WAD Karplus-Strong hand drums

The Karplus-Strong technique uses a white noise source in concert with band-pass and low-pass filtering and a very short delay time to create a chuffed attack, commonly used for plucked string and struck drum sounds.

This patch uses the TKB's A, B, C, and D row outputs to trigger the hand drum sound. Each row controls its own slope generator, each of which is configured with different rise/fall times. This results in each row's drum strikes having a slightly different character. Each row of the TKB has only a couple stages, spaced far apart, configured to send triggers. This allows other rows to send their triggers in between the triggers of any given row. The combined and uniquely configured triggers add to the illusion that a hand drum is being played.

Three of these slope generator triggers, collapsed into a single signal via a Processor, control the rapid opening and closing of the X Fader into which the white noise source is fed. The X Fader's output is band-pass filtered (using a VCFQ) before being sent to IN 1 of the Wilson Analog Delay, set with a very short delay time. The VCFQ's filter's frequency range is modified by three of the four slope's GATE outputs (similarly collapsed via another Processor), giving the hand drum a hollow sound, as if it was being struck near its rim.

The GATE output from one of the slope generators is sent directly to the IN 2 (DC) input of the WAD, which creates the low drum sound.

The TKB's ABCD output is also used, and this sends all of the triggers as they occur to the WAD's VC DELAY input, causing a bend in pitch for any drum strike that occurs when a trigger from ABCD occurs. This further contributes to the uniqueness of each drum strike, whether it's a high or low drum.

The low-pass filter on the WAD itself is used to roll off the higher frequencies, which otherwise result in a highly resonant metallic effect. This effectively damps the drum sounds, making them more realistic.

Finally, the TKB is patched to reset after 15 stages, which gives the loop an Eastern-edge rather than straight 4/4 time. At the tempo being driven by the DTG (~78 BPM), this results in a 15/8 time signature; although it sounds more like a slightly loping 6/8.

Lacking a Dual Channel Stereo Mixer on my system, the WAD's DELAY A and DELAY B outputs are sent to separate X Faders so that I can pan them downstream at my recording device. Since the delay time on the WAD is so short in this patch, there's a significant amount of phase cancellation (e.g. turning one or the other delay outputs down even slightly introduces a whole lotta bass). And the interesting phase relationship of the two outputs gives an ears-inside-out feeling that I find appealing.

In the course of working on this patch, I stumbled on a cool thing. I was actually looking for something to modulate the WAD's frequency and accidentally found that a second DTG hitting the WAD's VC F input would completely close the filter when the DTG's cycle went low -- creating a break in the loop and completely stopping all sound at the final output.

Further experimentation revealed that if I also mult the Quad Slope's various GATE outputs (currently being combined and sent to the band-pass filter's VC F input via that second Processor) to the second DTG's TRIGGER input, those triggers sneak in while the DTG's cycle is low. This creates a rhythmic glitch during those breaks akin to pulling a plug and reinserting it (noisily). Between that and the Eastern flair of the hand drumming, I'm reminded of the particular affectation for glitchy breaks in Muslimgauze's music.

Here's the full patch and video, which includes the interruptions in the loop:

No comments:

Post a Comment