[Joe Marler]

I have the Pegasus R4 Thunderbolt 1 and am familiar with their certification. Mine also has the Toshiba drives. I don’t know the best choice on replacement drives.

Seagate gets a lot of bad talk due to one particular meta study of one model of 3TB drives: https://www.backblaze.com/blog/hard-drive-reliability-stats-for-q2-2015/

This one situation may be blown out of proportion. I’m not sure reliability comprehensively tracks the *brand*. I have about 150 terabytes of hard drives (various brands) used on video projects, and have had occasional failures from all brands. Even solid state SSD drives can have significant failure rates, although they are usually more reliable than spinning hard drives: https://media.bestofmicro.com/4/T/302141/original/ssdfailurerates_1024.png

In a prior IT job around 1990, we had 100% failures from a large batch of Maxtor hard drives. It was so bad we had to fly out to CA and talk to the manufacturer. In those days one of Maxtor’s top competitors was Priam, and we had excellent reliability from those. Today Priam is out of business and since 2006 Maxtor is owned by Seagate. Not much logic there 🙂

This has little direct relevance to any current models, but it shows how certain batches or model numbers can be either good or bad, and it is almost impossible to know ahead of time.

This large meta study by Tom’s Hardware found that so-called “enterprise” drives did not necessarily have better reliability than consumer-grade drives at the same capacity level: https://www.tomshardware.com/reviews/ssd-reliability-failure-rate,2923-2.html

I’ve had good results with the Promise Pegasus but I face the same situation as you. In the future I may use the OWC Thunderbay 4. If a hardware RAID chassis (like the R4) fails, you can only put those drives in another same-type box. By contrast if a sofware RAID like OWC fails you can put the drives in any other similar chassis. https://eshop.macsales.com/shop/Thunderbolt/External-Drive/OWC/ThunderBay-4

My main advice is plan ahead and allow time for “burn in” testing using utilities like ScannerZ: https://scsc-online.com/Scannerz.html or DiskTestR: https://diglloydtools.com/disktester.html

BTW here are some test results I did on the RAID 5 sync rate of the R4. In general I’d suggest not using a 128KB stripe size.

https://joema.smugmug.com/Computers/Pegasus-R4-RAID5-Sync-Rate/n-p2Nn4p/i-pdz6p2j/XL

[Joe Marler]

Yes FCP X and Premiere both are heavily multithreaded and can use multiple cores effectively. You can see this by monitoring performance with Perfmon (Windows) or Activity Monitor or iStat Menus (Mac) when doing various editing operations. Often you’ll see all available cores at high utilization. Encode/decode/export of long-GOP codecs like H264 is mostly CPU-bound so in general the more cores the better.

Minimum Apple RAM then using OWC RAM is a common procedure. It does not invalidate your warranty.

Yes the M395X is the best GPU in the new iMac. Whether on FCP X or Premiere it seems you can never have too much GPU, so getting the top available choice for your budget is usually a good idea. This is especially so on iMacs which cannot upgrade the GPU.

Fusion Drive is not bad but for video editing you usually want your media on external storage. If media is external it raises the question of why use FD since you don’t need the space. I have both FD and SSD iMacs and with external Thunderbolt media I really can’t see much perf difference attributable to the boot drive. I don’t usually see much difference in boot time of various apps. However when moving lots of data you can encounter situations where FD “falls off the cache” and drops to nearly HDD-level performance. SSD has more consistent performance, so I’d generally recommend a 512GB SSD and external fast storage.

I don’t think you’ll see much perf. difference between Thunderbolt 1 vs 2 for your situation. The RocketStor dock looks interesting but I haven’t used it.

[Joe Marler]

[Shawn Miller] ”
“Meanwhile on a Mac I recommend to simply use EditReady, transcode to ProRes 422 LT and you will be able to playback full quality full res 4K on a laptop. There’s no quality loss from XAVC-S.””

Yes I tested that also. If I transcode the H264 4k material to ProRes, then even multicam is quite fast on Premiere CC — on either Windows or an iMac. I just wanted to make sure there was no other solution, hence the discussion.

[Joe Marler]

[Shawn Miller] “Your PC has half the RAM and a slower GPU…”

Based on Windows Performance Monitor, RAM is not an issue. This is indicated by the global perf. counters “available bytes”, paged and non-paged pool, plus the Premiere Pro per-process counters “private bytes” and “working set bytes”. So whether it has 16GB or 64GB would not make any difference in this specific case.

The GTX-660 GPU while a few years old is a big PCIE-bus card that pulls about 240 watts — more than my entire iMac. The GTX-780m in the iMac is essentially a laptop GPU. It is the “m” version. In AnandTech’s test, they found “the GeForce GTX 780M…ends up delivering performance that falls roughly half way between the GTX 660 and GTX 760”. So the PC GPU is faster overall than the 2013 iMac GPU: https://www.anandtech.com/show/7287/analyzing-the-price-of-mobility-desktops-vs-laptops/3

As a sanity check I installed and ran Premiere CC 2015 on my 2015 iMac which has a 4Ghz Skylake CPU, M395X GPU and 32GB RAM. It didn’t make much difference. It was a little faster but the overall situation was roughly the same.

[Joe Marler]

[Aindreas Gallagher] “PPRO CC can cut 4K AVCHD coming off a GH4 like, you know, butter – on a three year old laptop. it did for me…PPRO CC works as well or better, with more formats, than FCPX, on older hardware. I think most people on here who use both would agree on that.”

That is not my experience, but maybe it depends on your definition of “old”. I just re-tested some 4k material from a Sony A7RII, Panasonic AG-DVX200, GH4 and DJI Phantom Vision 3 Pro on both Premiere CC 2015 and FCP X. The Windows machine has a 4Ghz i7, 16GB RAM, 1TB SSD, GTX-660 and 7200 rpm media drive. The iMac is a 2013 model with 3.5Ghz i7, 32GB RAM, GTX-780m and single 3TB Fusion Drive with programs and media. I’d say the hardware is very roughly equal — the iMac CPU is slower but a couple of generations newer. The Windows machine GTX-660 may be faster than the 780m in the iMac; at least it did better on LuxMark tests.

I tested single and multicam projects on both machines. In general Premiere was quite sluggish at JKL timeline editing on the native 4k files, even on a single stream. It seemed a little more responsive on the GH4 material than the A7RII and DVX200 stuff, so if you’ve only tested one 4k camera, maybe that’s it. On 4k multicam, the timeline and viewer update response slowed for both Premiere and FCP X. However FCP X was still far more responsive. E.g, when fast forwarding in the timeline, Premiere updated the program and multicam monitors about once every 4-5 seconds. By contrast FCP X updated them 5-10 times per second — that’s with Premiere on 1/4 res and the FCP X viewer on “better performance”.

IOW FCP X was faster at JKL editing 4k multicam at full resolution than Premiere was at 1/4 res. Typically on FCP X you’d just use proxy for 4k multicam since it’s so easy and built in. Using proxy mode, multicam 4k was fluid and very fast; I’m tempted to say “like mercury” 🙂

With Premiere you can render the timeline but in my tests it took longer to do that than FCP X took to generate proxy files. Then if you apply an effect to the Premiere timeline it’s no longer fully rendered and it slows down. Of course you can transcode externally before importing to Premiere, but…

It says on the Premiere overview video it “allows editors to work with 4k and beyond, without time-consuming transcoding”, and “never needing to render until your work is complete”: https://helpx.adobe.com/premiere-pro/how-to/what-is-premiere-pro-cc.html?set=premiere-pro–get-started–overview

Having edited for years on CS5 and CS6 before switching to FCP X, I agree the “no transcode” thing worked pretty well on SD and HD but it’s not working that well on H264 4k. Maybe there is something wrong with my Windows machine. Maybe I need a much higher end machine, although if you can fly through 4k AVCHD on an old laptop “like butter”, maybe not. This guy had a super machine and he was forced to externally transcode for A7RII 4k material: https://www.eoshd.com/comments/topic/18574-a-story-about-4k-xavc-s-premiere-and-transcoding/

If Premiere can edit native 4k H264 multicam with “butter smoothness”, what hardware, configuration or procedure is required for this?

[Joe Marler]

Below are some GPU benchmarks I ran in cooperation with another 2015 iMac 27 owner having a M395. Note some of his numbers are way off — I don’t really trust those, so I would discard those extreme variations. He ran out of time to pursue further testing.

In general the M395X is maybe 10-15% faster than the M395. You can pretty much never have too much GPU performance so IMO it’s worthwhile.

https://joema.smugmug.com/Computers/IMac27-M395X-vs-M395-Tests/n-VS3FhM/i-cSSd4cn/X2

[Joe Marler]

[Craig Alan] “…can FCP X even use the original card files as “original media” . I had always thought these needed to be decoded”

I believe yes except for the special cases of AVCHD and XAVC. In those cases the “leave files in place” import option is not allowed — even if the folder structure is first copied to a hard drive. If the video files themselves are copied outside the structure (generally not a good idea) then it can import those file types with “leave files in place”.

[Joe Marler]

[inanc tekguc] ” these archive files of FCPx can only be read by FCPx.
I just switched from Premiere to FCPx and I can imagine switching again in the future”

The FCP X archive files can be read by any computer. It is just a package wrapper. On a Mac you can click on the archive file and select “Show Package Contents” and it reveals the original camera card structure. If you copy the file package to a Windows PC it just appears as the original camera structure.

However my documentary group does not use this method for several reasons:

(1) In the field we back up, duplicate and verify each day’s work from each camera. This consists of copying the entire folder structure then verifying the source/destination with Beyond Compare: https://www.scootersoftware.com/ We could theoretically also do this with the current FCP X archive feature, but this workflow was in place long before FCP X had the current archive functionality, plus there are other factors (below).

(2) Our shooters sometimes fail to reformat their cards and creating a camera archive would create duplicate clips and waste space. That said there’s a good argument for just copying everything and de-duplicating it later.

(3) Some of our cameras shoot combined stills and video, and the FCP “create camera archive” function puts the stills and video into the same file package, but we want the stills handled in a separate workflow. We could dig into the archive and copy them out but that violates the idea of not messing around inside the archive.

(4) In the often rugged field environment we may not have a Mac laptop with FCP X installed, or it might be unavailable. We might have to do the daily backup copy using a Windows laptop.

That said there are significant advantages to using camera archives, as discussed by MacBreak Studio:
https://www.youtube.com/watch?v=rv1gRTUXcDw

Re Carbon Copy Cloner, I think the checksum mode is normally not used. You have to invoke that.

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]

If I understand that workflow correctly, it would not be practical for us. NEAT video NR is incredibly slow and compute-intensive, esp. on 4k files. Since our typical documentary shooting ratio is 30 or 50:1, it would lengthy pre-processing, when ultimately only a tiny portion of that material is used.

That is why we usually defer things like video NR, stabilization, flicker reduction, etc, until late in the edit. However the idea of “baking those in” does solve the problem of FCP X re-rendering the slow effect when other edits are made.

[Joe Marler]

[John Rofrano] “I like to correct for things like noise and shake first and then render out a pristine file and use that in my project.”
Do you mean a rendered file for each clip or one consisting of multiple edited scenes that you then process with Resolve, or re-import that into FCP X? Do you use automatic scene detection, and if so what about transitions?