Forums | MacLife

I imagine the compressor results are due to elements of PPC Quicktime that haven't quite made it over to the universal binary version.

MPEG-2 encoding is slower. There's some news...

AltiVec wasn't all RDF.

Last edited by andgarden (2006-04-13 7:19 am)

These results really show the Core Duo's true colors when running properly optimized software.  I think Apple claims a 2-3x advantage over the PowerBook in FCP, and that might actually be a fairly reasonable claim.

It may be over a year old, but it still costs $2,000.  Even the low-end MBP, also $2,000, should be able to hold it's own against the G5 (if you include Motion, otherwise it's still pretty tight).  In the computer world, desktop systems are supposed to offer a significantly better value than notebooks, not the other way around.  ATM, the G5 might still the better buy for a lot of users, but only because of Rosetta and expandibility.  Also, as the article mentions, FCP has had five generations of optimization on the PPC, versus the Intel's one.

Last edited by Mr. T (2006-04-16 9:28 am)

Mr. T wrote:

It may be over a year old, but it still costs $2,000.  Even the low-end MBP, also $2,000, should be able to hold it's own against the G5 (if you include Motion, otherwise it's still pretty tight).  In the computer world, desktop systems are supposed to offer a significantly better value than notebooks, not the other way around.  ATM, the G5 might still the better buy for a lot of users, but only because of Rosetta and expandibility.  Also, as the article mentions, FCP has had five generations of optimization on the PPC, versus the Intel's one.

Apple has been making QuickTime for x86 for quite some time, I doubt they do not know how to optimize it. And saying that, FCP's video engine is based heavily around QuickTime, so.... Lets just say I hope Intel's desktop variants live up to the hype.

-mark

avkills wrote:

Mr. T wrote:

Also, as the article mentions, FCP has had five generations of optimization on the PPC, versus the Intel's one.

Apple has been making QuickTime for x86 for quite some time, I doubt they do not know how to optimize it. And saying that, FCP's video engine is based heavily around QuickTime, so.... Lets just say I hope Intel's desktop variants live up to the hype.

All I was really saying in that last sentence is that performance improvements are often cited as features in product upgrades.  I don't know for a fact if this has been the case with FCP, but any such improvements likely wouldn't have been too significant anyway, so it probably wasn't worth mentioning.

Anyway, QuickTime was a unibin right from the start.  If optimizing FCP for Intel was as trivial as optimizing QuickTime for osX86, it would've debuted on the same day as the Intel iMac.  But it's not that simple, and Apple didn't want to release an underoptimized professional product.  Secondly, such a statement implies that Apple consciously crippled the PPC version of FCP by not properly optimizing the PPC version of QT, and that's pretty hard to believe.  If Apple could've had double the performance in FCP all these years, they get my nomination for the "Idiot of the Decade" award.

MacBoy4139 wrote:

And QuickTime for Windows is faster than on OS X.  Especially in the encoding for iPod.  In fact, I wouldn't be surprised if on a MacBook Pro, QuickTime encoding would be faster when booted under XP (through Boot Camp) than on OS X.  It's not always the hardware.

You're comparing OS X (PPC) encoding to Windows (x86).  Since there's two variables that are changing, you can't conclusively say that the performance improvement is due to "OS technical nuances," which, in most cases, is quite insignificant.

Last edited by Mr. T (2006-04-17 3:33 pm)

MacBoy4139 wrote:

And QuickTime for Windows is faster than on OS X.  Especially in the encoding for iPod.

My reply was to the first part of your post, where it seems like you were comparing two different OSes on two different architectures.  I only say that because, in the second part of your post, you speculate as to what you think would happen if you ran tests in both OSes on the same hardware.

In other words, it seems like your conjecture is based on the fact that QT on Wintel is faster than QT in OS X (PPC).

QT7 is indeed Cocoa according to this Arstechnica article.

It actually says that QuickTime Player was rewritten using QTKit, which is an Objective-C interface for QuickTime.  Regardless, it's possible that there might be a performance gap on identical hardware, but we won't know for sure until someone tries it.

The point is QTKit is just a Cocoa (Obj C) interface for QuickTime.  QuickTime itself wasn't completely rewritten in Cocoa, just the player was.  However, since QTKit's functionality is derrived from the QuickTime API, even the player shouldn't really be vulnerable to Cocoa's inherent performance loss.

I won't deny that (I was only considering QT's playback capabilities, esp h.264).  I'm just meant that QTKit should be fast.  Anything the App does on it's own will certainly suffer.  BTW, I hope Compressor is faster than QT.