Monitoring Latency & Disk Load
Posted: Mon Feb 13, 2006 6:19 pm
Got a weird issue here I was wondering if anyone else could confirm there's either a) a problem or b) that I've missed something obvious and hence a complete idiot
On projects where I'm doing serious tracking and production, ie typically in excess of 40 channels of audio (drums, vox, synths and multitracked instrumentation), I'm noticing that as the project gets larger, regardless of CPU load, I'm getting increasingly bad monitoring latency response from my VSTi's. Even with my ASIO buffers down as low as the project can smoothly run at, typically 256, I'm getting something like 100ms latency.
Now at first I thought this was merely a CPU issue - but the CPU isn't even above 50% on most projects once stuff's been frozen. I even tried loading up a fresh live set, with no audio tracks, and just added as many different VSTi with monitoring 'in' as I could until just before the CPU crapped out - but even then the monitoring latency, for a 256 buffer was correct and totally playable. Hence I'm beginning to think that heavy disk load is actually the culprit.
Can anyone help? Anyone working on projects of the same size that can test this?
Sys: WinXP SP1a, P4 2.6C, 512Mb DDR400, 120Gb SATA and 200Gb Firewire. Audiophile 24/96 PCI. Live 5.03.
On projects where I'm doing serious tracking and production, ie typically in excess of 40 channels of audio (drums, vox, synths and multitracked instrumentation), I'm noticing that as the project gets larger, regardless of CPU load, I'm getting increasingly bad monitoring latency response from my VSTi's. Even with my ASIO buffers down as low as the project can smoothly run at, typically 256, I'm getting something like 100ms latency.
Now at first I thought this was merely a CPU issue - but the CPU isn't even above 50% on most projects once stuff's been frozen. I even tried loading up a fresh live set, with no audio tracks, and just added as many different VSTi with monitoring 'in' as I could until just before the CPU crapped out - but even then the monitoring latency, for a 256 buffer was correct and totally playable. Hence I'm beginning to think that heavy disk load is actually the culprit.
Can anyone help? Anyone working on projects of the same size that can test this?
Sys: WinXP SP1a, P4 2.6C, 512Mb DDR400, 120Gb SATA and 200Gb Firewire. Audiophile 24/96 PCI. Live 5.03.