FC7 Slower than FC6

[Chris Korrow]

This is a little bit troubling, I'm just about to upgrade comp & FC5, It'd be too bad to upgrade & then upgrade again ASA the new version comes out.

I usually skip an update.

-For reasons that remain a mystery for the moment, the newest version of Final Cut Studio (7) is significanty slower than previous version (6) -- at least where Compressor and Motion are concerned. Hopefully, this will be addressed by Apple's software engineers in a future point release.

Final Cut Studio 6 versus 7

[William Hohauser]

That is curious and I would wonder exactly how the tests were set up. QuickCluster needs to be set up in very specific ways or it behaves badly. Was the recent audio firmware update installed or not? That changed how Compressor worked on my 8-core machine with QuickCluster which I had to reset a different way.

Did they also try renders within FCP? I found those to be faster but not by much. However, I would wait for NAB next month to see if another version of Final Cut Studio is on the way. Whatever you decide, FCP 7 will be faster than FCP5.

[Chris Korrow]

Seems like since the results were the same on the imac that it must be a software issue.

[Nigel Barker]

There is simply not enough detail explaining how the tests were performed to know whether these results are valid or not. In particular there is no explanation of how someone else could try & confirm these figures.

[Chris Korrow]

Point taken, though it seems like barefeats is pretty reputable and independent tests help keep us informed. (i.e. audio bug in 09's)

Unfortunately I need prores for the a new project and can't wait for the next update.

Thanks William for letting me know that you notice no difference as far as day to day work, very helpful.

[Nigel Barker]

The reason that I am so suspicious of those results is that they show such a large difference in performance that it seems very unlikely that nobody else would have noticed that FCP7 is so much slower than FCP6. FCP7 has been shipping for about eight months so if in real life tasks using Compressor or Motion were taking literally twice as long I really think that we would have heard some shouting long ago.

Incidentally in my experience for editing FCS7 is faster & seems to spend less time rendering than FCP6.

[Chris Korrow]

Well once again, this is really good to hear, and sets my mind at ease about upgrading.

This is why I posted this here, - to get some level headed opinions.

Thank you!

[Robert Lane]

While there's a lot of detail missing in exactly how the test was setup the results really aren't off the mark. My own testing (beta FCP7) showed the same thing, that Compressor 3.5 was taking just a tick longer to do encodes than the older version.

And that's not totally surprising; FCP had a few more features added to it but nothing was done to improve the core engine that directs things like encodes, renders etc. across the board. It's one of the main reasons I've always suggested to those who have FCS2 to hold off on the upgrade path until the "real" FCS update is released - hopefully this year sometime.

The other thing is that I doubt RobART (owner of BareFeats) setup a virtual cluster. There's almost no need to really do that anymore now that the faster front-side bus is in command and there are so many cores churning away. Besides that, so few people who attempt it even fully understand how to it properly and as W.H. noted, if not done right you're in for trouble.

Until FCS gets it's much needed "full monty" rework internally to take advantage of all the things Snow Leopard has to offer (most notably things like Grand Central Dispatch) you won't see any performance increases from FCS2 to it's current iteration.

If you've got FCS2 there's no compelling reason to upgrade yet. If you don't have FCS at all then get on board and start learning it so when the "real" upgrade comes - soon? - you can then get that upgrade for cheap.

[Chris Korrow]

Thanks Robert,

I'm running the original FCS & a G5 PM. I'm hoping the new MP will be announced this Tuesday as I am in desperate need of an upgrade - was never really satisfied with the SP 09's.

I'll probably get the current FC since I kinda need it yesterday and run it for a while. Either way with a new pro it'll be leaps and bounds above what I've got now.

Thanks again to all, as that test made it seem way worse that it actually is from what you all are saying.

[Robert Lane]

Chris, as you're probably aware - and are hoping for - the next-gen of Mac Pros (and Macbook Pro's) is slated for release "soon". I definitely would not purchase a new system until they do come to market. There may be more going on "under the hood" in these new systems than we expect and if so would be worth the wait.

And no, not I nor anyone outside Apple's hollowed walls knows exactly what's coming but many recent changes in CPU/RAM/FSB tech has happened in the past 12 months that hopefully will be included in this next round of hardware offerings. Indeed, Steve Jobs made a hinting statement saying that these next systems would be "next generation..." tech.

Time will tell, but still worth waiting to see before dropping hard-earned cash.

[Chris Korrow]

Already got my projects etc. copied over to new HD & heck even got all my firewood cut for next year so yes, I've ben putting it off for a couple of months now.

I have hopes of an announcement tomorrow as do many. There have been 9 MP listed on Apples refurbished pg for the last few weeks & it seems like everyone is holding off to see what's coming up.

Faster CPU & GPU is pretty much a given, but I sure hope they they add some more ram slots in the SP.

[Nigel Barker]

Originally Posted by Robert Lane

While there's a lot of detail missing in exactly how the test was setup the results really aren't off the mark. My own testing (beta FCP7) showed the same thing, that Compressor 3.5 was taking just a tick longer to do encodes than the older version.

The results on BareFeats show Compressor 3.5 taking more than a 'tick' longer than Compressor 3.0. They show it taking around 50% longer (up from 72 seconds to 100 seconds on the Mac Pro). That is why I find it likely that these tests are either wrong because the testing methodology is flawed or don't reflect real life experience.

[Robert Lane]

As I say, I can't offer commentary about how RobART setup his testing - and you should email him directly if you really want to know what he did or didn't do - but again his publicly published tests shore up my own internal testing that in fact C-3.5 is slower in most encoding operations, sometimes it's literally just a tick longer and on other parameters it's significantly longer.

Again the hope is that when the full update to FCS is released that this and the plethora of other most-wanted/must-have improvements are completed by Apple.

[Chris Korrow]

Well since I'm getting ready to drop a chunk of change on a complete system overhaul and as I've been waiting for the new MP's to be released, I've spent some time researching what to expect in my upgrades.

Thanks to all & here's my assessment.

Since I consider myself more of an artist than a tech person, I appreciate the independent testing that folks do as it helps to keep the companies honest and me informed (Apple denied the audio bug till a fix was eminent).

As far as FC, - Seems like day to day performances are negligible between the two, but there is a good possibility that some tasks may take a little to considerably longer (no back-up testing yet to prove or disprove).

It seems like the performance issue is not a sweeping program wide issue (which someone like me may assume by looking at the barefeats results).

The bottom line is that FC7 seems to be an upgrade purchase that I can feel good about.

So once again Thanks.

It would be interesting to see other people running some tests though.