You may not want to read the following if you require a direct thunderbolt 3 connection. For me, I needed to put the machine somewhere quiet and so 10GBe was the way to go.The first thing I did was buy Sonnet's Solo 10G (Thunderbolt 3 to 10GBe) which is rock solid btw in Catalina - in comparison to Apple's built-in 10GBe ports that cause disconnects upon heavy load. The latter was also the reason I wasn't upgrading to Catalina all this time. And Catalina has been a big upgrade on SMB 3 speeds (and a downgrade to many other things!). However up to Mojave I could only get around 800MB/s Read/400MB/s Write, where as with the same hardware and Windows (bootcamp) I could get 1000MB/s Read. Catalina optimised this and I get 980R/600W. Nowadays I don't force spotlight indexing by the OS. They had added built in support for spotlight search (only files, not folders), but since late Mojave this broke, and QNAP haven't addressed it yet although multiple tickets have been opened the past year.
Because Apple is dropping the ball constantly these days. Are you using native only plugins or red giant/neat etc.. as well? 10.4.4 introduced this re-rendering bug and have been going mute on the feedback forms. They used to reply once in a while, now they run on auto-pilot not fixing any bugs that show up. Report it as I have and hope for the best.
For anyone following this, the slowdown issues seem to be resolved in the latest macOS 10.13.4. I have been in contact with both QNAP and Apple and before today the issue was ongoing. We can now work on the system again with multiple users. For reference I will post the configuration that is working for me:
FCPX 10.4, macOS 10.13.4
MacBook Pro Retina 2017 15" Maxed out
QNAP TVS-1282T3 firmware 4.3.4.0537 Build 20180331
QNAP's 40Gbps 2m Thunderbolt 3 cable
8x8TB HGST NAS 7200rpm drives
1 Static volume (Qsirch on, spotlight indexing on, multimedia apps off, network trash off)
Mounting via SMB 3
iMac 5K Retina & Mac Pro 2008 (with dosdude's patch to install 10.13.4 and an NVIDIA GTX 960) connected via 1Gbps ethernet for now. Waiting on NAB for new 10Gbps adapters.
veto files = /._*/.DS_Store/
Disabled DSStore files on network
defaults write com.apple.desktopservices DSDontWriteNetworkStores true
Hey everyone from Athens!
A problem with thunderbolt 3 and FCP X is driving me nuts and was looking towards your wisdom and experiences to help track this issue down. Mounting the NAS via Thunderbolt 3 (40G cable/SMB 3) and trying to create/name/delete a new folder inside an FCPX event, or even attempting playback (normal HD H.264 files from a Canon 5Dmk3) I get beachballs all the time. This is on a fresh install of High Sierra 10.13.3 and on the Macbook Pro 15" 2017 with Touchbar. Library and footage is on the NAS and SMB signing is disabled. When mounting to the same exact location via a normal 1Gbps Ethernet cable (and smb 3) everything runs smooth as butter! I've created a short video of the problem. Perhaps with your knowledge you can help in figuring out a solution. I've tried almost every trick in the book. Is this the case for you as well via Thunderbolt 3/10GbE? (In the video when I say SD footage I mean HD )
Here's a video of the problem: youtu.be/v8I9clZOrMM
This has been going on forever in FCP X ever since multicam existed! I sent feed back about this last week, please do so as well, because this 'clicking' makes it to the export!!
Open terminal and type this: sudo tmutil thinLocalSnapshots / 10000000000000 4
If you have time machine, and high sierra, macos keeps local "snapshots" for backup purposes. Supposedly they are automatically freed when space is needed, but I have seen cases where they don't release the hdd space. Let us know if this worked for you!
Do try to calibrate both using matrix based, and table based profiles. One of the two works on all fcp x versions. Also try v2 and v4 ICC variations. I think table based, v2 icc is the most compatible calibration.
Thank you for taking the time to give such an indepth view on the core processes. Since the footage is optimised (transcoded to prores) the clips don't require rendering. Also, when adding the color corrector, the clips while unrendered play back smoothly. That's what was weird for me. I am wondering if switching to an amd card, fcpx will use the VCE logic to accelerate playback?
This is weird. I have added a 960GTX (unflashed) inside a Mac Pro 2008 - everything else works great. I have some 4K material, and I have optimised it. When I throw the material in a project, the playback is really stuttery. Now, I noticed that adding a color corrector effect on the footage, makes it run real smooth!??. It seems that playback on this system is not (really?) accellerated by the graphics card, but when I add an effect on, it might pass it through the Nvidia and play it back properly. Any ideas on how to get this working the right way?
As you guessed, it was quite a complex project. It's great to get feedback from people not involved in the production and your thoughts on how the track was like is a nice idea to have in mind for this year. All five years this event took place we were shooting with a bunch of cameras for a whole week, plus the live 2 hour stream. Usually 600GB-1TB footage to sort through. We had a deadline of usually 8-10 days from start to finish for the whole 20' Highlight clip and usually stringed it towards 10 days for the mix and color. Thank you for all your kind words.
Four years later...
A quick (but not so clean way) that works for me (for shorter edits) is to duplicate the multicam clip in place, detach audio from the duplicate clip, select the proper audio channel and use fades to go from one mic to the other...
[File Attachment: ScreenShot2016-06-12at14.51.40.png]
30p inside 60i timeline will create a "progressive look" effect (a bit stuttery on pans) which news channels don't like very much. Safest way to go is to shoot 60p and edit inside a 60i timeline.
Hi bogiesan! Like everything, there are many ways to go about organising and editing. I opened this thread in order for one to show their way of organising. This can help many new comers or disorganised editors looking for a fresh start to have a proper foundation. Though conventions may have been perfected, they can't be applied to all workflows and in the end one has to find their own way of working. After 10 years of editing, I have a workflow I'm completely happy with, and at the same time I'm always looking to refine and improve it - think lean management on a process. We do like getting into specifics here at FCP.co, so why not show us your way of naming things?
Hi everyone from sunny Athens!
Art of Motion is a parkour event taking place every year in Santorini, Greece. Last year, as part of the team I edited, mixed, graded and mastered this short film for Red Bull in FCP X. And yes it was footage overload!
I'd love to hear your thoughts!
Hi! I'm actually not a fan of large names either, I am a fan though of important information about a file - on the file, and I guess that won me over!!...
Lets pretend there is a Red Bull event called Art of Motion 2016 (there actually is). Now Red Bull requires several different edits for that project, such as an action clip, and event clip, a story clip, etc... The word "project" above is a placeholder for "Art of Motion 2016". So my above example would read "16-0605 RB AoM 2016 Action Clip v 1" or "16-0605 RB AoM 2016 Event Clip v 5".
I love the idea of perhaps adding letters to version numbers: reading something like "16-0605 RB AoM 2016 Action Clip v 1b" perhaps using the number as a client delivery version and letters as internal revisions.
I do run into a few "issues" I haven't worked out yet such as footage shot in the same card but spanning several days. Does it still all go into one folder (with the wrong shoot date) or split up into days?
With your way, don't you have identical file names with different projects?
Hi fcp.co from sunny Athens!
Wanting to continuously improve my workflow, I like to pick other people's brains from time to time to check if someone is doing a certain process a smarter way, i.e. less steps, better naming, etc... Thinking this is a good place to share each other's ways of doing things, what's your way in terms of folder structure and file naming practices? We may also talk about what's good and what's not so good.
My way of doing things are using a refined post haste folder structure. It starts with reverse date naming, then a one word client short name then the name of the project: so for a red bull project created today it would be "16-0606 RB Project".
All files (especially exports) keep the whole name above and add onto it a clip name (such as "action clip") and a version number. So an export for the client would become "16-0606 RB Project Action Clip v 1.mov"
So the reverse date (16-0606) works like a serial number (usually it's the creation date, or the first day of an event) and stays the same throughout the project.
The media folder and files follow the same concept with a few twists. The reverse date is the shot date. This keeps card folders organised by day. Add onto it a camera short name (for color correction info) and a card number. So for footage shot yesterday on an A7SII for the project above the folder (and file names) would look like this:
Folder: "16-0605 RB Project Action Clip A7SII C01"
File: "16-0605 RB Project Action Clip A7SII C01 0001.mp4"
That's it for now! We could go into the whole workflow, but let's start at the root shall we?
Always with best regard to this community!
Compared to 10.2 the color correction performance is actually half even on a maxed out new Mac Pro 2013. Color picking or drawing masks is painfully slow! Not to mention on my laptop and my old MacBook Pro. Try stacking three or four color corrections after the primary grade and color pick. The reason I'm complaining is because in 10.1.4 this was a very fluid process . Is it just me?
Thing is that Sony cameras being the example, they do record a lot of this metadata. Using exiftool one can see it. Apple, Adobe and Avid software doesn't. Lack of a standard wrapper format could be an issue here, but there seems to be no tool to even work around this in a semi-automatic manner. Yes the metadata age is in its infancy, but the technology is all here and every camera manufacturer is embedding it in their own way. This is not an issue with photography these days fortunately and that is a great example of good netadata usage.
I believe metadata exists to avoid problems, improve our decisions in post and probably a million other reasons I haven't thought of. Why is it near impossible to retain video metadata all the way through to editing?
Pretend we are in a shoot. One of the cameras is the Sony A7S, another the FS700. Importing through FCP X 10.2 most metadata is thrown away. Metadata such as s-log tags, focal length, etc... In fact, no software other than Sony Catalyst Browse will see this metadata and even this software won't export this data in any way attached to the video clips or otherwise.
So many workflow enhancements can come from having metadata. From custom clip renaming to faster and more accurate color grading. And this is just having the same brand/format cameras in a shoot. Imagine bringing different brands into the mix - which is the case most if all of the time.
Raise your hand if you really need a simple yet powerful metadata management app that you don't need a 500€ subscription to it like SilverStack - which doesn't really work with many formats in any case!
What are your thoughts on the subject?