I also did a load test about a month ago, with somewhat different results. I basically followed some testing procedures I found on the web at the old LUG site. I grabbed a 2 minute chunk of sound off a CD, put it in my audio editor and inverted it and saved the inverted copy. I then made multiple copies of both of these files (e.g., song-copy1.wav, song-copy2.wav, ..., inv-song-copy1.wav, etc.). Made a new blank song, and started adding audio tracks, 2 at a time, alternating the original and inverted copies. Each track was on a different audio playback device and used a different copy of the sound file. If everything is working right, you get silence, assuming that the track pairs have identical effect processing. I was able to get 24 mono tracks on my system (see below for details), with EQ on each, 3 bus effects and a few of inserts here and there. Inserts seemed to take up the most processing power. With EQ and bus effects enabled, but no inserts, at 24 tracks the CPU utilization was about 50% on the meter, and disk speed about 30%. As soon as I started adding inserts, CPU and disk usage started to climb tremendously. Of course, I was using the Platinum Verb as the insert :) However, even thought the meter occasionally went into the red, I did not get the dreaded CPU too slow message, and playback completed properly. I did find one small bug - once the CPU meter reaches the top red bar, it gets stuck. It will never go down again. I had to restart Logic to get the meter back. I also found out that you get different results if you use the same audio playback device (track instrument) for multiple tracks. You can get WAY more tracks this way, and if you are not using effects or the effect settings are the same for more than one track, you can save some processing power. This is useful if you are recording several takes of the same instrument and want to make a composite track, fading and cutting between different takes. I was able to get 8 or 10 tracks running on the same audio device, along with about 12 other tracks on different devices, without the CPU even getting sweaty. My system: Pentium II 400 ASUS P2B board, 100 MHz bus 128 MB RAM System drive 8.4 GB Quantum Fireball, running on the built in IDE interface Audio drive 19.5 GB Western Digital Expert (guts by IBM), which has high areal density and runs at 7200 RPM, running on a Promise 66MHz Ultra ATA controller Plextor CD-RW attached to an inexpensive Adaptec SCSI card AWE 64 Gold sound card (I know, I will get something better soon...) Clean install of Win98 SE, networking disabled, and no software such as word processors and or compilers. I use Partition Magic to keep 2 Win98 systems on the same drive, one for music and one for software development. LAW 4.0 - I haven't installed the upgrades yet because of reported issues with bounce. I am very pleased with the performance of this system, even though I am not getting quite as high a track count as others. It's more than sufficient for my needs at the moment. I mean, just think, 5 years ago 24 tracks of digital audio cost more than $100K!!!! Steve Keller Software Engineer, Online Electronic Arts, Inc. EMail: skeller@... Ph: 650/628-7056 > Joeri Vankeirsbilck <joeri@...> wrote: > > Anyway, you wanted a track count. Well, I just tried to make some audio tracks that > were playing the same audiofile. I first wanted to check the load of the CPU etc and > not the harddisk. So, I made tracks... and tracks... and tracks... until I had used > all 64 audio STEREO tracks... all playing the same file. No phasing, just the same > sound... much louder of course (had to put the master at -50 dB)... and the system > performance meter showed this: Audio: 20%, harddisk 1%... The harddisk was of course > just 1% since I simply copied the same file to all tracks... but the audio > performance surprised me: 20% with 64 stereo tracks... not bad... especially since > the system still felt as fast and fresh as it was playing 1 audiotrack. Is this test > correct? I mean: 20% is wonderful, but is it because I was only using 1 stereo > audiofile for all the 64 stereotracks? Could anyone with a different system (Mac or > PC or whatever...) confirm/compare ? > > Ciao, > Joeri > -- > Joeri Vankeirsbilck > joeri@...