[sdiy] Panel fillers

ASSI Stromeko at compuserve.de
Sat Feb 22 00:05:56 CET 2003


On Friday 21 February 2003 22:03, Tim Ressel wrote:
> This is an interesting concept. Leave us expand on it.
> Here is what I am thinking: an NCO that is run off of
> the Midi clock using a PLL to make a 128x or so clock
> to run the NCO.

A DLL would probably work more nicely as MIDI clock has quite some 
jitter and you can't filter too heavily as you need to follow tempo 
changes fast when they occur. You can then chose to either filter the 
jitter or just force the NCO to follow the clock samples it gets from 
MIDI and drop or skip phases when the estimate was not accurate enough. 
It should also listen to MIDI START commands and hard sync to a known 
phase at each START and every 96th clock tick (assuming 4/4). This 
makes an LFO that is synced to the time signature, which can be used to 
do all sorts of wierd things or even sequencing. Saving the LFO it 
might also be useful to just create triggers at and into each bar and 
use these to start envelopes. When sequencing it is probably a good 
idea to optionally honor STOP and CONTINUE messages.

> Now for the fun part: The divide-by portion get
> interesting. the midi clock is 24 ppqn. Divide by 6
> yields 16th notes, divide by 12 gives 8th notes,
> divide by 9 gives dotted 16ths. Is anyone going to
> want a divide by 10? Should this an option? Or just go
> with multiples of 6(9)?

It really isn't that much more work to divide by arbitrary integer 
numbers. How are you going to clock septuplets otherwise? Some people 
would even want to use rational and irrational divides - did you ever 
hear of Conlon Nancarrow? :-)


Achim.
-- +<[ Q+ & Matrix-12 & WAVE#46 & microQkb Omega sonic heaven ]>+ --

Q Instrument & FX Signal Flow:
http://homepages.compuserve.de/Stromeko/Qstuff/Qfx.pdf



More information about the Synth-diy mailing list