Page 28 of 38
Posted: Wed Mar 16, 2005 7:37 am
by Machinesworking
sniffio wrote:Adam Jay,
which OSX do you prefer the best for Live4 ?
CAn it be either 10.3.5 or 10.3.6 ?
10.3.6 is a total dog, 10.3.8 doesn't seem to have too many distracters.
The optimization will work,
it's just that the Live4's meter won't show it.
You can see it in Activity Monitor.
It will save you more RAM, too.
But if the meaning of optimization=Live4's meter,
guess my won't work.
The meter is not that stable if you know what I mean.
Sniffo, think about it, cutting down the size of the OS can't really save you any noticeable amount of RAM. I've done this test, total scientific approach. Tested before and after optimization with a "typical song for me", altivec fractel carbon like the author recommended, and the Live 4 test.
All three tests came back exactly the same, no noticeable improvement in performance!
I also tested each with just the clean system instal, before removing fonts and languages, then defragging the HD, no difference.
I reduced my samples to 88 to 44 samples=0ms
If I did not do the optimzation, the audio glitch will be there,
although the % will rise.
With the optimization, yes, the sound quality will rise.
I recommend the optimization for the users who need the extra edge of the power for real-time-use.
Otherwise, I wouldn't even bother to notice it.
What I think you're experiencing is a clean system instal VS a dirty system. That and from what I can tell, most of the people who are experiencing dropouts aren't accounting for spikes due to FX and Soft synths. My experience has been that if I keep the machine below 50% CPU, I don't have problems in live situations.
The simple fact of the matter is the guy who came up with that whole Electronic musician article you swear by, is NOT a programer! He doesn't really know what he's talking about, he had only good intentions, but at this point he cannot back down and admit he is Don Quixote. I'm not the only person that has done his optimizations to no good use, others have experienced the same. Reality is most people who swear it works did the optimizations BLIND, they DID NOT test before and after optimization. Those of us who did, found no improvement.
Posted: Wed Mar 16, 2005 4:25 pm
by ultrasource
I was going to post my OS X "optimization" results thinking I would have something worth posting BUT..... I spent waaaay to long "optimizing" over the weekend with very little improvement. 2% better on the cpu meter, at best.
At least I found out my m-audio firewire audiophile truly sucks compared to my powerbook's built-in audio on a clean intall.
Posted: Wed Mar 16, 2005 5:12 pm
by Machinesworking
ultrasource wrote:I was going to post my OS X "optimization" results thinking I would have something worth posting BUT..... I spent waaaay to long "optimizing" over the weekend with very little improvement. 2% better on the cpu meter, at best.
Yeah basically the only thing holding back OSX is CPU speed, and that's not really anything Apple can do something about, until IBM makes a Centrino style G5 chip.....in ten or so years!
People are thinking of OSX in terms of OS9, where Extensions could be trimmed, and you could get better performance that way.
OSX dynamically allocates resources CPU wise, so you can have multiple applications open, and the only resource being used by applications not running in the background, is the minimal RAM that the application requires. The only truly significant CPU pig I've run across system wise is the airport card, turn it off and you get a little more CPU. The safe thing to do in general is to turn off the internet before doing a show, you don't want a glitch because your airport card is receiving signal from the internet cafe next door!
Posted: Wed Mar 16, 2005 7:01 pm
by AdamJay
Machinesworking wrote: OSX dynamically allocates resources CPU wise, so you can have multiple applications open, and the only resource being used by applications not running in the background, is the minimal RAM that the application requires.
this is why Renicing Live comes in real handy.
..
Posted: Wed Mar 16, 2005 7:12 pm
by raapie
Centrino is great, but Sonoma is the big next step.
Buying new laptop based on Live 4 Performance Test
Posted: Fri Mar 18, 2005 8:23 am
by achaikin
So I'm in the market for a new laptop, high-end, for my onstage Ableton Live work.
Based on the results of the Live 4 Performance Test, I'm sticking with Windows, much as I really want to switch to Mac.

I do lots of live audio recording onstage, lots of tracks, lots of VST plugins and effects... I need my CPU% low, my RAM high, and my hard drive(s) fast.
I'm considering Dell, HP, MusicXPC, and DigitalAudioWave.
Dell doesn't make any AMD laptops... bummer. I'd really love an Athlon 3700+. But the Pentium M's are no slouch.
Both MusicXPC and DigitalAudioWave claim that their machines are optimized for music production -- quieter, all the MusicXP tweaks, dual hard drives (keeping system and audio drives separate), dual boot setups for music vs. regular use, etc.
The MusicXPC M2 (
http://www.musicxpc.com/products/m2/m2.htm) comes with two 7200rpm hard drives -- mmm. But the machine is a Pentium 4 3.2 GHz, which means 21-31% in the L4 Tests, and it's wicked expensive ($3900 MSRP).
DigitalAudioWave (
http://www.digitalaudiowave.com) has Athlons up to 3700+, Centrinos up to 2.1GHz, and lots of other zippy goodies. But the company looks small and a little sketchy.
Dell and HP are not selling their laptops optimized for music, as MusicXPC and DigitalAudioWave claim to be. But the littler companies probably can't deliver the kind of service-and-support muscle that the bigger companies can.
Your thoughts?
Posted: Fri Mar 18, 2005 8:26 am
by AdamJay
my thoughts
"optimized for music" PC's is just marketing jargon.
are there any "optimized for music" macs ? nope, does that mean you can't make music with them? nope again.
check out the HP zv5000z with AMD64 3700
there are a bunch of users on this forum with it and not a single complaint has come up... even though they aren't "optimized for music"

..
Posted: Fri Mar 18, 2005 9:02 am
by raapie
sure marketing effects are used but setting up a computer for music is different from one for playing games. for example with music you want background processes to get priority but for games you want application priority. this is a HUGE difference in performance.
12" powerbook: performance test results
Posted: Sat Mar 19, 2005 4:27 am
by taximouse
1) 49%
2) Apple Powerbook 12"
3) OS 10.3.8
4) 1.5 GHz G4
5) 512 MB
6) built-in audio
7) 5400 rpm
PB 15" 1.67
Posted: Sat Mar 19, 2005 5:08 am
by drush
1) 40%
2) Powerbook 15"
3) OS 10.3.8
4) 1.67 GHz G4
5) 1.5gb
6) built-in audio
7) 5400 rpm
Posted: Sat Mar 19, 2005 5:39 am
by toneroll
18 %
dell inspiron 9200 notebook @ 2.1 ghz pentium M
2 gig cheap RAM
onboard 7200 rpm HD
maudio firewire 410 @ 512 (will check out lowaer sample rates in a day or so)
windows xp sp1.
it jumps every now and then, but only to 19 and 2o %
i am using that speedswitchxp anyone think this might be causing the jumps?
Posted: Sat Mar 19, 2005 10:25 am
by benaline
I doubt that Speedswitch is making the difference...my test results were mostly 19, occasionally going down to 18 or up to 20, its probably the songs inherant cpu variation, I tried all the tweaks in the book to get it lower in a sad and vain attempt to beat a high score that exists only in my mind, all to no avail. In fact it was interesting that all the recommended xp tweaks made absolutley no improvement to my score compared to the computer straight out of the box. Maybe the impact of these tweaks is becoming quite minimal as pc's spec and build quality increases.
btw, You could always try closing speedswitch down from the taskbar before doing the test to be sure.....
Posted: Sat Mar 19, 2005 6:19 pm
by toneroll
yeh ill try that soon as i get back, im at my dads and seen this bumped and thought oh yeah i did the performance test the other night.
its not like i really want to "beat" a high score, i just want to know that i've got a finely tuned machine, for live. even tho it is , and is more than capable of a skip load of plugs vst's etc, its just one of those nigglin things in the back of my head if its not serviced to perfection.
(these are things i do when my ears get tired or i lose all inspiration)
Posted: Tue Mar 29, 2005 3:27 pm
by Sulas
1) 35%
2) desktop pc
3) windows 2000 prof sp4
4) athlon xp 1700+, 1,467GHz
5) 512mb ram, ddr266
6) Emu aps running kxProject drivers
7) Seagate 120Gb ata100 7200rpm
Posted: Wed Mar 30, 2005 5:42 pm
by afg
1) 20 - 21 %
2) desktop
3) Windows XP SP1
4) AMD Athlon 64 3500+ (2.2GHz / socket 939)
5) 1.5GB DDR333
6) Creamware Powerpulsar + Pulsar2
7) Western Digital Raptor 74GB / 10,000 RPM
Note: I own a GigaByte K8NSNXP-939 mainboard. Due to instability issues when the RAM is driven at DDR400, I switched to DDR333 operation. The PC is very stable now and not slow either.