I'm a video professional, been using FCPX for years. Just upgrade from a 2017 15" (2.9Ghz Quad i7, 16GB RAM, 4GB Radeon 560) to a 2019 16" (2.3Ghz Eight core i9, 64GB RAM, 8GB Radeon 5500).
I hate bringing silly question to these forms, but I'm stuck. Maybe a point into the right direction for research at least is what I'm looking for.
For some reason playback on unrendered video is nearly impossibly slow. Choppy at best, often lagging/freezing for ~3 seconds when moving from one clip to another during playback. It's like the effect of editing off the worlds slowest hard drive (but I'm editing off the internal SSD). Whether there are adjustments/efffects or no adjustments. 4K or 1080. (all 8bit). Same exact performance (lack of). Better quality or better performance previewing. Proxy hardly even helps.
I literally have had to render out my timeline every few minutes to check my timing for cuts, its impossible to check unrendered, too choppy.
Are there any other things I can try? I have graphics switching off, so it should default to the Radeon 5500. I did notice in monitoring that often enough the intel chip triggers - should it be using both or no?
Feels like there's a hardware/software malfunction. Editing was faster on my last computer.
noahrtaher wrote: ... I have graphics switching off, so it should default to the Radeon 5500. I did notice in monitoring that often enough the intel chip triggers - should it be using both or no?...
It it equally slow with automatic graphics switching enabled? Turning that off might inhibit use of the integrated GPU which is where Quick Sync decode/encode acceleration is located. However I thought that T2-based machines use the T2 for that, not Quick Sync.
It should be much faster. Make sure you are on 10.4.7 or later; there were significant performance upgrades in 10.4.7.
Also be absolutely certain you are not using the Chrome browser. It behaves improperly and saturates the VideoToolBox framework which locks out FCPX and makes it slow and erratic.
If it is none of those things, state what camera this came from and the codec specifics.
Hey! Thanks for the info about QuickSync. Looking into that.
I have tried with both graphics switching on, as well as off. Same result. New development though...
I believe it has something to do with HLG footage from my Sony a7iii. I shoot most of my weddings with the picture profile, HLG3 + rec2020 HDR. Final Cut Pro recently received some updates involving HDR support, as well as updates to its handling of Sony XAVC codec (listed in 10.4.7 and 10.4.8 release notes). I wonder if there are errors in the new updates that are causing performance issues.
I mention this because higher bitrate 4K footage from my GH5 are absolute butter, edit incredibly fast with realtime unrendered playback even with effects, LUT, film grain. Funny enough, so does footage from my Sony when using Cine4 profiles. Heaven. Definitely what my $3500 16" was made to do. This was not always the case that HLG was slow! I bet the issue has something to do with the new updates. Maybe a bug in performance.
Ronny is correct, it should work OK. Before we switched to all-ProRes acquisition, my doc team shot (and I edited) literally hundreds of terabytes of 4k Sony XAVC-S. We usually used SLog2, not HLG, but we did HLG test material and I don't remember any problem.
10.4.7 is definitely much faster than 10.4.6 on most H.264 codecs including Sony XAVC-S.
However it is conceivable there could be a narrow problem isolated to Sony XAVC-S using HLG3+ Rec2020 HDR. I have not tested that.
The fact it also behaves that way in proxy mode is very suspicious. Proxy is only 1080p ProRes "Proxy", which is a very lightweight fast codec. Even a 2013 2-core Mac Book Air should be fast on that. This could imply a deeper problem.
Be sure all material is on local drives, and these are HFS+ not ExFAT.. If you are using a USB-C SSD drive, be certain you are using the original cable. Those new drives are highly sensitive to cable type and will often silently step down to USB 2.0 speed. Run Blackmagic Disk Test on all of your drives as a sanity check.
Be *certain* you are not running the Chrome browser. Also - make a duplicate snapshot of your timeline, open that, then select all clips and strip all effects with Edit>Remove Effects. That will help determine if an effect is involved.
I'd test this myself but I'm leaving on assignment this morning.
Thank you for the offer to check out some of my footage Ronny, just sent it over retransfer. All 4 are Sony a7iii HLG3 rec2020. Clips C0005 & C0077 are 4K 24 frames, Clips C0135 & C0213 are 1080p 120 frame. I edit on 1080p 23.976fps timelines.
At this point I really think it could be a narrow issue - but if you're able to edit the footage well... it would be nice to figure out how I can get there too.
Okay, I just made a screen recording showing two different 1080p 24fps wedding edit timelines. One wedding was shot entirely in CINE4 rec709 (first one shown), the other was shot entirely in HLG rec2020 (second one shown). I tried to make it obvious in the screen recording how capable the rec709 footage is to edit. I even take off all LUTs and effects on the HDR footage, with no improvement. And showing that I can keep all adjustments on the rec709 footage just fine. To be clear this performance difference was never the case.
Something with HDR seems to be the only thing I can narrow it down to. I also made an Etre Check Report for you guys to look into but I've done it before as well, and it came back positive.
Thank you for sending the footage. I have done some testing on different machines, and I am unable to reproduce your issue.
In the video I am attaching, you will see that I can play both the 1080 HLG and the 4K HLG footage, with grading and effects applied, on an unrendered timeline without any issues. The video was recorded on a 2014 MBP running Catalina and FCP X 10.4.8 on an HDR timeline with playback setting to Better Quality, you should even get much better results on your MBP.
I'm sure that there's nothing wrong with your footage nor with your combination of OS and FCP X. The only thing I can think of that could cause these issues for you is the use of outdated plugins. I saw that you have Neat Video, that's a performance hog as it has not been optimised for modern GPU-based workflows.
If other people want to chime in, any ideas are welcome.
The video example you uploaded is helpful, although a bit different in use case. You were editing in an HDR 4K project, and I normally edit in a 1080p rec709 project. Either way I think you're right - you aren't having the issue. I am beginning to think the issue I'm having could possibly be related to the way that the software FCPX is responding to the new 16" MacBook hardware. Not sure. I was working around a similar thread in apple support communities, and had a moderator there identify it as a bug.
Anyways, I definitively discovered recently as shown in my "Screen Recording 3" video, that the main underlying issue is that playback - even in the media browser - are unable to playback in realtime. Frames are dropped even just previewing footage from the media browser before it gets to the timeline for editing. For the rec2020 HLG Sony a7iii footage. It also spikes GPU usage to 160%. Is it possible that there's a corrupt HLG playback cache file or something somewhere?
For now, the workaround is to use proxies - which I discovered DID actually provide a way for me to edit and playback the footage. It's not a fix, but a workaround for now. Hopefully Apple will patch this soon.