[Erik Lindahl]

It depends on what you do. A year back I tested doing transmission masters in FCP7, FCPX and PP (I presume CS6). FCPX fastest at rendering one sequence out hands down, FCP7 beat PPRO. HOWEVER, FCPX’s rendering to SD-resolution was worst of the pack. Further more, FCP7 has batch-exporting. FCPX has a background encoding list but you have to export times per sequence (sequence handling has been boosted in 10.1 but still no batch export). This oddly makes the old-timer with out GPU acceleration and 64-bit the best tool for the job.

In other situations FCPX has blown my mind with its realtime performance. Still haven’t felt a super-urge or need to swap from a tool that works just yet. By the time I so, Resolve 11 might be a better solution or my self as we primarily do online / finishing work.

[Erik Lindahl]

Well 400 MB/s isn’t that poor for 1 SSD over USB-3. The test has a RAID-0 setup hence using multiple ports. It’s impossible to get 800 MB/s over 1 USB-3 port.

On the MacPro, to achieve something similar the only solution would be to add a USB-3 interface to one of the thunderbolt ports.

[Erik Lindahl]

Love when people blurb things like X cores or X Gb’s out with out backing it up. I just did a UHD-project that rendered FASTER in FCP7 compared to PPCC on a new MacPro. 32-bit vs 64-bit, CPU vs GPU etc. The later isn’t always better. There are reasons to move on for sure but newer isn’t always better. And FCP7 can use way more than 1 core, but probably not 1 CPU of today’s 12-core beasts.

[Erik Lindahl]

What are you transcoding them to?

[Erik Lindahl]

To answer your question we’re just about to install a new 8-core system. A few things to think about:

– The 12-core system can be a waste in a lot of cases. The 6-core and 8-core machines for a lot of cases order better bang for the buck or just simply more bang.

– 512 GB or 1 TB internal SSD will offer better I/O than the 256 GB.

– I’d defiantly go for the D700’s for future proofing.

In terms of our system the only issue we had was that the system froze when it went to sleep at the same time as the screen was sleeping. Seems to be a know issue. Aside from that we’re just planning the I/O properly to max out the thunderbolt ports

BUS 0 = GUI SCREENS
BUS 1 = PROJECT DRIVE SSD (1300 MB/s)
BUS 2 = MEDIA RAID (800 MB/s) + VIDEO I/O

I’d imagine we primarily get external media from the net or USB-3 drives. The above should work smoothly. The machine we got (8-core, D700, 64GB, 1TB) is impressively silent as well. Wouldn’t have an issue having a system like that in the studio.

[Erik Lindahl]

Is this really true? Now I’ve only compared a 2008 8-core system with a 2013 8-core system and not PremierPro specifically.

From my tests AE takes roughly as little as 30% of the time – ie 60 min render is down to 18 minutes.

From my tests in Episode Pro transcode take around 40% of the time – ie 60 min render is down to 24 minutes.

Both of the above are primarily CPU, I/O, RAM etc stress, not so much the GPU. If you’re comparing something currently accelerated by CUDA (3D RayTracing in AE for example), yes, an older system will be far faster. But even apps like Resolve require a very much upgraded old system to be compare:able to a 2013 system and you more or less are required “unsupported” solutions. But yes, a maxed out 2012 12-core MacPro CAN be quite upgraded.

[Erik Lindahl]

[Andrew Kimery] “A new Thunderbolt, a new connector yet still no optical. If this rumor holds true it will mean three iterations of ThB in 6 years. Kinda makin’ me feel like you rushed this to market Apple…”
You mean Intel? I don’t see how having 10 Gbit or 20 Gbit connectors on your iMac, MacMini and MacBook Pro can be a bad things during the revisions of Thunderbolt.

– TB1 and TB2 har 100% identical aside from the fact TB2 will give you in the realm of 70% better throughput and 4K support.
– TB3 is 100% compatible with TB1 and TB2 and if the above goes again, we’ll see storage reaching 2250 – 2500 MB/s, dual 4K screen support (or 1 4K screen and 1 storage device at the same speed as TB2 from one port) as well as 100W power that can probably drive a lot of external devices like video interfaces and simpler storage solutions.

I don’t see a huge problem. The only thing that will be a head-ache are the two connectors. And I think TB3 will at the earliest come out first quarter 2015, most likely is somewhere around third quarter 2015. If we’re “unlucky” we’re talking spring 2016 if TB3 is tied to the release of “Skylake” seeing “Broadwell” has been delayed.

[Erik Lindahl]

I fully agree. Frequent feature updates and bug fixes are very welcome. I just hope the features they push actually work in the real world. Masks have been some what iffy in AE previously in conjunction with filters, PrPro has had a horrible mask / matte engine. The whole mask-tracking feature sounded AMAZING but in reality it’s so-so useful in something like AE. For an editor like PrPro I can see some function for sure but still quite an odd solution in this day and age. Text templates sound amazing for some workflows. Just hope it works in a sane way and performance isn’t horrible.

[Erik Lindahl]

…and to be honest, this release doesn’t that that massive.

[Erik Lindahl]

Does this mean Premier Pro / AE / AME will have the ability to import and playback DCP’s also?

Have you updated masks in any way with this release? Previously using a mask on a layer as it’s alfa / matte could cause a lot of unwanted artifacts with certain filters – even built in ones. Also, the previous “tracking” feature of masks was quite terribad seeing you went back to the 90’s and merged the tacking and maskin data into each other. Why not use something like Mocha as your tacking back-end? Less keyframes, the ability to re-use tacking data across multiple sources (not just masks) – the possibilities and bonuses are limitless.

Can we (finally) select any layer or mask as our matte for a given layer? Or is this asking too much?

Will this release improve the situation of thing between AE and PrPro? I’ve been quite surprised how poorly things translate between the apps when this is one of your major selling points. You still in the latest version call different filter differently in the two apps an as stated, the workflow between the apps could be tons better.

Will we ever see anything new in the workflow-area? I.e. having one of your apps splitting media into a set directory structure in a given manner for mor efficient post workflows and versioning’s? Will we see some kind of scene and edit version support in Premier? What about useable image sequence support?

Sorry for the bombardment but I always hope a lot of these things will be fixed at some point.