[Joe Marler]

[Jeff Leland] “if I set the preferences to leave media in place, and consolidate that media into the library, I use NO additional disc space – even though I see two additional folders created that seem to copy the original media as well as optimized media into the original media folder – correct?”

Hard links will NOT be used when *importing* media on the same volume within a library or to an external location managed by that library. This is even for cases where no re-wrapping takes place. Additional space will be consumed.

Hard links are only used when FCPX is copying media from one managed location to another on the same volume. E.g, copying media within FCPX from one managed “internal” library to another, or from a managed “internal” library to a library with external media defined in Library Properties>Storage Locations, or between two libraries using external media defined in Library Properties. Note that a library using external media is different from a library where media was imported with “leave files in place”.

This can be verified by using the console commands ls -i (to see file’s inode) and ls -l ( to see files’s reference count). If the inode is the same, it’s actually a single file with multiple hard links pointing to it which gives the appearance at the Finder level of multiple copies of that file. Likewise ls -l will show the reference count, which increments once for each hard linked “copy” of the file.

The file’s reference count will increment if copying data within FCPX between managed libraries (showing it’s really the same file using multiple hard links). Even though Finder shows two different media files in that case, there is only one physical file and no more space is consumed. Also ls-i will show the two files having the same inode.

The reference count will not increment when importing media, showing it’s two physical files, each taking up space. The two files have different inodes.

So the idea is incorrect that hard links will avoid additional space consumption if importing media on the same drive volume.

[Joe Marler]

[Patrick Jones] “I noticed that the Render was crazy fast, and it was punhy. Although I still hate with a passion the magnet timeline, and it does slow me down on my creative process, it was sort of worth it.

6 months in, and I feel it’s not rendering as fast.. Nothing changed.

I still work off my RAID 5. Have over 50% available. My Libraries are under 10 Events (usually 3 Project with each Library). Should I upgrade my Interal to an 1+TB SSD”

I’ve edited a 20TB 4k H264 documentary on a 2015 iMac 27 like yours, using proxies, with material on a Thunderbay 4. It worked mostly OK, and that was without using background rendering.

When you say “render was crazy fast”, then “it’s not rendering as fast”, do you mean export or rendering the timeline with CTRL+R? If export, what codec, and what is your editing codec?

Have you added any compute-intensive effects? Those will slow things way down and have little to do with I/O.

You can check your machine and software config by simply creating a test library, importing a 4k/30 H264 test clip, add no effects or edits, then exporting that as 4k/30 H264 this way: File>Share>Master File>Settings, Format: Computer, Video codec: H.264 Faster Encode, Resolution: either 4k or 1080p.

If that is as fast as expected, then maybe it’s something you’ve done to the production timeline or library. As edits gradually accumulate in a timeline, things can slow down. To some degree that’s normal.

If you were previously editing 4k on a 1080p timeline, then switch to a 4k timeline, the render files will become 4k so certain things will be slower. You can usually edit 4k on a 1080p timeline, then if you need 4k output, switch the project characteristics back to 4k before you export.

However for 4k H264 the 2015 iMac is not as fast as the 2017. Apparently due to improvements in the Kaby Lake version of Quick Sync or in how FCPX uses this, it’s about 2x faster at H264 operations, including timeline scrubbing and export.

[Joe Marler]

[greg janza] “My PC system is relatively contemporary even though my CPU is a few years old I have a ton of ram and I work off of NVMe drives. I tested a mxf wrapped 4k H264 file shot on a CANON, EOS C300 Mark II. Premiere has no issues with this file at all in full res much less 1/4 quality. JKL is without hiccup. It also makes no difference whether “Enable hardware accelerated decoding” is turned on or off. Premiere performs identically.

I’ve also worked on the new imac pro. It’s a very nice computer and it’s very fast. “

Well, that’s very interesting. On most of the 4k H264 codecs I’ve tested, Premiere 2018 timeline operations are quite sluggish on both my 2017 top-spec iMac 27 and 10-core Vega 64 iMac Pro. This includes XAVC-S material from the Sony A7RIII, UHD 4k from the Panasonic DVX-200, and many other cameras.

Even FCPX on an iMac Pro is no speed demon on this material, but it’s much faster than Premiere on any Mac I’ve tested. However even with FCPX I usually find it necessary to transcode to proxies, and when those are used in Premiere it also is lightning fast.

If you were using XF-AVC Intra, an NLE can handle that much faster than the Long GOP version. I tested some All-Intra 4k H264 material from a Canon XC-15, and it was very fast on both Premiere and FCPX.

[Joe Marler]

[Kyrina Bluerose] “…There were just way too many clips to work with to multicam them all, so it seemed to make logical sense to make a project with each group, and then make the final product using “highlight reel” from the others….I’m still at the point I’m trying to make the program do what I want it to do, since I don’t yet best know how to work within the program!”

I’ve done documentaries with over 150 multicam interviews and many other multicam b-roll sequences. In general don’t use projects (aka sequences) as an organizational tool. It’s better to use favorites, rejects and keywords on the MC clips themselves, not on the parent clips. In fact after making the MC clips it’s best to reject all the parent clips and always have the viewer in “Hide Rejected” mode.

This greatly de-clutters the Event Browser and avoids accidentally using a parent clip in the timeline, which then requires manual back-tracking to find that range in the MC clip.

There is a limitation in the current FCPX keywording/rating system for MC clips in that you can’t state what camera angle. You assign a keyword or rating to a clip range, which may include multiple angles. However to most editors it will be visually obvious why that range on a MC clip was marked.

The Event Browser cannot display clips in a specific order, so it can’t be used as a “string out” substitute. However you can easily define various “highlight reels” by using keywords on various clip ranges. Then those reels can be added to a project in whatever order you want, and the clip order re-adjusted.

The overall goal is do as much of your organizing within the Event Browser as possible, then (and only then) start adding content to a timeline. Don’t begin your organizing by dumping a bunch of stuff on several timelines and copying between them.

[Joe Marler]

I don’t understand that video. He showed the PP 2018 12.1.2 update notes saying H264 *decoding* is supported on the iMac Pro and MacBook Pro. Yet he is editing C200 RAW which has nothing to do with decoding H264 — that is a separate code path.

Not related to his issue, but in PP 12.1.2 on my iMac Pro running macOS 10.13.6, I can turn Premiere>Preferences>Media>”Enable hardware accelerated decoding” on/off and I can’t tell any difference. They are both slow, much slower than FCPX or Resolve on the same hardware. In general PP has always been slow at 4k H264 decoding. As of 2018 it is vastly faster at *encoding*, ie exporting to H264, but decoding is still very sluggish on Macs.

I’d be interested if someone with PP 12.1.2 and a contemporary Windows machine could try JKL timeline responsiveness at 1/4 res for a 4k H264 timeline with “Enable hardware accelerated decoding” both on and off, and observe any difference. It is conceivable H264 decoding works *better* on Windows than Mac. However this wasn’t his main issue — he got side-tracked mentioning it, and it has nothing to do with C200 RAW decode performance.

His issues were:

(1) PP 12.1.2 timeline perf. for C200 RAW is much slower on his Windows PC vs iMac Pro. That is interesting, but (to my knowledge) RAW cannot be accelerated by Quick Sync, AMD’s UVD, etc. Maybe some aspect of RAW conversion such as de-bayering is better implemented by some Apple framework that Adobe used on macOS. I think Apple did some performance work at the AV Foundation layer to facilitate RED RAW performance — that was well demonstrated when the iMac Pro was released. Maybe some of this carried over to other RAW formats. However I don’t think it’s because Adobe developers are “prioritizing the Mac platform”.

Every RAW conversion pathway is unique, just like it is for stills. For stills we are accustomed to Adobe Camera RAW or other converters handling this, but for video it’s not as standardized. You can’t grab a new camera, shoot some RAW video and expect all NLE software to immediately work perfectly. One hope for ProRes RAW is it would unify RAW post processing, and when a new RAW camera is released it would more likely work immediately (at least in FCPX). By “work” this means both function plus provide consistent performance.

(2) He reported slow export performance of C200 RAW to (H264?) on his Windows PC vs iMac Pro. He wasn’t specific what export format, whether ProRes or H264, but he said the iMac was faster.

(3) He reported slow export performance of Sony XACV-4k from PP 12.1.2 on Windows PC vs iMac Pro. At 05:22, it appears he’s using the 2160p Youtube preset for export. I have tested XAVC-S 4k export performance many times between PP 12.x and FCPX on an iMac Pro, and Premiere is often faster, as of the 2018 version. After seven years of not using Quick Sync or AMD’s VCE, Adobe is finally using this and it’s fast. They are still slow on the decode side, IOW for timeline operations but for export it is (at last) fast. However Premiere has many different export options and it’s easy to accidentally invoke one which slows things down.

I don’t have a high-end Windows machine to test, but I personally doubt that PP 12.1.2 is slow exporting 4k XAVC-S to single-pass VBR 4k H264 at 30 mbps. For that task Premiere 2018 is faster on my 10-core iMac Pro than FCPX 10.4.3 is. In fact Apple really needs to get their act together on FCPX 4k H264 encoding on the iMac Pro — that one task is slower than a 2017 iMac, likely because their use of AMD’s UVD/VCE decode/encode hardware needs optimization.

[Joe Marler]

[Eric Santiago] “With a drive path column, I can at least go back to the original Library (I kept a copy) and see where each of these files is actually sitting.”

In the library inspector, FCPX itself will tell you what drive volumes media files are on, just not the paths to those files on a single screen.

Final Cut Library Manager from Arctic Whiteness will export a .csv file showing the path to each media file: https://www.arcticwhiteness.com/finalcutlibrarymanager/

Provided the file names don’t change you can normally relink media files after moving them to another drive.

So one method is using Library Manager to find all locations of media files for a library, then move the files from the various USB drives manually to a single Thunderbolt array, then re-link. This assumes all external media.

The other method is creating a new library on the target drive, defining a new media storage location on that drive, drag/drop the events within FCPX to the new library, then consolidate. However that will create a different media tree structure and may also rename media files if there are naming conflicts.

You can also have a mixed situation where some media is inside the library and other media is outside the library. E.g, you imported some media with “leave files in place” and other media was copied to the library. I think in that case copying to a new library storage location on the new target drive and consolidating is the best method.

[Joe Marler]

[Anne Davis] “The interviewee is not happy with the end result and I’m hoping to be able to soften the focus either within FCPX or with an external plug-in or program.

I’d appreciate any suggestions of filters, etc. that might be helpful to achieve this result.”

As Jeff said, Digital Anarchy’s Beauty Box is OK. However the best skin processing plugin is Imagenomic Portraiture. It was long the best such product for stills, but is more recently available for video. Unfortunately it is expensive but it’s simply the best. It is very compute-intensive to run, so if you are processing a 5 min interview clip that will take a while, even on an iMac Pro.

As with all compute-intensive plugins such as Neat Video or Digital Anarchy Flicker Free, it’s best to defer running it until editing is almost finished. Since some trial and error may be involved in getting it “dialed in”, it’s best to blade a small range and run it there. Then once you finalize the settings run it on the entire clip. That will take a while.

It can produce a very natural looking result, not some artificial softened look.

https://imagenomic.com/Products/VideoSuite

[Joe Marler]

When relink fails it doesn’t give you much info. However if you pick only one file (which failed previously) and relink that, it will give more specific info about why it failed. Try that, then examine the file metadata on the source and destination machines using MediaInfo, Invisor, etc.

The library contains metadata about each media file. During relink this is compared to the metadata read from the disk file being relinked. Unfortunately I don’t know if an easy way to examine the stored metadata, except to export an XML, open it in TextEdit and search for the filename. That will lead to something like this:

<asset id="r28" name="C0099_28672" uid="F51E6A6ED48F2F90D9A86998A9001A05" src="file:///Volumes/TBay8TB_SSD_True/CMA/CMA060518/HarleyA6500/C0099_28672.MP4" start="0s" duration="6186180/60000s" hasVideo="1" format="r27" hasAudio="1" audioSources="1" audioChannels="2" audioRate="48000">

If you export an XML on your destination machine, then for a problem file compare the metadata within the XML from both source and destination XMLs, this will show whether the file metadata (from FCPX’s viewpoint) is the same or not. If it’s not the same the question is why not. In that case the next step is examine the source and destination media files themselves using MediaInfo, Invisor, etc. If they are the same, the question is why is relink failing.

BTW the latest version of Final Cut Library Manager can produce a .csv list of all media files used in a library, including pathname, but it doesn’t list metadata. It would be nice enhancement if it did.

If you are using external proxies, ie proxies generated by FCPX but stored outside the library, relink will be unreliable in that case. That is a known problem I filed long ago.

Are both media sets on both machines on HFS+ partitions? If one is a portable exFAT hard drive, maybe that’s a factor.

[Joe Marler]

[Trevor Asquerthian] “Some folks play directly to air from NLEs every day. Some are more reliable than others ;-)…”

There is no limit to what people will do, but that doesn’t mean it’s a good idea. I personally would never play back material from any regular NLE — even at a film festival — much less for broadcast. That includes FCPX, Avid, Premiere, Vegas, Edius, Resolve, etc.

In case other readers didn’t catch it, the EVS IPEdit product you’re using is not a traditional NLE. It is a broadcast “playout system” with some editing features: https://evs.com/en/news/evs-introduces-play-while-editing-ipedit

Broadcast playout systems often have ability to reorder playlists, edit head and tail of upcoming clips, etc. It appears IPEdit had more extensive features in this area.

Broadcast playout is a highly specialized area. The product management of any traditional NLE has a list of features to consider for upcoming versions. Each of those new features entails technical risk, development cost, test cost, support cost, and projected payback. I seriously doubt any of the leading NLEs are prioritizing concurrent editing during broadcast playout.

It’s possible you could do your own “science project” and cobble together something using the collaborative editing features of existing NLEs. However they are really not designed or tested for broadcast playout. You’re probably better off going to NAB and making inquiries, since this is a more broadcast-workflow issue than a NLE issue.

There are specialized apps for professional media playback such as DT Videolabs’ Playback Pro:

https://www.dtvideolabs.com/dt-videolabs-playbackpro-professional-video-playback/

For larger film festivals or stadium-size audiences, playback might be two redundant computers, each running specialized playback software, and a production video switcher handling failover. For broadcast, the term is “playout” and it would normally be file-based playout servers such as this: https://360systems.com/pdf/maxx_2420ex.pdf

This new Ross playout server is another example: https://go.rossvideo.com/tria-news-playout-server

Clutch Broadcast PlayOut Server: https://www.clutchautomation.com/clutch-playout.html

A cheaper Mac-based playout server: https://renewedvision.com/provideoserver/

This one is Windows-bases software, but you must add your own I/O cards: https://www.bhphotovideo.com/c/product/1016995-REG/deyan_automation_systems_aslite_single_channel.html

[Joe Marler]

[Trevor Asquerthian] ” I need to be able to play out a timeline whilst I am making edits.”

This is possible in FCPX, see below 60 sec demonstration. However it is designed as a productivity enhancer when editing, not a mechanism to play a timeline directly from the NLE to broadcast while editing the same timeline.

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

While Thomas Grove Carter makes it look very impressive, there are several limitations which rapidly appear. E.g, you can cut the timeline and do a slip edit behind the moving playhead, make that a compound clip and watch it appear — all while playback continues. But you can’t play that new clip or edited portion of the timeline while the main playhead is moving (even with Window>Show In Workspace>Event Viewer enabled) because FCPX only supports a single active playback stream.

In theory within certain limits you could edit a timeline in FCPX while playing direct to broadcast. But in general you don’t want to play from *any* NLE direct to broadcast — with or without concurrent editing.

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.