|
| quote: | Originally posted by MERiDiAN5i2
First off, I avoid VSTs that are CPU hogs. Some plugins are just total CPU whores, and unless they do something really special, I avoid them like hungry sharks. If I must get sounds out of them, I play the sounds I want to a WAV and load it up in the sampler.
Playing sampled sounds is usually very low CPU, so samplers are your friend. If an instance of an instrument is only being used for several sounds, render single hits of the instrument into WAVs and load them up in your sampler. Play the samples instead of the instrument. Samplers are your friend, and so is a ton of good sample packs!
Freezing tracks. I forget how this works in fruityloops, but working in Nuendo/Cubase it's really easy... Just click the snowflake in the VST panel, wait a little bit, and the instrument is rendered to WAV and the VST closes, but the track stays in place, just not changable; basically zero CPU for that track. If you want to make a change, unfreeze, make changes, edit, whatever, refreeze. I love how Nuendo/Cubase freezes stuff so easily. I always freeze stuff I'm not planning on working on in the next hour or so, which is quite often half or more of the tracks.
Use effect plugins sparingly. Effects can always be added at final mixdown from rendered WAVs, and I generally unload them while I'm working on the track, bring them back into the mix if I need to hear them. Besides, this practice helps you to get things to sound good without lots of EQs, filters, and other effects. The better your work sounds without tons of effects, the better you can tune it at the final mixdown 
Be careful of how many voices are in use. Some instruments totally hog voices depending on settings, such as with lots of sustain or long release, you may find an instrument is playing alot more voices than it needs to be. Some VSTs allow you limit the number of voices in use, and this can greatly reduce CPU utilization often without much notable different in output, especially if the instrument in question is "down in the mix". Nexus is a good example of a voice hog, especially with pads, although I really like Nexus.
Why do you need 10 of the same instrument open? Can't you use automation to bring this down to a lesser number, such as automating settings and program changes? Or is it that you are trying to make the whole track from one instrument?? It's quite possible to make a very full sounding track from only 8 instruments, if one of them is a sampler and you automate program changes and dont bring in a whole seperate instrument for 2 bars of sound...
Also, your deficient sound hardware probably has alot to do with this. That "HD" audio is both a CPU and interrupt hog. Uberjunk. Get yourself a good HARDWARE (not software-based crap like the HD audio) sound interface that supports ASIO. My old (~2004) 3GHZ Pentium 4 with a decent ASIO interface can do lots more than my recent-model Core Duo 2 laptop, mainly because the laptop is using the pathetic "HD" audio. Seriously though, a good audio interface is a must to deal with production work.
Most importantly, make sure your PC is fully optimized for being a digital audio workstation. This basically entails a bunch of system settings and ensuring that nothing unnecessary is running on the PC. Search around here, I know there's a number of threads on optimizing your PC for digital audio work, this a bit broad of a topic to cover. |
My God you sure as hell hit the spot. I never really looked into the fact that the number of voices will kill CPU. Gonna check that out. I'm usually very generous with the effects but for things that require long releases, I export to wav.
Good information to know. Thank you.
___________________
Life isn't about finding yourself. Life is about creating yourself
|