|
New information added - Please scroll to the bottom of the page.
Click here for a schematic of the 3 stage event timer with pulse outs
Click here for a schematic of the 2 stage event timer
Click here for a picture of a the two stage Event Timer/Comparator module prototype
Click here for an audio sample
Click here for another audio sample
Click here for yet ANOTHER audio sample
and ANOTHER still!
There are functional descriptions of these samples below
The initial idea for the Event Timer came to me from a Reaktor module of the same basic function which I found to be extraordinarily useful.
Have you ever needed to time an event AFTER another one begins? One way of doing this is using a dedicated pulse delay, or another device which can be configured to do the same thing, such as the stage gate output of a sequencer or an envelope generator with an end of cycle pulse output. This works fine until you change the length of the original event. Because the delays can't track the original, each time such timing modification is made, the delays may need to be re-configured accordingly. Fine if you're working with single events - if you want to fire a slew of them off of different lengths in series however, you're screwed.
The Event Timer ends all this set-up as it WILL track its reference signal. It's basically a multi channel envelope detector. Instead of being time dependant, this device senses the amplitude of it's input against a pot variable level threshold setting. Each of the independent comparators have a manually adjustable triggering threshold. At this point, the delays you dial-up are locked to the original and will track it even when the length of the original event is modified
There is a twist to this circuit however that even the Reaktor device didn't do - each comparator has both Q and Q not outputs, can run independently from one another as separate comparators with different inputs OR they can be linked together so that the all recieve their signal from input 1 and more important, they become serialized to one another (the threshold of comparator 2 is linked to that of comparator 1, the threshold of 3 is linked to that of 2). Comparator 1 has a free range (can be set to any reference level from ground to VCC), but the threshold level of subsequent comparators down the line can't be abjusted lower than the one before it in the stack as they use each oither's pot reference for their lower adjustment limit - much like a parametric EQ or fixed filter bank. In this way, you can extrapolate two or three different voltage levels from one signal --> one taking the low(er) range, one taking the mid and one taking higher range of the same input.
Each comparator's link function is switch selectable. You can link 1 and 2 and keep number three fully independant with its own input it you care to
In application, picture the attack phase of an envelope generator (see diagram 1 below). If you put that EGs output into this module and set it in combination (serial) mode, you could get a high out when it's halfway to full voltage (mid attack) and another when it reaches it's full amplitude (full attack). In this way, you could trigger other envelopes during the cycle of the first that would track regardless of the overall length of the envelope.
AUDIO SAMPLES
Both of these samples demostrate tthe COMBINED mode.
Audio Sample 1
The first audio sample is nothing you couldn't do with the gate outs of a sequencer, but does give a clear indication of the dfevices' operation. It's three events that are timed together by the decay of the first sound. You'll hear how the track each other then the decay of the first eg is shortened. I'm using the Q NOT outputs on this patch
Audio Sample 2
This is where the Event Timer really shines. It's a simple patch that's being mated through a VCA with a long attack envelope (call this EG 1). and have the linked two other ADSRs to EG 1 using the Event Timer that control pan and timbre. The first comparator is triggering an ADSR that opens a filter, the second comparator's ADSR pans the sound. The distinct difference between this and using a sequencer or gate delay it that all three ADSRs stay high as long as EG1 does, regarless of how fast the overall event lasts (notice that all three parameters (AMPLITUDE, TIMBRE and PAN) fade together at the end, yet are cascaded at the beginning. I'm using the Q outputs on this patch.
This sets up allows an interesting pan algorythim that's synced to the overall amplitude of the event. Notice the sound fades into the LEFT Channel initailly, then quickly pans to the RIGHT and comes to rest at the CENTER for a sustained period (as long as EG 1 femains at full amplitude), then pans back to LEFT as the overall amplitude drops. Note that ths pattern stays intact even when the overall time of EG 1 is altered dramatically.
|
Diagram 1 ~ Audio Sample 2 Panning:
|
|
|
|
Audio Sample 3
Admittedly, there's nothing you couldn't do (or notice you're not doing) here by other means outside of an event timer, but i liked this so much I wanted a link to it! Here the second comparator is turning the echo on and off and the first is panning this sound - or better put, triggering a randomly VC controlled EG to pan Both are sampling the random voltage that's creating the non rythymical pulses. Love that FM sound you say? This comes from my VCO design - as yet to be posted on this site. Yeah, I love it, too!
If you build this, you should probably feed the + side of the pots from a trimmer/divider (I ended up doing this) because you'll find the usable threshold range falls well less than VCC. Without a attenuating trimmer you'll find that half your threshold pot has little to no effect and the first half is very sensitive. Also, I've added a buffering op amp between comp 1's pot wiper and the switch to keep the level of the second pot from effecting the first when they're linked. If you look at the schematic you'll see the potoential problem). It will stil work like this, but the level of the first comparator will shift when you throw it into combination mode from an offset caused by the second pot.
I'll update the schematic with both of these changes in place in the future.
This circuit is easily adaptable so more comparators can be added in the string. You might want to do this. Three linked gates would be quite interesting.
Audio Sample 4
This patch is right out of the Peter Grenader songbook as it is a gesture in the piece Secret Life which will soon be explained in somewhat great detail on the MUSIC SAMPLES section of this site. A perfect example of the work that is sometimes required for simplicity, as the patch behind this sample clearly indicates click here, as no fewer than three envelope generators, two sequencers and two VCAs are required to make it happen.
In this patch, you can see the Event Timer (again set in in combo mode) and all its patch points shown in red indicating it's doing four things: Turning the phrase on and off, telling it when to pan and when to change frequency as follows:
A) The first comparator's Q out is turning Milton on (allowing it to count). The sequencer is doing nothing but passing the metric pulse of the main sound through it's OBEDIENT CLOCK output.
B) The Q NOT output of the first comparator is turning Milton off (telling it to stop counting).
C) The Q output of comparator 2 is triggering the envelope generator that's doing the panning
D) The Q NOT output of comparator 2 is advancing a second sequencer to stage 2 which transposes the pitch a half step down.
You will notice five different lengths of phrases in this sample, the duration of which each being assigned via a S+H that's sampling noise and controlling the overall length of EGs 1 and 3. But you will also notice that even though the length of the phrase varies, the timing of the controls (pan and the 1/2 step transposition) do not, thus tracking the dynamics of the phrase itself> For this you need an Event Timer.
To get an idea when these controls are being introduced, you will see a diagram showing the envelope and the threshold set points directly above the Event Timer in the patch diagram.
|
