Here is a quick add of snares with Metal Foundry, with Avatar pieces. I don't have a screencapture, so I will just walk you through manually.
On the "Construct" screen
Right click the main snare, and look at the "Voices and Layers". I have the "gradient" (middle hit layers) set to 20, voice limit to 99. This will be the across the board on any snare that is used for middle velocity layers (blasts in this case)
The kicks are the same, except for the avatar kick (1st pad on the right). I have random turned off, and "hard" set to "5", which since Random is off, is only triggering the "hard" kick sample labeled as 5 in the engine
On the "Mapping" screen
Click on "Midi Nodes" on the right, then right click "MidiNode2"
Below that is the "Members" screen. All of these will be triggered on D(3? D1 on the SD2 keyboardI don't remember how the math goes)
I have the "main" snare set for center hit 0-127, so it will be triggered across all ranges, I have the "edge" set from 0-100, so it will only be hit during blasts or lower velocity rolls/ghost notes, etc, and the "rimshot" set for 100-127. The same idea carrier to snare 2 (X3) which is the blast supplement snare. Snare 3 (x2) is the hard hit snare, which I always use the GMS Hasch Maple sample because of the ring. It will only be hit during 100-127 velocities, so hard or very heavy handed doubles (I like to leave a little bit of room for dynamics if the fill isn't supposed to simulate a double-stroke roll).
This was put together without any EQ or mixing, and the snare sounds will come together more when you mess with compression, specifically the Hasch Maple. I always set this for stupid fast compression to get rid of it's transient, but leave the body and ring in, so it will "tuck" underneath the main sample and the transitions between layer cross over will be less evident.
This is pretty much how I setup something specific to Death/Black metal, since there won't be much variation between blast and the standard hit. If you are working with something where the snare is being much more used, and you need better control over the dynamics than setting velocity layers will do, then I would suggest setting all drum samples 0-127, printing the tracks to .wav, and controlling the volume of each snare via volume automation. I generally only leave rimshots at higher velocity, since they can sound gross at lower volumes (almost pending the snare out of tune, especially with multiple layers) and most drummers won't rimshot if they are doing intricate work (ghost notes, or whatever).
