-
FCP-X and the Odd Couple intertwine.
Have you noticed that part of the confusion of X is that one person says it’s “fast” then another says it’s “slow.” One says it’s “shallow” and another calls it “deep.”
It delights one user and drives another crazy.
Why?
I have a theory I’d like to toss out for discussion.
I’ve been moving projects around lately because, my 6 terabytes is maxed out and I’m doing so much new work that I’ve been running out of disk space in my system. I’m using the Duplicate and Move commands a lot – and I’m also spending a whole lot of time … waiting.
The process has been an odd combination of, frustration, exploration, anger, relief, pleasure, and madness. (Actually the only real pleasure is that when I finally can’t figure out if X is still working or is hung up – when I re-boot – it’s still re-booting gracefully without losing any progress!)
The big problem is that simple things that I used to be able to do relatively quickly under the Finder – seem to drag on and on under the FCP-X. And while waiting for these operations to get done I started asking myself why.
It reminds me a lot of the frustration of copying CF cards from my camera via Disk Images. Say I have a chock full 16 gig card and disk imaging it takes 5 minutes to complete – if I then ask it to image another card the same size, but containing just one single 10 meg photo – making a disk image of that might take three and a half minutes!
It’s weird, but the “file copy overhead” is such that there’s little connection between the size of the file and the time it takes. I suspect that the formatting, preparation, structuring, file copying, confirming, – whatever – is the big time sink – and that the data actually being moved isn’t really that big a deal.
Is something like that going on under the hood in how FCP-X handles it’s database construction and proofing?
In the day after my 10.0.2 to 10.0.3 upgrade, I spent 24 hours watching beachballs and getting frustrated while X seemed to take forever to process all my existing projects and let me get back to work. Then, nearly magically, after a day of misery – suddenly, things went back to being quick. That was weird.
I’ll freely admit that I don’t understand enough about the under the hood operations of modern software to more than wildly guess what’s going on, but I’m starting to suspect that FCP-X hates disorganization. I’m beginning to suspect that FCP-X has a true split personality. In some areas it’s the Felix Unger of NLEs. (see: Neil Simon’s “The Odd Couple”) ridiculously rigid about demanding everything be organized and in place before it’s happy – but then transforming into Oscar mode at other times, happy to face disorganized file types and to have you toss a bucket of them into a timeline and to not particularly care how or why you arrange and re-arrange them.
Looking back at the beachballs and delays – they seem to come up when I’ve told X to do something, changed my mind, tried to “un-do” it and then do something different.
If I was writing the Odd Couple” scene describing this, I’d have Oscar run into the garage and tell Felix to move all the red boxes from the shelves in the garage into the car in the driveway. Felix would start methodically moving all those red boxes onto the floor of the garage, preparatory to the move. Then Oscar runs back in and says “not the red ones, the green ones!” and runs back out.
What does Felix do? What he does NOT do is ignore the red boxes and simply start moving green boxes. No way. Consistent with his character, he’d have to methodically re-stock all the red ones to their original locations before he’d even consider moving on to the green ones. In fact, he’d not just re-stack them, he’d make sure the stacks and rows are perfectly even before he’d even touch a red box.
This is starting to seem to me like FCP-X’s approach to the behind the scenes ordering of it’s database structure. And I’m starting to suspect it’s why it sometimes seems so painfully SLOW in some things.
If you do things that require it to re-order the database – particularly if those things might affect other parts of its relational system operating behind the scenes, – things can bog down fast.
The difficult thing to understand is that it might be this same strict organization requirement that makes it so fast in other operations.
In real life, we all understand that like a well organized tool bench, when things are put away properly, they’re easy to find and use subsequently.
We all have our own individual tolerances for disorder.
I think FCP-X’s database hates disarray. And will spend time making sure everything behind the scenes is orderly, so that when you want it to go grab “this” clip- it confidently knows exactly where to look.
I’d love to hear from others as to whether this fits into what they’re experiencing.
Thanks for reading.
“Before speaking out ask yourself whether your words are true, whether they are respectful and whether they are needed in our civil discussions.”-Justice O’Connor
-
25 Replies
-
Hey Bill,
I’m been avid visitor of the this forum. Been using FCPX since the release. I do agree that FCPX is very picky on the organization of projects. I currently, moved a indie feature I’m editing from 7 to 10. Why you say? I was comfortable enough with X to try. Plus, I wanted to be able to refer back to the original raw files a little easier. Before I got on board as editor the files were transcoded to ProRes4444 which was crazy for H.264 files. Initial the transfer went well using 7toX app but as soon as I started to moving files around manual and duplicating projects without thinking the system started to bug out. I’ve spent the last 3 day reorganizing my files. Now I have everything very organized. FCPX is purring like a kitten hot knife thru butter. Big lesson be very organized and if you’re moving large amount of files or transcoding do it after hours so you don’t keep stopping processing in the middle.
-P
-
Performance wise, 10.0.3 has been a giant step down for me.
You can watch this happen, FCPX tries to save every keystroke. You can read it in the manual, if it’s not every keystroke or “detected change”, it’s every 15 minutes by default with no way to change any of this behavior via preferences.
When the Event or Project changes, FCPX attempts a duplicate and save. It seems that’s it’s not smart enough to wait to save, and then confuses itself when it has to do multiple things. I mean try saving after every keystroke in any other program, NLE or not, and see how that goes. It’s not very efficient.
I, too, dont understand the inner workings, I can only theorize.
I think that Apple is trying a new way of working and it’s not quite there. I am sure it was well intentioned, and essentially we are working right through the middle of debug mode.
This gets even worse on SAN configs as they can have some trouble writing a bunch of little files, quickly.
The fastest way I have experienced to move FCPX structures are via Finder after quitting FCPX.
Despite all the FCPX “quirks”, instability will be the one singular trait that will truly drive me away. Reliability is very important to our shop. I am not feeling very reassured at this point, hope it gets better.
I don’t think that FCPX dislikes disarray, I think that it’s trying to be too smart and doesn’t know when to play it cool.
Optimization.
Jeremy
Please excuse any iOS misspellings.
-
X seems to be designed to work one of two ways. Either work on a single production at a time and then consolidate and archive. Or, always have everything on-line in your two libraries, a la iMovie, iPhoto or Aperture. Neither approach seems very viable over the long run in the real world. Ironically, Apple had a functioning solution in FC Server. Similarities between it and X are purely cosmetic and cursory. X feels like the programmers were aiming at a mash-up between FCP “legacy” and Avid media management and haven’t got it right yet.
Oliver
Oliver Peters Post Production Services, LLC
Orlando, FL
http://www.oliverpeters.com -
[Bill Davis] “The big problem is that simple things that I used to be able to do relatively quickly under the Finder – seem to drag on and on under the FCP-X. “
Bill, let me be the first to congratulate you on finally acknowledging that even for you FCPX is less then perfect. I can now take your praise of X a little more seriously.
Herb Sevush
Zebra Productions
—————————
nothin’ attached to nothin’
“Deciding the spine is the process of editing” F. Bieberkopf -
yea it sounds to me like thats making excuses for buggy software.. dont get me wrong, 10.3 was a big leap in the right direction with the obvious additions of multi cam and beta monitoring, but its still buggy as all hell.
Neil Goodman: Editor of New Media Production – NBC/Universal
-
[Neil Goodman] “and beta monitoring, but its still buggy as all hell.”
Really?
I’d be interested in your reasoning supporting the above comment Neil.
I’ve finished and delivered 17 paid projects to date and not one of them missed deadline because of anything I can attribute to the software.
Yes, actual learning is required to operate this software. And no, it’s not perfect. But to describe it as “buggy as all hell” nearly totally miss-represents the state of the software in my experience as a daily user.
To my thinking “bugs” are repeatable flaws in software code that are reproducible, universal, and stop the user from being able to do things they need to do. “Bugs” technically aren’t missing features, or elements that software A does that software B does not.
Do you use it regularly? If so, where are these “bugs” causing you grief? Are they show stoppers, or merely annoyances? Have you missed deadlines?
We’re all here essentially learning about the software so I’d like to know what you can factually contribute to the discussion.
You say you’ve experienced this software as “buggy” and I’d like to know how and why.
Thanks.
“Before speaking out ask yourself whether your words are true, whether they are respectful and whether they are needed in our civil discussions.”-Justice O’Connor
-
Since 1.0.3 I’ve found the spinning beachball moments to be almost intolerable. I’ve been scrupulous about optimizing drive and I always have plenty of RAM available.
There seems to be a major flaw with the constant writing/saving. Also booting and loading the projects and events seems to take longer than previous versions.
It feels like it’s thinking too much and has lost a lot of responsiveness IMO
And then there’s the issue of having to re-render some projects every time you open the project.
-
[Bill Davis] “I’ve been moving projects around lately because, my 6 terabytes is maxed out and I’m doing so much new work that I’ve been running out of disk space in my system. I’m using the Duplicate and Move commands a lot – and I’m also spending a whole lot of time … waiting.”
I haven’t seen any kind of slow down in 10.0.3 yet, but so far I’ve only been working in a couple of events. One has 18 1 to 3 minute projects in it and the other is a 7toX transfer with several 90 to 120 minute projects. I’ve been using the disk image method of organization on my 6TB RAID 5. Each job has its own sparseimage containing the event, project, media, etc. I mount only the sparseimage I need to work on. When I’m done for the day, I use Carbon Copy Cloner to save the changes to my mirrored archive RAID. At that point, I have an identical sparseimage in 3 places. If I ever need to move a job, I just copy or move the appropriate sparseimage with the finder. I can work in FCPX while that is happening.
Jason Jenkins
Flowmotion Media
Video production… with style!Check out my Mormon.org profile.
-
For me, X became nice and breezy when I turned off the organizing, transcoding, and analyzing on import and the background rendering. It was real useable after that! Lost some of the selling points, but oh well. Just my two cents.
-
IMO, Bill’s initial post was well thought out, and like all of you, I have no inside knowledge about how the database structure works. Having said that, my intuition has told me from day one that the database is the 800lb. Gorilla in the room… that it controls everything… i.e. “traffic cop” style.
I’ve felt (without proof) that it was(is) responsible for some features being missing initially, and possibly to this day. By that I mean, until it can be developed in a way that is “database compliant” it won’t be included. We have been asking for the ability to keyframe the color correction filter from day one, yet we still seem to be waiting. I can only speculate that they haven’t figured out how to do that within the universe that is the traffic cop database?
To Bill’s definition of a bug, which seems a bit constraining, I would suggest that a bug can be intermittent and not always reproduceable. If it was, chances are it would have been caught prior to release. The AutoSave bug in FCPX is one of those. If it was always 100%, they would have fixed it by now. The titler bug apparently was, and has been resolved. I also think a bug can be quite unacceptable without stopping the program completely.
FWIW, I do seem to be getting more spinning beach balls under 10.0.3…even with a new project, but it may be something unique to my environment… I honestly don’t know. I think that their perceived increase in frequency may be why some people think 10.0.3 is slower?
I also take heed to Paul’s comment that you might want to be careful about when you decide to starting doing things that might trigger an internal reorganizing. While not a problem in the off time, it could be a disaster if you’re under a tight deadline. Some things in FCPX might be best left to after the project is delivered?
Mitch Ives
Insight Productions Corp.“Criticism may not be agreeable, but it is necessary. It fulfills the same function as pain in the human body. It calls attention to an unhealthy state of things.” – Winston Churchill
Reply to this Discussion! Login or Sign Up