Thoughts on Mac post future? (and why I think the next Mac Pros will rock for RED)

[Bruce Allen]

Hey folks

With all of the FCP X stuff going on, people seem to have lost faith in Apple's ability or desire to make the next Mac Pros good and "pro".

Umm... here's why I think they won't go low-end. They'll be damn good. And the 2012 Ivy Bridge ones will be good too.

2K IPADS: If the next iPad really is 2048x1536 resolution as rumored... Apple's going to want to make a machine that prosumer-level folks can use to make 2K resolution movies for it on FCP X without breaking a sweat. Sure, us pros can make 2K stuff now with low-end hardware and software - but that's because we do clever manual steps, caching, have fancy RAID cards, etc. But the Mac Pro is gonna have to be fast enough for the average fruit moron to get fluid 2K output EASILY- which means we can likely push it to 4K if we're clever :)

TALKING OF 4K: All of those high-res icons in Lion, plus the fact that FCP X is targeted towards output straight from the gfx card... don't you think that fits in well with a very high-res display for a Mac Pro?

THUNDERBOLT: Up till now, a huge weakness of Mac Pros has been lack of expansion. Thunderbolt really helps fix this. The fact that high-end iMacs had 2 Thunderbolt ports is extremely encouraging.

GREATER GPU EMPHASIS: Do you really think Apple has spent all of this time on OpenCL, FCP X, etc... and isn't planning to allow some half-decent gfx cards on it? Sure, previous Mac Pro graphics cards have been mostly pathetic - but these are good signs.

TRADITION OF VALUE (RELATIVE TO OTHER WORKSTATIONS): Historically, when they're released, they've been good value vs the PC workstations. Sure PC desktops beat them for price/performance... but the expensive Dell and HP workstations get beaten too!

THE MAIN CPUS ARE STILL GOING TO BE FAST: Sandy Bridge is damn fast and Ivy Bridge will be even better. No way Apple's not going to put the best Intel has in there.

Back in 2007 folks criticized me because I said that the folks screaming for 4K onlines on a Mac were being unrealistic. But now it's 2011. Personally I might skip this generation - I'm in the "writing / storyboarding / working my ass off to raise funds" stage of work right now. But at the very least, come 2012 I am getting an Ivy Bridge Mac Pro. And I damn well will make something good in 4K in 3D, 48fps. On my desktop :)

I'm not sure what I'll edit it in (likely Avid but honestly - whichever editing package copies the best features from all of the other ones wins :). Not sure what I'll be doing the effects on (Nuke - or my old favorite After Effects if they suddenly fix their problems). I'm pretty sure I'll be grading it in Resolve though and doing sound in Pro Tools. And pretty sure I'll be using a Mac Pro.

What do you guys think?

Of course, might still use some cheap PCs with gaming GPUs for 3D rendering. Plus I'm watching to see if nVidia can really build that "ARM plus CUDA supercomputer-on-a-chip" thing they're working on. But I don't see anything beating a Mac Pro as a general purpose post workhorse for a while...

Of course quite a lot of this is thanks to Resolve. If they make a PC version of Resolve, then it'd be quite possible to move.

Bruce Allen
www.boacinema.com

[Alexander Ibrahim]

I agree with you Bruce.

I think that what FCP X looks like in 6 months is much more important than what it looks like today.

I think we can expect at least two more generations of very serious workstations from Apple in the Mac Pro tradition.

I think 3 generations out though it will start to matter less and less. I mean today right this red hot second the fastest Mac you can buy is the high end iMac. On most tests its faster than a 12 core Mac Pro ... that's how much impact Sandy Bridge has. (I think that Storm, Smoke, Resolve put the lie to this ... but how many Mac Pro users use those applications!) The current MacBook Pro 17" is faster than any Mac before it except the 2010 Mac Pro's with 6 or more cores! That tightens the window on how quickly we have to amortize the full value of a Mac Pro.

My current 8 core 2010 Mac Pro was paid off the day I bought it ... if it wasn't I couldn't have bought it. Seriously. I predict that that will be the normal financial case going forward after roughly Q2 2013.

Of course, maybe Dragon and the RED F35 and RED 645 cameras will push that date further out. Maybe regular finishing in 4K will push it out a bit for us ... but that's the trend.

As an extreme example of that trend ... how many of you were in digital post in the 1990's? The iPad 2 would have been the top supercomputer in the world in 1993-1994, and its more powerful than desktop computers well into the 2000's. It's very likely that the iPad 3 will perform 1080p editing in iMovie or "FCP X for iPad" better than any G5 PowerMac with FCP 6, and that the iPad 4 will match 2008/2009 Mac Pro's with FCP X.

As always, our tools will melt away in importance ... but increasingly what we, the artists, can do will become more important.

[Gene Crucean]

I fully agree that the next Mac Pro's are going to be worth the wait. Think JJ Abrams (if anyone gets that reference/connection I'll be pretty stoked as only a few of you probably are in a position to know).

But I strongly disagree that typical desktop computers (mouse and keyboard) are going away any time soon... if that was implied. I might be mis-interpreting a comment or two. Having said that, I do think content creation will get stronger and stronger in the tablet sector. This stuff will mostly appeal to a slightly different audience than us however.

As far as FCPx is concerned, I love it. I hate that they chose to release it with some of these missing features that we all "need". But I get what Apple is trying to do, and it's ballsy. It's extremely ballsy. I also don't really understand why any of those who call themselves a professional, would get soo bent out of shape over a 1.0 release not being perfectly ready to roll. Even if it was perfectly ready to roll, you "should" be doing a lot of testing before rolling it out into any kind of pipeline/studio. There are no excuses for not doing that. Regarding how Apple's future looks in the post arena... http://blog.nicedissolve.com/2011/06...-fcpx-workflow . That gives a tiny insight into it. The future is bright, trust me.

Which makes me feel like there are a loooot of FCP users at home that call themselves professionals, and set their expectations so high they had nowhere to go but down (I just did that with Super8 heh). People that don't really have any other workflow issues or concerns and are just plain ready to move forward to a completely new version. I think they are the ones that are taking this version so personally.

[Bruce Allen]

Thank you Alexander! I pretty much agree with your points 100%.

As far as FCPx is concerned, I love it. I hate that they chose to release it with some of these missing features that we all "need". But I get what Apple is trying to do, and it's ballsy. It's extremely ballsy. I also don't really understand why any of those who call themselves a professional, would get soo bent out of shape over a 1.0 release not being perfectly ready to roll.

Well, if they want it to be treated like a 1.0 release, they could call it something different. Their choice. The thing says 10.0 when you load it up, for heavens' sake!

They wanted to cash in on the Final Cut Pro name but they didn't want to bother to make it load old Final Cut Pro projects. Please name another professional app that kept the same name but can't load old project files. Heck, imagine if OS X couldn't read files from OS 9!

In the SuperMeet they said "This is a project in FCP 7... and this is the same project in FCP X - look, so beautiful!"
Deliberately misleading, I say.

Even if it was perfectly ready to roll, you "should" be doing a lot of testing before rolling it out into any kind of pipeline/studio. There are no excuses for not doing that.

Well, as a pro, I was kinda hoping that the basic functionality would be in FCP 10.0. You know, so that they could bug fix it for version 10.1. And then I test it and roll it into my workflow. So now I have to wait frikkin' two versions.

Regarding how Apple's future looks in the post arena... http://blog.nicedissolve.com/2011/06...-fcpx-workflow . That gives a tiny insight into it. The future is bright, trust me.

What in my thread made you think I disagreed with this?

Bruce Allen
www.boacinema.com

[Gene Crucean]

Bruce, I'm sorry. I was in agreement with most of your post. Was backing you up more than anything.

But personally I don't care what it's called. It's a "complete re-write". That's heavy stuff no matter how you cut the mustard. Being a programmer also, I know just how much work it takes to get a product like this out the door. Just my 2 cents. That's all.

[Alexander Ibrahim]

Thank you Alexander! I pretty much agree with your points 100%.

Thank you, but I'm just trying to be reasonable about FCP X and the future.

I think too many are evaluating FCP X and the future of computing as a re-do of the past.

Well, if they want it to be treated like a 1.0 release, they could call it something different. Their choice. The thing says 10.0 when you load it up, for heavens' sake!

I sort of agree, but mostly disagree.

A rose by any other name ...

FCP X is a brand new 100% newly written codebase. They have very little carried over from FCP 6/7 or iMovie.

Now, I think this is obvious from looking at FCP X itself. More importantly, Apple has been pretty clear since NAB that FCP X was a complete re-write.

So, no matter what marketing said, or what they called it, we should have known it would lack a lot of important features on its release.

It was pretty obvious to me that the FCP X demo at NAB was pretty close to a rigged demo. Its pretty obvious to me that everything said before FCP X's release was tightly controlled and approved by Apple PR and marketing.

I'll admit, while I expected FCP X to lack important functionality, I never expected it to lack XML support.

I want to be clear ... I'd accept XML support. I don't need the ability to directly import FCP 6/7 projects. Seriously don't need it. I can deal with exporting XML from those apps and importing it in FCP X.

In the SuperMeet they said "This is a project in FCP 7... and this is the same project in FCP X - look, so beautiful!"
Deliberately misleading, I say.

I agree with this.

I want to say however that Apple made that comment based on automated tools they used.

Those tools were not released with or included in the FCP X package.

There is probably a show stopping bug ... something that makes it trash some common case in FCP projects.

Now ... as a pro, I'd be VERY happy to get that tool anyway as it stands today. So long as the problem areas are mapped out.

There is a solid chance that I can reconstruct and lay out an FCP 6 project so that it can be handled by that tool ... worst case scenario by rendering and relinking portions of the project.

Well, as a pro, I was kinda hoping that the basic functionality would be in FCP 10.0. You know, so that they could bug fix it for version 10.1. And then I test it and roll it into my workflow. So now I have to wait frikkin' two versions.

Agreed!

[Bruce Allen]

I want to be clear ... I'd accept XML support. I don't need the ability to directly import FCP 6/7 projects. Seriously don't need it. I can deal with exporting XML from those apps and importing it in FCP X.

Sure! XML would be fine. Or a documented API or a roadmap... Python... just... well, anything!

When FCP first came out, I remember it being very developer-friendly. Hence the 3rd party support. Hope they do the same thing with FCP X pronto.

Bruce, I'm sorry. I was in agreement with most of your post. Was backing you up more than anything.

Sorry! I think it was my fault actually - I guess I'm just so used to being criticized as an Apple hater on all of the other threads that I assumed it as such in this thread too :)

Anyway, what are you guys planning on? FCP X? Augmented with which tools?

I'm fascinated to see what they do with Logic. For audio stuff... the whole "tracks" thing just makes a lot of sense. So I don't think they'll do a magnetic timeline there... but you never know!

Also, RE: the whole video output thing... even though I love Blackmagic for what they did with DaVinci, I a little on Apple's side here. I wish DaVinci just had fullscreen output from the nVidia card like Scratch, etc. Either that or get rid of that slight lag while they copy it from the GPU -> the Blackmagic card. I color corrected a feature with mostly Dreamcolor & Apple monitors connected to DVI. Looked great and it won best international film at the the Santa Barbara Film Festival. So I don't buy the idea that everything has to be HD-SDI.

If FCP X spurs Apple to sort out their drivers (Windows has 10bit Displayport output, stereo 3D output - we need this on Mac!), great.

Bruce Allen
www.boacinema.com

[Geoff McLeod]

I think they should ask Mr Jannard to take over the running of Apple's pro development department. I'm sure he's busy enough, but Apple needs someone like that to drive things (and communicate with their customers). In the meantime, we wait.

[Quentin Brown]

Although this thread got a little derailed into talking about FCPX it does seem to be an indicator (or confirmation of what we suspected for a while from various indicators) that Apple are not so concerned with the needs of the very high end users.

Rumours are that although the next MacPros will have Thunderbolt and be rackmount they may have as little as 2 PCIe slots!
Perhaps Apple are betting the farm on Thunderbolt but it seems a little premature to me and even if they release an affordable rackmount PCIe expansion box that uses Thunderbolt any one card will be limited to the 4xPCIe speeds of any one Thunderbolt channel. Try running Resolve on that with a RedRocket card! Fail! You need a 16x slot for the fast GPU, a secong GPU, an 8x slot for the RR and a blackmagic output card plus a raid connection. And I may be wrong but I thought Resolve on Mac only handles up to 2K with 4K support limited to the full Linux version?

Anyway I hope the rumours are wrong but in light of Apples direction of late I suspect not...

[Bruce Allen]

Although this thread got a little derailed into talking about FCPX it does seem to be an indicator (or confirmation of what we suspected for a while from various indicators) that Apple are not so concerned with the needs of the very high end users.

Rumours are that although the next MacPros will have Thunderbolt and be rackmount they may have as little as 2 PCIe slots!
Perhaps Apple are betting the farm on Thunderbolt but it seems a little premature to me and even if they release an affordable rackmount PCIe expansion box that uses Thunderbolt any one card will be limited to the 4xPCIe speeds of any one Thunderbolt channel. Try running Resolve on that with a RedRocket card! Fail!

If I had only 2 16x PCIe slots, I'd do this:

Slot 1 - superfast GPU
Slot 2 - PCI extender w two 8x slots (redrocket & GPU 2)
Thunderbolt - RAID
Thunderbolt - backup RAID
Thunderbolt - Blackmagic 3D card

Or:
Slot 1 - superfast GPU
Slot 2 - PCI extender w two 8x slots (redrocket & RAID)
Thunderbolt - GPU 2
Thunderbolt - backup RAID
Thunderbolt - Blackmagic 3D card

Though to be honest, I think the CPU & GPU & software combo might be so damn fast that I won't need a RedRocket.

I'm fine as long as I can get interactive HD stereo. That's more than enough of a "proxy" to set creative intent. For the final 4K 48fps stereo render, I'm okay if it's not realtime!

And I may be wrong but I thought Resolve on Mac only handles up to 2K with 4K support limited to the full Linux version?

Mac Resolve is not limited. I trust them on this.

BTW next year, Ivy Bridge will have PCIe 3.0. 16x PCIe 3.0 = 32x PCIe 2.0

Bruce Allen
www.boacinema.com