FCS 3 (FCP 7) Wish List

RyanBellaCine

Well-known member
With the recent release of Premiere Pro CS4, It has gotten me thinking about what FCP can't currently do. And what I think it probably should be able to do. Going beyond the lack of native P2 & other new tapeless formats, what else would we want to see in the next generation of software?

For me, the foremost criticism of FCP I would have right now is the lack of multi-tasking support. With thorobread Multi-Core Mac Pros, this should be feasible to say the least. A few points I have:



Why can't I capture video in the background while still being able to edit other footage already captured? Or continue working on another project? Tape editing is dying. but it's far from dead. DVNTSC is the majority of what I work with. And the forecast is that it's here to stay for a few more years. Being forced to Hurry up and wait while a tape captures is a huge time waster. The computer can handle doing both. So why can't the software take advantage of that? Even as I type this, Final Cut is busy capturing an hour long MiniDV tape. I'm "stuck" browsing the net ;~) when I could be working on and preparing the last tape I captured if only FCP allowed this.

Real Time effects rendering. Not talking about the dynamic playback. I'm talking about actually rendering footage. When I press Apple+R or Option+R, Why should I be stuck waiting for the render bar to finish? Can't it render in the background while I continue working? And if it can do that, why not make it automatic? I shouldn't have to tell it to render at all. I have at least 4 processor cores which could handle that task no sweat while I continue working.

Exporting: Again, when exporting a movie it no longer takes much computer horsepower, depending on the format of course. But with multi-core machines, FCP should be able to do this in the background. let me move on to the next project.

For me those key things that increase efficency and take advantage of multicore systems. And if they don't show up in the next generation of software somehow, I think it would be a huge negative critism. I don't see a reason why software can't do this.

Perhaps we will see the new version of FCP once Snow Leopard goes gold master. From what I understand, The OS may have much to do with the efficency of multi-threading software. And Snow Leopard is said to really attack the multicore programing issues.


One thing I do like about the current FCS is the virtual clustering in Compressor 3. I use it religiously. And for the most part it works great and is a huge asset. We have 4 Mac Pro editors here linked together via Xsan & Xserve on a fiber cable network. Having the ability to borrow a few cores from the other 3 machines, when the other editors using the machines don't even notice a thing is just awesome. I can render our streaming h.264 files in less roughly 1/3rd the time it would take if i had to export from within FCP. Which doesn't even multi-thread h.264. (which it should!!) And eveyone else can continure working without a hitch. Brillaint!

I'd be interested to hear what other folks are looking to see in a new generation of NLE software.
 
Last edited:
Does CS4 allow you to capture in the background and continue editing...WITH THE SAME APPLICATION OPEN? The application is busy...you can't have it do two things at once. If you think you can, why don't you prove it by learning how to code and doing it yourself? Avid doesn't do this either...nor Media 100 AFAIK...does CS4? That'd be a trick...

Rendering while continuing to work? Again, the APPLICATION is busy. it isn't about the processors and having more of them working, it's about the software you are using is already doing something. Again, does another NLE do this? Renders are quick for me. And if they aren't, it's an excuse to get up and stretch and MAYBE walk to the bathroom or something...get a drink of water...get my mind out of things for a moment. Use these moments in positive ways...not swearing at the computer.
 
Capturing in the background while you edit will severely tax your hard drives--that's where the bottleneck is. You need to decide which is more important--an ironclad capture without dropped frames, or squeezing a few more edits in while you capture. Granted, DV isn't intensive for drives, but start accessing all kinds of media and audio/etc... while you edit will.

You can't just point to the unused processors and expect them to provide all the necessary horsepower to do more with your system. You need to consider everything that can cause a bottleneck and take that into consideration. Plus, if you're going to encounter dropped frames--does it drop the frames from the capture or your playing back an edit? Either one will be unacceptable, and you'll find yourself back where you were before--capture first, then edit. I agree with Shane--I don't know of a single NLE that allows you edit while you capture. Multicore rendering would be great, but that might require a rewrite of plugins.

Regarding background rendering--totally possible today. Avid Liquid does it, and I believe there a few others--Vegas, maybe? I use Avid and would love to see background rendering, and it should be possible for both Avid and Apple to implement, given enough time to code it. Is it easy to do? Can't say, I'm not a programmer, but it exists in other apps so it's at least possible somehow. Just maybe not with FCP or Media Composer's code base--may require a complete write of the app. Time will tell there.


Michael.
 
Does CS4 allow you to capture in the background and continue editing...WITH THE SAME APPLICATION OPEN? The application is busy...you can't have it do two things at once. If you think you can, why don't you prove it by learning how to code and doing it yourself? Avid doesn't do this either...nor Media 100 AFAIK...does CS4? That'd be a trick...

Rendering while continuing to work? Again, the APPLICATION is busy. it isn't about the processors and having more of them working, it's about the software you are using is already doing something. Again, does another NLE do this? Renders are quick for me. And if they aren't, it's an excuse to get up and stretch and MAYBE walk to the bathroom or something...get a drink of water...get my mind out of things for a moment. Use these moments in positive ways...not swearing at the computer.

Vegas allows you to open as many instances of the program as you need and can do both these things "simultaneously." I work this way on a regular basis.
 
OK...background rendering I'll give you. But still, isn't that asking much of the hard drives? Writing and reading at the same time. Surely rendering will slow down if you do this, or you'll get dropped frames while playing back footage.

HARD DRIVES are the bottleneck...forgot about that.
 
While this isn't the idea situation for everyone, Here, we have a Fiber Cable Xsan server which is built for multiple simultaneous editors crunching the drives at once. So yes, hard drive workload & bottlenecking is an issue. But there are solutions out there already to address that.

But Even if not everyone can take use of such a feature right off the bat, there are situations that can take advantage of such features. I dont see any reason why Software technology can't be pushed to the limit as long as the systems can handle it. Technology is one area where "trickle down" actually works... Eventually it could become mainstream

While I'm not even going to claim to know squat about programing, As others have pointed out, it can be done. On the non video side, Lightroom has been able to render previews, AND export "prints or final versions" in the background all while still allowing continuous work. Lightroom is a non-destructive software program. As are NLE's. And again, Apple has stated they are seriously looking at the issue of multitasking on the OS level. Once they solve that, it might open the flood gates for applications to follow.

It's well agreed on the technology front that the next big steps in computing power will involve an explosion of multi-core processors. So it makes lots of sense to put all the long hours of investment into software programing to take advantage of them and make them multi task able.
 
Last edited:
There shouldn't be an hardware reasons why you can't render in the background or capture a single video stream in the background, IMO. W/o missing a beat I can rip a DVD, transcode stuff in Compressor, and work in a large (100+ of DV footage hrs) project in FCP on a 'lowly' 4-Core machine w/5gigs of RAM, and 3 1TB internal SATA scratch drives. The hurdle, I think, comes on the software side and enabling FCP to be a good 'traffic cop' so that these processes are properly regulated. I'm hopping that when we see a from the ground up re-write of FCP (the current core of FCP is from '98/'99) we'll see a lot more flexibility in the program than we have now.

Shane, I don't think having advanced programing skills should be a prerequisite for making an upgrade wish list.:beer:


-A
 
Lightroom being able to do this makes sense...that is a photo editing application. Video editing demands much more processor power and drive speed than photo editing...a lot more.

Encoding video and CAPTURING video are two very different things. Encoding can be handled by a processor or a bunch of them...and it is simply rewriting data from one place to another...nothing really intensive writing wise. But capturing video..writing that to disk, and then asking the same hard drive to also play back video is asking a bit much of the drives.

But...whatever. Ask away. I just want the basics figured out better. CORE stuff needs a little tweaking still.
 
Oh I should also mention that the work we do does not involve short films. Most of our work involves hours and hours of continuous meeting & conference footage which needs to be transcoded, edited, and streamed. While we do short 3-5 min pieces as well, mostly, the time it takes to capture, render & compress for streaming/ DVD is huge. Tying up our Editors on a 25min -2 hour long export or render is a huge loss when we could be working on the next project's timeline in the meantime. I'm sure our situation isn't the only one like it in the production world.

EDIT:


As far as Capturing while while editing, I've already had Compressor 3 run a Virtual cluster farm where my own machine was using 7 instances, borrow another 7 from another editor, and also borrow 2 instances each from another two editors, WHILE THOSE EDITORS ARE BEING USED WITH FCS. I don't know what the other editors were working on exactly. But they never noticed a thing. ( I certainly would have heard about it since I'm the go to guy when ever there is a problem) So i know it can be done horsepower wise. On my machine I'm also sometimes burning a DVD Studio Pro project, Working with Photoshop CS3 and browsing the net... all at the same time! Not so much as a hint of struggle - Mac Pro 1st Gen 8 Core 3Ghz with 16GB RAM, Xserve RAID, and local FW800 LaCie drives.

This all was happening while all computers were using the SAME XSAN XSERVE RAID. so all of the 18 instances in Compressor 3, and at least 3 FCP sessions running, did not phase the hard drives one bit. It can be done. Maybe not for the low budget, freelance, or shadetree editor situations, but what reason is that NOT to push the envelope of technology further? Where the support base is already here in some form or another.
 
Last edited:
Encoding video and CAPTURING video are two very different things. Encoding can be handled by a processor or a bunch of them...and it is simply rewriting data from one place to another...nothing really intensive writing wise. But capturing video..writing that to disk, and then asking the same hard drive to also play back video is asking a bit much of the drives.
Obviously drive speed and video format would factor into how well this would work, but shared storage systems have multiple clients reading and writing to the same drives at the same time w/o problems. If your setup meets the hardware requirements the key would be FCP being a good 'traffic cop' and always giving the capturing video stream 'priority' to the HDD even if that means reduced RT functionality while the capture is going on.


-A
 
Have we forgotten about the SmoothCam filter? The filter works on a background process (spawned by FCP) and allows you to continue working in FCP. From a user perspective nothing changed and they don't see any new application starting up.

FCP was not written as a concurrent application that is able to use multiples cores properly. While the OS can split FCP processing across multiple cores to spread the load, the application itself does not take full advantage of the availability of multiple cores by creating multiple work processes that can offload certain task without interrupting the user.

Not only does this lock the user out during rendering, etc., but it also limits the full power of the application as there are a lot of tasks that can be sped up by splitting them across multiple processes. In theory, rendering and capturing can be offloaded. At the very least, rendering can be done through the existing QMaster and Compressor infrastructure (which is concurrent and makes good use of multiple cores).

Unfortunately, in reality it is very, very hard to write a good concurrent application, especially one so large as FCP. In particular, concurrent applications can be very hard to debug and certain edge case bugs can be very hard to spot and test for. But, FCP will no doubt become more "aware" of multiple cores in future releases.
 
Buy another computer if you need to capture and render while editing at the same time- helping to bring you tomorrow's FCP today.

Noah
 
Now why would I want to spend another 7k on a new tower ( not counting decks & rewiring costs for another XServe/Xsan Fiber Cable setup) when a cheaper software update would solve that? Besides, your completely missing the point. Its about what software can do and how it can evolve. Not about finding a workaround for right now.
 
OH MAN....Cheez...when I first read your response, I didn't see a "t" in front of "tuck," I saw an...uhm...another letter. Making it a VERY different word.
 
what I really want to see in FCS3 is:

1: fully backwards / forwards compatible. heaven forbid I can save a project in FCP 7 & have someone in FCP 5 open it w/o having to jump through 18 hoops. (if you can do this w/ MS Word, Photoshop, etc. why not final cut PRO?)

2: full support for M2T files in the log & transfer (ala Sony's HDV CF card recorder)
 
Instant edit button- for those projects you get talked into editing...

Make it better button- when you know your shot was crap the minute your looked in the viewfinder, it would reassure the client to know that you can always fix it in post...
 
Back
Top