sample length vs. # of vel. layers
I’ve just invested in Linplug’s RMIII. The advantages over the RMF (that I wanted) are: 1. multiple velocity layers and 2. Multiple outputs.
The problem I’m immediately faced with is the RAM usage (eg. 16 crash cymbal samples @ 2Mb per sample) as all the kit’s samples are loaded into RAM. You can imagine how this eats up the RAM with even a fairly small kit.
Obviously there is a trade-off possible here in that I can:
1. Reduce the number of velocity layers per drum (but get less realistic hits) and,
2. Trim the tails off the samples. eg. the crash cymbal lasts over 10 seconds. I’ve listened to that crash tail at high volume and it sounds awesome but perhaps in a typical musical context it’s not noticable.
What are peoples thoughts/experiences?
Render to wav files and then record on top of those. you can always go back as redo the drums, re-render them. That way you can use all of those layers.
edit: oh, and add more RAM.
Thanks for that. I guess I could do render it but one of the main reasons I like working with midi drums is the ability to make changes quickly.
So it’s really back to my original question - Would it produce a more musical sound by shortening the drum hit wave files (and it’s mainly the cymbals) or using fewer velocity layers per pad?
Shorter decays would be my preference, especially in a denser mix where other instruments will mask the fact that the samples don’t decay like in real life. Velocity layers are so important for realism in developing the groove to sound like a real person is at the kit. Unless you are going for 80’s phoney pop drums, velocity layers are your friends. Just make sure when you chop the decay you don’t go too nuts. You can probably shorten a cymbal to 2-3 seconds and have it be quite acceptable. But it will sound phoney if it is so short that it just sounds like a short ‘psht’.