Switching to adobe

[Randy Painter]

Looking at switching away from apple (iMovie, FCE), to premiere plus. Have only used apple software for editing the last 4 years. Sold my Mac pro and built a nice westmere system. Got tired of apple's high prices on Mac pro's and lackluster updates. Haven't even used windows in 6 years. Is premiere better than FCE? Good selection of built in transitions and stuff. Will be working with AVCHD files. Appreciate and comments or suggestions. Tnks

[Bruce Watson]

Is one better than the other? That's more of a religious argument than I feel like addressing. Suffice it to say that the BBC thinks Premiere Pro is sufficient for their needs (IIRC they just bought 2000 seats or so). There's no doubt it will do the job, but it does it differently and may or may not fit the way you want to work. More personal choice, that one is.

For AVCHD files with Adobe, you'll want CS5, and for that you'll want Win7 64 bit. And for that you'll want Intel i7 processors. I get good (not great, but certainly serviceable) results on an i7 laptop with just 6GB of RAM. I suggest 12+ GB of RAM would be better. And at least three hard drives -- boot/programs, source/projects, and rendering.

I'm using the Production Premium package myself. I find it to be pretty slick, and if I keep learning I may one day succeed in scratching the surface of what this software can actually do. It's vast and powerful. I have no doubt that FCP, Vegas, etc. are more or less equally vast and powerful. It's amazing to think about actually -- wasn't too many years ago that most of this was done with razor blades and splice blocks, and the capabilities we take for granted weren't even dreamed off.

[Randy Painter]

Just built a nice dual xeon quad core westmere, 2.66 ghz, 12gb ram system with win7 64 bit. So that's covered. Wanted a Mac pro but for over 5 grand, decided to build it myself. A little under 3 grand. Just want to find a decent software package that plays nice with AVCHD and uses the GPU to speed up rendering. Apple is to busy with the i toys that the video software has fallen behind. Don't need top of the line software but don't want to struggle either. I enjoyed the timeline in iMovie 6HD, but the newer iMovie sucks flat out. Used FCE 3.5 a little but went back to iMovie 6. Would have stayed if apple would have kept the iMovie 6 layout and made it 64 bit, but didn't happen. Guess I can dowload the trial and see how it goes. Tnks

[Pete Bauer]

Originally Posted by Randy Painter

Guess I can dowload the trial and see how it goes. Tnks

That's the right answer. Just about all the NLEs have a trial version. In the case of Adobe, not all features (like MPEG decoding) work in the trial versions but most major features do so you can see if you like the workflow, which really is something nobody else can tell you. And besides, the responses to these kind of questions unfortunately usually are short on factual statements and degenerate quickly into "platform wars" which we don't do here. Best wishes.

[Ann Bens]

No point in downloading the trial it does not handle AVCHD.
You can buy the package with a 30 days money back garantee.

[Pete Bauer]

I disagree. When one switches software, it is still quite valuable to see if one is comfortable with the changes in interface and workflow. AVCHD is fully supported in the retail version and, as is widely reported, works extremely well on fast systems like Randy's. That won't be the "buy / don't buy" issue because it is a given.

Sure, it would be preferable from an end-user standpoint to test-drive the full version, but even though that's not the case, there's a lot to be gained from a test drive using a different format.

[Mikel Arturo]

May be this helps a little:
Adobe CS5: 64-bit, CUDA-Accelerated, And Threaded Performance : How Should You Accelerate Adobe?

IMHO, CS5 is the best version of Premiere Pro. I have worked with Premiere since 4.2 and this version is rock solid, fast and "easy scalable" (because is easier to change your graphic card than change your CPU and then your Motherboard, ram, etc.). And Adobe says that Mecury will go better with future graphics cards+better Nvidia drivers+Premiere updates.

HDV editing is fast as editing standard DV with CS4 (without Mercury, on a quad core system with 8 Gigas RAM). It's equal. Well, if you edit in SD you can't believe the speed.

If its true that a mere Nvidia Gforce 240 can activate CUDA acceleration, you can't go wrong (I don't have tested this, but you can read the forum to get more info).

And one big thing: don't close eyes to ONLY one video app (this is an Apple users fault, with Final Cut). You can combine FC & Premiere, and select one for every work, or mix them.
My combo is Premiere CS5 & Corel VideoStudio X3.
For fast work, I use Corel.
For hard work, Premiere+After Effects+Photoshop+Encore (well, THIS IS PREMIERE PRO, all the apps. working together).

[Andrew Clark]

Originally Posted by Randy Painter

Just built a nice dual xeon quad core westmere, 2.66 ghz, 12gb ram system with win7 64 bit. So that's covered. Wanted a Mac pro but for over 5 grand, decided to build it myself. A little under 3 grand.

** Hey Randy, glad you saved the $$ but, just to be fair to the Apple campers......you're not going to get anywhere near the quality build, neatness and sturdiness of any Windows only tower computer as you would with a MacPro tower. I know, I've searched and as a previous 2009 MacPro owner, there is nothing out there that competes with Apple in the quality and craftsmanship they put into their (hardware) products. Yes it's true they don't have the latest and greatest connectivity (USB 3.0, e-SATA) and graphics card options, but hey, you can't have it all....then again...

As for their software....well....not bad....but not great either....at least not yet.

[Randall Leong]

Originally Posted by Mikel Arturo

If its true that a mere Nvidia Gforce 240 can activate CUDA acceleration, you can't go wrong (I don't have tested this, but you can read the forum to get more info).

Yes, my result with a GT 240 is now on the PPBM5 results list on the PPBM5 site.

One caveat with the GT 240: Go with a 1GB DDR5 version, not the DDR3 version or the 512MB version, for best results. The result I put up is the top overall performer (382 seconds) on the PPBM5 list among those systems with only 6GB of RAM. The disk setup on that system could have been faster, however.

[Peter Chung]

Nice results, Randall! Thanks for submitting your benchmarks. Amazing that your measly GT 240 with only 96 cores beat out some systems with the GTX 285, GTX 460, and GTX 470 in the GPU benchmarks! Your system doesn't seem like no rust bucket, does it?

What I don't understand, though, is how your system beat out the two systems under yours, an i7-980X with a seemingly better Disk setup, and an i7-975X in CPU and Disk I/O. They both utilize RAID and have separate disks for OS, Project, and Outputs, whereas your system only uses 2 drives with no RAID. What is going on here?

[Randall Leong]

Originally Posted by Peter Chung

Nice results, Randall! Thanks for submitting your benchmarks. Amazing that your measly GT 240 with only 96 cores beat out some systems with the GTX 285, GTX 460, and GTX 470 in the GPU benchmarks! Your system doesn't seem like no rust bucket, does it?

What I don't understand, though, is how your system beat out the two systems under yours, an i7-980X with a seemingly better Disk setup, and an i7-975X in CPU and Disk I/O. They both utilize RAID and have separate disks for OS, Project, and Outputs, whereas your system only uses 2 drives with no RAID. What is going on here?

I discovered that the Velociraptors (at least the 150GB and 300GB versions) have issues with being run in RAID arrays, while the i7-975X is using only 5.0.1 rather than the 5.0.2 that both my GT 240 system and the "underperforming" 980X is running. And in my tests 5.0.1 consistently performed slower than 5.0.2, especially in MPEG-2 DVD encodes. Also, the 975X probably used first-generation 1TB hard drives with 200GB platters (the first edition of the Hitachi 7K1000), which never had high sequential transfer speed to begin with (while my particular 1TB drive in that GT 240-equipped system is a WD Black with 334GB platters).

And further down the list is a 980X with a 483-second overall result that had seemingly highly-ranked scores in everything - but the disk score is really slow. That system claims 7200 rpm speed on the 2TB hard drive. But since there are currently only three "true" 2TB 7200 rpm SATA hard drives on the market (the Hitachi 7K2000, the Seagate Barracuda XT and the WD Caviar Black), it's more likely that that 980X system with the molasses-slow disk score used an early 2TB WD Green, which has never had outstanding sequential transfer speed to begin with. (I knew this because my particular 2TB WD Black came through with a AVI Disk score of only about 105 seconds in my "Steamer" rig compared to the 400+ seconds of that Green.) Furthermore, although the WD Green claims a rotational speed of between 5400 and 7200 rpm, in reality it spins at just above 5400 rpm.

Even more amazing is that the GT 240 system of mine is equipped with only 6GB of RAM while the two that had performance deficiencies that ranked just below me had twice as much RAM (12GB).

And my system is nicknamed "Rust Bucket" because it's currently housed in a case whose thumbscrews are no longer holding in the side panels (the holes became too enlarged to fit those thumbscrews securely, as it often happens with lightweight aluminum cases).

I also looked at the very top scores, and discovered that two of the top four overall systems used less-than-optimal memory configurations. One of the systems apparently used 14GB via five DIMMs (probably mixed 2GB and 4GB DIMMs set up such that one of the 4GB DIMMs is installed into the channel that has one slot deliberately left empty, so that 12GB gets mapped to triple-channel and 2GB gets mapped to single-channel) while another most likely used four 4GB DIMMs in a Flex memory controller mode. In both cases, some of the memory had been running in triple-channel and the rest in single-channel.

[Harm Millaard]

Randall,

The scores you reference, Studio North and BillG, with suboptimal memory configurations were done intentionally. We have no explanation for this behavior, but upping memory beyond 12 GB makes a huge difference, but the unexplainable thing is, that 14, 16 or 24 GB makes no discernible difference. Studio North and BillG have done the test with different memory configurations, but the only performance gain came from more than 12 GB and for the rest made no difference.

Why? I don't know.

Jim's results are under investigation but he has now run the test twice with identical results. He will send me additional info tonight to see what is going on here.

[Greg Laves]

Macs have the latest greatest everything but don't have USB 3.0, e-SATA and the greatest graphics card options. When Macs started to really improve they started using more PC hardware, like Intel chip sets. As time goes by, Mac will introduce more PC hardware. I have edited on another friends Mac and he always cautioned me to manually save every 10 minutes in case his crashed. Why would he say that if they are so perfect? I like Macs but they are just another tool. For the money I would spend on a Mac, I could have a really awesome PC, my CS5 Production Premium upgrade and a new HD DSLR.

[Harm Millaard]

Not intending to start a PC-MAC dispute. I think Peter made a very correct post on this topic and for the rest it is a very personal choice what fits you better.

OTOH, I just wonder how many people know what MAC stands for?

One says great build quality, the other says something else, another one says limited hardware options, I say: "Maximum Amount Charged" (MAC) on your credit card.

[Randall Leong]

Originally Posted by Harm Millaard

Jim's results are under investigation but he has now run the test twice with identical results. He will send me additional info tonight to see what is going on here.

As of now, I have just seen Jim's results. It turned out to be something funky with the settings on his system. His retested system (at 264 seconds total, with a disk time of 118 seconds) now ranks just below my "Steamer 3.67" rig (which was originally built as the original "Steamer" rig with an ATi graphics card and 6GB of RAM, and has since seen several major component upgrades). He would have beaten me had he gotten a RAID 0 disk array for the project disk. But even then, his value rating would have been noticeably below mi "Steamer 3.67" system because I paid just $200 USD for my i7-920 while the i7-980X still sells for close to $1,000 USD. The current price difference between the i7-950 and the i7-980X can almost pay for an eight-disk (8TB total) SATA RAID array with the least-expensive Areca hardware PCIe SATA RAID controller card that can handle at least eight disks.