfect world all cameras on a given production would shoot even multiple frame rates such as 24/48/72 or 29.97/59.94/89.91. Admittedly this isn’t always possible.

– Ideally before importing any media, make sure all camera filenames are *globally* unique across the entire production. Not doing this can cause spurious duplicates if XML is loaded, also having a unique filename greatly streamlines post production turnover. My doc team simply appends a 5-digit unique serial number to each offloaded file using the tool “A Better Finder Rename”: https://www.publicspace.net/ABetterFinderRename/index.html

However Finder itself has batch rename ability as shown in this brief tutorial on MacMost Now: Finder itself has batch rename ability. This short tutorial explains how to use it: https://macmost.com/step-by-step-using-the-mac-batch-rename-tool.html

– Do as much work as possible in the Event Browser before adding material to a timeline. Fully leverage FCPX database features; try to avoid thinking in terms of bins. This video “One Smart Collection To Rule Them All” shows one possible approach:
https://youtu.be/sjuCfJFhdo0

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]

[Andrew Johnstone] “I am about to start work on a feature doc and much of it will be shot in 4k. With that in mind I am looking for advice on max library file sizes for FCPX.”

I have edited a large documentary on FCPX with 8,500 4k H264 clips in a single library. This included 220 camera hours, about 130 multi-camera interviews, occupying 20 terabytes. Most of the media was in just three events. Machine was a 2017 iMac 27. The media was on two redundant 32TB OWC Thunderbolt RAID arrays. Actual library file size was small, about 120MB since all media, proxies and cache were external to the library.

In general the performance was OK but I was using proxies. FCPX proxy management can be difficult since they are tied to the original disk volume name where they were created, also proxies cannot be relinked. However they work OK.

An alternative approach is relink to the external proxies as original media, edit, then relink back to full-res media for final conform. There are several relink constraints and this should be tested on the specific codecs and file types before committing significant work.

If using lots of effects an iMac Pro is better because it is quieter under sustained high CPU/GPU load. A more recent doc I worked on used an approx. 2 hr timeline and had many hundreds of effects, including about 50 clips with Neat Video. It took several hours to render on my 10-core Vega 64 iMac Pro. Media in that case was about 14-16 TB, spread across a 4-drive SSD array and several smaller SSD drives. Media, proxies and cache were all external to the library, using “leave files in place” import and other media placed in designated locations using the library inspector>Storage Locations>Modify Settings.

[Joe Marler]

[Chad Greene] “Joe, if I read your email correctly, I believe you are suggesting we start with our 4K footage, then relink to Proxies for editing?

No, I was just explaining one way it might have happened. It is understandable you’d want to start with proxy resolution in this case.

[Chad Greene] We do set our Project setting manually so they are correct.

In limited testing I’ve done it seems sometimes relinking to higher res media changes the project resolution and sometimes not. Maybe it doesn’t change if going from NTSC material to higher res, because it would be considered a category change. When relinking from FCPX-encoded 854×480 23.98 progressive to 1920×1080 23.98p, it seemed to change.

[Chad Greene] …The event media and the video in the timeline are both showing the proxy frame size even after relinking to the original media. This small clip frame size is what I want to reset. Any ideas how I can do that?”

As I said, you can copy/paste all clips in the low-res project to a manually-created 4k project. You can relink before doing this or afterwards.

[Joe Marler]

Instead of starting with 4k resolution, relinking to externally-generated proxies, editing, then relinking back to 4k, IF you started with proxies this will fix the timeline at the settings of the first clip added (640×360 in this case), if using an “automatic settings” project.

Simply relinking the clips back to 4k will not change the project characteristics. This in turn limits what resolution you can export.

You will need to create a new 4k manual project and copy/paste the entire timeline over. CMD+A to select all clips in the old timeline, CMD+C to copy, switch to new timeline, CMD+V to paste.

If all clips are the same frame rate and pixel aspect ratio this should work. If the proxy clips have different frame rates this could cause shifting of some edit points due to rate conforming issues. Larry Jordan mentioned this here: https://larryjordan.com/articles/fcp-x-conform-different-video-frame-rates/

Externally-generated proxies can work, but each case is unique. It’s best to first test it on a small dataset, scale up, then re-test the end-to-end workflow using all representative codecs and media types before committing to full scale production.

If there is any doubt about the video characteristics of the external proxies, this can be checked with a tool like Invisor. This allows side-by-side spreadsheet-like comparison of multiple files. E.g, you want certain key items like pixel aspect ratio and audio parameters to match up between your proxies and the original media: https://apps.apple.com/us/app/invisor-media-file-inspector/id442947586?mt=12