geddonn

Hi guys, I'm trying to hack together a simple 2 step Envelope Generator (AD) from the 7-step envelope generator example patch. I'm trying to make it react exactly like the ADSR~ object with Sustain and Release set to 0.

Unfortunately I can't seem to make the envelope retrigger without first sending a zero value.

Also I can't seem to make it retrigger until the envelope has finished.

Is there a way to do this simply? Possibly more simply than the 7-stage envelope generator method, I feel it's maybe overkill for a 2 step envelope.

Hi guys, I'm trying to hack together a simple 2 step Envelope Generator (AD) from the 7-step envelope generator example patch. I'm trying to make it react exactly like the ADSR~ object with Sustain and Release set to 0. 

Unfortunately I can't seem to make the envelope retrigger without first sending a zero value.
Also I can't seem to make it retrigger until the envelope has finished.

linear-ad-envelope-generator

you can use e.g. function with line~  or curve~...

without gen~:
you can use e.g. function with  line~  or curve~...


Thanks for your reply. I know several ways I can do it in MSP and this is a good one, but the point was to achieve it in gen so I can make a sample accurate envelope generator.

Aha, thanks, I guess I can use a pulse train to get it out of the max and into MSP too.

Now I guess my only problem is how to get the envelope to reset to zero quickly if it is retriggered while it's still ramping.

Aha, thanks, I guess I can use a pulse train to get it out of the max and into MSP too.
Now I guess my only problem is how to get the envelope to reset to zero quickly if it is retriggered while it's still ramping.
Any ideas?
I'm gonna have a rummage.


Here's a way of doing it. This is not as flexible as the seven stage envelope in terms of number of stages, but (I think) it has the same retriggering behavior as adsr~. I've also added the legato mode, which I find more useful for monosynths.

Also, you can prematurely short-circuit the attack (i.e. if you release the envelope before it's reached the decay phase, it will decay from there)

I highly recommend checking out some of the threads on envelopes on kvraudio. I've found them greatly helpful.

FWIW, envelopes can be surprisingly hard to write, especially with features such as those found in adsr~. I'm working on an analog style ADSR now that has a retriggering model similar to the Moog mini. (where rapid retriggering of notes would cause an increase in volume)

For no-frill two-stage envelopes you can look at models that are much simpler: a pulse-wave generator (I'd look at sig~ 1 into +=~ into

Here's a way of doing it.  This is not as flexible as the seven stage envelope in terms of number of stages, but (I think) it has the same retriggering behavior as adsr~.  I've also added the legato mode, which I find more useful for monosynths.

I highly recommend checking out some of the threads on envelopes on kvraudio.  I've found them greatly helpful.

FWIW, envelopes can be surprisingly hard to write, especially with features such as those found in adsr~.  I'm working on an analog style ADSR now that has a retriggering model similar to the Moog mini.  (where rapid retriggering of notes would cause an increase in volume)  

For no-frill two-stage envelopes you can look at models that are much simpler: a pulse-wave generator (I'd look at sig~ 1 into +=~ into 

Is it possible to change the attack/decay time WHILE the ramp is in progress?

Is it possible to change the attack/decay time WHILE the ramp is in progress?
Help appreciated!


Sure! I'd encourage you to check out the Bp.AHD and Bp.AD2 in BEAP for some examples. (Especially the latter one)

Sure!  I'd encourage you to check out the Bp.AHD and Bp.AD2 in BEAP for some examples.  (Especially the latter one)


Thanks a lot, that helped clear a lot of doubts. Boy was I wrong about making Envelope Generators :3 :3 :3 :3 :3

Here is a gen~ implementation of an ADSR envelope, first published in May 2014, with full documentation. It uses the same envelope-phase calculation method as the 7-stage envelope, and is openly expansible to any number of stages

http://www.yofiel.com/software/cycling-74-patches/adsr-envelope

Ernest, have you considered adding a Delay stage (for a DADSR)? Might be handy for LFOs, and a fun one to figure out.

Ernest, have you considered adding a Delay stage (for a DADSR)?  Might be handy for LFOs, and a fun one to figure out.


It really isn't very difficult to add a delay stage in gen~. One simply puts a delay object on the gate-on input, with time set in sample periods.

It really isn't very difficult to add a delay stage in gen~.  One simply puts a delay object on the gate-on input, with time set in sample periods.


I think it's slightly different, since the attack of the envelope is delayed but the release is not. Also, the wait period might interact with the release in a monophonic setting. For controlling the amplitude of a pitch LFO, a hard reset to 0 would probably be fine, but this might not be ideal for an amplitude LFO's envelope. My guess is that you could probably just model it as an additional stage in the envelope which has 0 amplitude.

I think it's slightly different, since the attack of the envelope is delayed but the release is not.  Also, the wait period might interact with the release in a monophonic setting.  For controlling the amplitude of a  pitch LFO, a hard reset to 0 would probably be fine, but this might not be ideal for an amplitude LFO's envelope.  My guess is that you could probably just model it as an additional stage in the envelope which has 0 amplitude.


My design does both of those things already. Delaying the gate-on signal would not affect the gate-off signal, and it already has a pre-attack phase that only happens on gate-off events while the envelope is on, to prevent clicks.

My design does both of those things already.  Delaying the gate-on signal would not affect the gate-off signal, and it already has a pre-attack phase that only happens on gate-off events while the envelope is on,  to prevent clicks.


if i may ask: why is it a problem that you are required to send 0 before the list?

anyhow, if you dont need to modulate the envelope in realtime, you could use a buffer~ with the pregenerated envelope inside.

@Ernest, I'm doing analog-style envelopes triggered at signal rate so gate-on and gate-off are usually in the same stream and then a separate stream containing triggers is used to indicate new envelopes. I took a look at it and the delay triggering at signal rate ended up being pretty easy to add to existing envelopes; the key is not using a delay to begin with.

For anyone curious, here's an example of a signal rate triggered DADSR. It uses an asymmetrical rampsmooth~ (with arguments of 88200 1) in combination with a threshold so that it takes extra time to trigger the start (2 seconds), but the release is triggered immediately. This also allows it to be cancelled without being triggered.

@Ernest, I'm doing analog-style envelopes triggered at signal rate so gate-on and gate-off are usually in the same stream and then a separate stream containing triggers is used to indicate new envelopes. I took a look at it and the delay triggering at signal rate ended up being pretty easy to add to existing envelopes; the key is not using a delay to begin with.  

For anyone curious, here's an example of a signal rate triggered DADSR.  It uses an asymmetrical rampsmooth~ (with arguments of 88200 1) in combination with a threshold so that it takes extra time to trigger the start (2 seconds), but the release is triggered immediately.  This also allows it to be cancelled without being triggered.

Hello, Peter I checked your analog adsr post really great work and Ernest I had a look at your adsr which is also excellent.

I recently made a simple AD envelope off the back of the 7-segment example in the gen help library. However I'm not satisfied with the resulting curves produced by scaling the output, with scale.

* I noticed a similar discrepancy between outputs in the 'like curve' example in the gen library.

Ernest I remember that you mentioned creating an envelope with curve values that are consistent over varying lengths, I wonder if you may be able to help me?

Here is a copy of the patch there is a comment and two presets illustrating the issue.

* a side note, when I look at the live.adsrui it seems as though the exponent is split across two sections of one ramp?

Ernest I remember that you mentioned creating an envelope with curve values that are consistent over varying lengths, I wonder if you may be able to help me?

Here is a copy of the patch there is a comment and two presets illustrating the issue.

* a side note, when I look at the live.adsrui it seems as though the exponent is split across two sections of one ramp?



This is roughly what I will go with, I did a few test's with easing functions but then I can't switch between linear and curved envelopes.

I will add logic to switch between curves for each segment, if anyone notices anything about this code they might change then let me know.

Hello, thank you for all your amazing posts.

I made something very simple, it is an AD envelope with individual curve control for attack and decay.

There is no smoothing on the parameters, but I figure you can implement that in your own way, really looking forward to diving deeper into envelope design but this is a start, any criticism is appreciated!

I made something very simple, it is an AD envelope with individual curve control for attack and decay.
There is no smoothing on the parameters, but I figure you can implement that in your own way, really looking forward to diving deeper into envelope design but this is a start, any criticism is appreciated!

I've noticed a problem, when a new envelope is triggered before the previous one is finished it jumps to the new point rather than ramping to it.

Here is a simple version of an AD envelope I have made, it does not yet have a re-trig mode like Ernest's, and is not as sophisticated as the other's here created by Peter Mcculloch. I look forward to diving into their designs later.

I'm working on the re-trig, but here is the current version, you can set the curve for attack and decay independently, one thing I noticed about the 7 segment example is that the shape of each curve is a bit odd, so I used pow and that seemed to give a more evenly distributed curve than exponent in scale.

Linear AD Envelope Generator