Lack of FCP/X share AppleTV performance, or just lazyness

I just upgraded from MacBook Pro 13" (2012) with i5, 16 GB and SSD to a MacBook Pro 15" (2016) with 2.9 GHz i7, 16 GB, Radeon 460 and SSD.

Bruce/X bencmark went from 240s to 42s with the new computer, but when sharing a 1:45:00 long movie the new computer wasn't faster. Roughly 2:20:20 in both cases. The 2.5 TB library was on a G-RAID Thunderbolt disk with 250 MB/s thoughput and file written to SSD (2016) or 5400 rpm harddrive (2012). El capitain for old computer and Sierra 10.12.2 for the new computer.

I then tried 5K RED footage on the new machine with both FCP/X and Premiere Pro CC 2017. Rather surprised when PP render in 8:20 and FCP/X in 21 minutes, with a file twice as big. CPU utilization in PP ca 800%, and FCP/X ca 300%. Rather snappy with both applications

No proxies, no opimized media in the last test. Both proxies and optimized media in the first one.

Clean install of OS and applications, migration of user data..

Your experiences or clues how to improve rendering speed appriciated?

casimir_artmann wrote: ...I then tried 5K RED footage on the new machine with both FCP/X and Premiere Pro CC 2017. Rather surprised when PP render in 8:20 and FCP/X in 21 minutes, with a file twice as big. CPU utilization in PP ca 800%, and FCP/X ca 300%. Rather snappy with both applications ..

I just did a brief test of a 7 second 5k R3D clip, exported to 4K H264 single-pass from FCPX 10.3.1 and Premiere CC 2017, both at 30 mbps.

FCPX took 21 sec, and exporting to 4K ProRes 422 from was about the same time -- roughly 3x real time.

Putting the R3D material on a 1080p timeline did not make much difference -- about 19 sec.

Premiere CC 2017 took 1 min 1 sec using the same export parameters -- 4K H264 single pass with 30 mbps bit rate. Both ran on the same machine, a top-spec 2015 iMac 27 with media on a Thunderbolt 2 SSD RAID.

Normally exporting from H264 UHD 4k to H264 1080p from FCPX is about 1/3 real time (IOW 20 sec to export 60 sec material) so for whatever reason it's much slower on 5k R3D material. However (at least on my machine) Premiere CC is 3x slower than FCPX for the same case.

How long was your RED footage? The ratio of program length to export time might be interesting.

The RED clip is 2 minutes and I did a 90s project / sequence as a test.

Would guess something with combination of CPU & GPU with FCP/X as PP was much faster for me.

I will try running compressor during night and tomorrow, and see if any differencies

If the 5k RED clip is 2 min and it took FCPX 21 min to export (10x real time) that definitely doesn't sound right. That is 3x slower than my 2015 iMac and a 2016 MBP shouldn't be that much slower. If Premiere CC took 8:20 (a little over 4x real time) then it's 2x faster than FCPX which I've never seen on anything.

If you could tell me the exact export parameters you're using I'll be happy to do more tests on both my iMac and my 2015 MBP. If you want to pursue this we should probably use the same source file. You could download one from here and I could use the same one: www.red.com/sample-r3d-files

I re-run Bruce/X test this morning and it took 42s as last time.

Then I shared the same 5K Red clip as before, 1080/25P and encoded H264 for performance. The fans kicked in and the CPU was running full speed and finished in 6 minutes for a 90s clip.

One difference was that I had the setting for H264 highest quality on the export yesterday so I re-run with highest quality and the export took 11 minutes. Near full CPU utilization this time compared to yesterdays 50% . The first 20% of rendering was faster than the last part. Probably thermal protection that kicked-in.

Another difference was that yesterday, I had external Thunderbolt drive, an Apple Thunderbolt display, Wacom tablet and a Focusrite Scarlett attached to the laptop, as this is my normal editing setup. But it's hard to see why CPU utilization yesterday was only 50% compared to 100% today, except for sheer lazyness

An additional test with optimised and proxy media and background rendering made the export time shrink to 3 minutes, i.e. twice the time for the 5K clip.

Lessons learn: There are a lot of parameters that influences performance on a new MacBook Pro running FCP/X.

An additional test with optimised and proxy media and background rendering made the export time shrink to 3 minutes, i.e. twice the time for the 5K clip.

Lessons learned: There are a lot of parameters, not only hardware, that influences performance on a new MacBook Pro running FCP/X.