question for kit xor hand drummers

[xoxos]

i'm developing a vst to generate polyrhythmic patterns.

as my plugs tend to be, the concept more or less implicates some 'most simple, most flexible' way i know to generate within a range of event. i've finished my generative algorithm and now starting the part that transforms the algorithm into events.

i'm not a studied drummer of any sort.

my current concept applies the engine to hand drum and kits with a bass event, a 'main'/'tone' event, and two nonspecific auxiliary events, which would be either two different snare hits in kit mode, or a slap and rim on a hand drum. oc application of algorithm to kit would be bass as kick drum and others as snare, and to hand drum so that bass does not occur simultaneously with other hits

if you're not bored by now.. as a drummer, do you have any technical criticism or suggestion, eg. if you would say "any decent snare drum rhythm should have four primary variant tones, these are the ways i hit the head.." i didn't find a whole lot of real depthy matter on snare drum technique

[pooshka]

not exactly on how to hit drums, but following links will tell you a lot about what is involved in snare drum sound.

http://www.soundonsound.com/sos/Mar02/a ... 58e2e44b6d

http://www.soundonsound.com/sos/Apr02/a ... 58e2e44b6d

if you go to www.soundonsound.com and search for 'synth secrets' you will get a lot of these kinda stuff. there are also synth secrets articles written for cymbals/cowbells and stuff as well.

i hope this wasn't too off topic. :p fun read nevertheless...

[Markleford]

Well if you're modelling actually playing, then I'd guess that it would be much simpler to think of it this way...

Kit: one ostinato always playing that can be one of two tones at a time (open/closed hat or ride/bell), overlaid on a more sparse accenting rhythm of one of two tones at a time (bass and snare). Those two lines can play on the same beat, but their internal tones can only be played one at a time. Oh, and occasionally a fill/roll happens where other tones can be triggered in a typically linear fashion (down the toms or whatnot), and hit a cymbal on the one to mark phrase boundaries. So far as variant tones, those are to taste: I'd pretty much do a round-robin pool of a few each for variety and call it a day, but nothing particularly "different" for each. Sure I lay on my snare rim quite often, but I wouldn't say it's integral to the rhythm: more of a random variation.

Hand: Well, it really depends on the type of hand drums, as you can do muting and palm techniques. For a conga or bongo or timbale (minus cowbell/agogo) setup, I'd say it's always non-overlapping tones (for argument's sake: sure you can strike both at once, but that's typically for accenting a main rhythmic motif in a "tutti" sort of manner) with each tone having an alternate hit for accenting (slap or rim, as you mentioned). Otherwise, most of the variation here is a matter of velocity switching in your sample set. Apart from that, I think you can get pretty good generative hand drum rhythms by running constant 16th notes in groups of 2 and 3, occasionally dropping one here or there.

Or maybe I'm full of it.

Or maybe I don't fully know what you're after.

- m

[xoxos]

ta for replies. the engine i'll have no trouble implementing in my own fashion in a music generating program, but thought i'd do a midi plug (probably sample playing alternate version too for fancy sample drillnig fills) figuring it would be widely received.

i figured the 2 alt hits method would allow fair flexibility in application for users to apply, since kit players don't do rimshots in every track et c. i'm decidedly cheesy on percussion lines, but for my 'fancier acoustic patterns' i'm usually set with 3 snare samples myself.

rolls are no problem.. prolly option to assign to two alternating, unique hits, or option to trigger the main/2 auxes.

so "only three snare sounds in a pattern" isn't a grossly offensive concept :p i know these sorts of qs are crap.. "why doesn't he just do it like he can/likes.."

[herodotus]

Just out of curiosity, what sorts of "polyrhythmic patterns" are we talking about here

In any case, I think that multiple samples of a single drum are less important to a "real drummer" sound than dynamics. And this varies from style to style. Two measures of a jazz drummers playing might involve 3 or 4 dynamic levels on each drum or cymbal. In metal and other aggressive forms of music there might be only one.

[xoxos]

my system uses two 'event-probability phrase generators,' eg. specify a loop which generates joined event probability strings, and add another one for good measure with 'probability crossfading,' which i figure should be able to produce 'perceptibly ordered but dynamic' sequences, say if you have one 16-pulse loop which crossfades via lfo to a 5-part loop, and resets at 16 measures :p

i admit i'm pretty unindustrious about being a competetive developer.. couldn't be arsed to d/l any demos of the expensive autodrummer packages, so more or less trying to ascertain whether releasing with a 3-snare-tone engine is going to be dead in water.

just to clarify, my concept with 'probability strings' is that ime random-generated rhythms are a bit clunky, cos you get "5 hard snares in a row then strong kicks on the offbeat" et c. too often, so i broke the events into strings, ~like arabic percussion, so you have say a 3-pulse string that could be like bass, rest, tone, then a 4-pulse string, and the event options on 3-pulse in that measure position should hopefully reduce the 'really silly random drumming' output. :p

[IIRs]

Q. How do you know when the drum riser is level?

A. Drool comes out of both sides of xoxos's algorithmic pattern generation plug-in's mouth.