MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Portamento (glide) over fixed time, regardless of interval?
MUFF WIGGLER Forum Index -> Eurorack Modules Goto page 1, 2, 3  Next [all]
Author Portamento (glide) over fixed time, regardless of interval?
Mr D
Hi folks,

For your consideration:

Imagine you want to have portamento (glide) on a mono patch (using a slew limiter?).

When you press C3, then D3, it should take 150ms for the voltage representing C3 to transition to the voltage representing D3.

Next, you press C1, then C4. It should also take 150ms to slide over those 3 octaves.

Most hard and soft synths will have a glide mode that behaves like this, as this is the most musical portamento mode.

I just bought the Doepfer A-171-2, and of course it doesn't work like this.

So, i'm wondering, is there a Eurorack module (slew limiter?) that will transition between changed voltages at a fixed rate, regardless of the size of the voltage step?

Somehow i doubt it, but was just wondering!
uebl
MN Maths will of course do it (although you can't really dial the timing in milliseconds..). What do you want to use it for, TB-303 sequencing?
huffnPuff
What you are describing is fixed time, not fixed rate - you want the slew to travel larger gaps faster so that it can reach it’s destination in a fixed time span.
Mr D
Yes, you're right, time, not rate............

Hmmm, i guess a Maths purchase is not too far down the line.............
huffnPuff
I’m not sure if Maths will do this, but it’s doable in software for sure.
I think the way to do it is with an exponential curve where the rate of change depends on the distance that remains - like charging a battery or a capacitor.

I’m sure there’s a way to do it in analog, I don’t know exactly how.
huffnPuff
OK, let’s think about this: you want the slew rate to change progressively so this means a voltage control slew limiter.
Feed the slew rate CV input with the target voltage minus current voltage and fine-tune to taste.
If the Doepfer module has voltage control over the rate you could try that.

Edit: if a higher voltage at the slew rate CV input makes voltage travel slower (it probably does) then you want to invert the formula above: current voltage minus target voltage plus some offset.

This way, like a train reaching its destination, the closer you reach your destination the lighter the deceleration is.
huffnPuff
ingnore.
Mungo
huffnPuff wrote:
the target voltage minus current voltage
This is the hard part, you'd use a sample and hold or analog shift register to memorise the voltages so you could have a steady difference CV.

Mr D wrote:
Imagine you want to have portamento (glide) on a mono patch (using a slew limiter?).

When you press C3, then D3, it should take 150ms for the voltage representing C3 to transition to the voltage representing D3.

Next, you press C1, then C4. It should also take 150ms to slide over those 3 octaves.

Most hard and soft synths will have a glide mode that behaves like this, as this is the most musical portamento mode.
Here is the trick, just use a filter:

You can compare the sound of a filter and a slew rate limiter in that video (panned left and right). I prefer the sound of a filter.
huffnPuff
Mungo wrote:
huffnPuff wrote:
the target voltage minus current voltage
This is the hard part, you'd use a sample and hold or analog shift register to memorise the voltages so you could have a steady difference CV.


I don’t see why? The target voltage is available from the sequencer and the current voltage (akward English, sorry) is the slew limiter’s output. Actual voltage might be a better term.

Using a filter probably makes more sense though.
Mr D
Thanks...........learning something here.

Ok, so use a filter.

I guess the filter slope and frequency will determine the glide time?

Will a filter at one fixed setting give me the same glide time from C3 to D4 as from C0 to C7?

EDIT: is this what people call a "low pass gate"?
Mr D
Anyone care to explain the physics behind this, i can't quite get my head around it?

So normally the pitch CV switches instantly. A filter won't allow that instant step, so turns it into a sawtooth?

And the lower the filter, the lower the frequency of that sawtooth? (actually just one cycle).

And the amplitude of the sawtooth determines the frequency jump?

Am i on the right track? ............. or.......?
diode_destroyer
The Kenton midi 2 cv stuff has fixed time modes. From Kenton Modular Solo manual: "Fixed time will attempt to keep the time taken for the slide to be constant, regardless of the interval between the start and finish notes. (In extreme cases this is not always possible.)"
danishchairs
Mungo wrote:
Mr D wrote:
Imagine you want to have portamento (glide) on a mono patch (using a slew limiter?).

When you press C3, then D3, it should take 150ms for the voltage representing C3 to transition to the voltage representing D3.

Next, you press C1, then C4. It should also take 150ms to slide over those 3 octaves.

Most hard and soft synths will have a glide mode that behaves like this, as this is the most musical portamento mode.

Here is the trick, just use a filter:

You can compare the sound of a filter and a slew rate limiter in that video (panned left and right). I prefer the sound of a filter.

Good question, great explanation and demo video.

I kinda like the sound of the two of them (glide using a filter and using a slew limiter) together.
Dcramer
Mungo is right.
I was working on a project recently where I needed portemento over ver tiny steps and using Maths wouldn’t work as it was way to slow over the large steps but you could still hear the stepping in the tiny steps.
I tried my Livewire Bessel Generator, and viola! It’s based on a filter circuit and perfectly tracks time so that the distance between steps doesn’t change the slide time.
Sadly it’s not accurate for 1v/oct but for my CV mangling it was perfect! w00t
pieter
Mr D wrote:
Anyone care to explain the physics behind this, i can't quite get my head around it?

So normally the pitch CV switches instantly. A filter won't allow that instant step, so turns it into a sawtooth?

And the lower the filter, the lower the frequency of that sawtooth? (actually just one cycle).

And the amplitude of the sawtooth determines the frequency jump?

Am i on the right track? ............. or.......?


Forget sawtooths (sawteeth?). The low pass filter will round the edge of a square wave. If that wave determines the pitch, you will now get a glide. The lower you set the cutoff frequency of the filter, the longer it takes to get from the starting voltage to the final voltage. The important part is that the cutoff frequency has an effect on the slope regardless of the voltage difference. That's why any jump now glides the same amount of time.

You can add resonance to the filter to get interesting pitch wobbles. thumbs up
wiperactive
I've recently been asking myself the question that OP Mr D put forth having realised my current slew limiters all behave in the constant rate mode while what I really want is the fixed time option often found in non-modular synths (which I generally prefer, especially within the realm of tight sequencing).

It appears - on the face of it - that the WMD VC Mini Slew should be able to do this, though like the Kenton converter, there may be limitations to the consistency of its "time compensation" mode.
diode_destroyer
Mr. D, are you sure the Doepfer doesn't do what you want? I would have thought that if you set both the switches to exponential mode and turn the upper shape knob all the way CCW and the lower shape knob all the way CW, then it should behave like a first order circuit (which has this time constant effect you want). But I don't have one so I dunno...
huffnPuff
The patch I described above, assuming that module has voltage control over slew, should work. It definitely works with Maths. Chosing this curve or that curve is not an instant fix; the slew rate should be set in reverse proportion to the actual (i.e. progressively diminishing) distance to the desired pitch.
This will work with pitch CV, except in some Doepfer modules you need a buffer between the slew limiter and the oscillator (no such issues with Maths).
desolationjones
I puzzled over this last year and the answer is simple: Exponential slew will be very close to fixed time, regardless of interval.
Mr D
About the Doepfer: to be honest i only got it yesterday. I tried it for about an hour but it just sounded so wrong and unmusical, regardless of settings, that i gave up on it. But i'm certainly gonna have another go.

I'm still a little puzzled. I don't understand why the slew rate is described as linear and the filter as exponential.

I agree that an exponential slew rate will seem a bit closer to what i want, as it goes much quicker to the final CV. But what i actually want is a linear curve that moves over a fixed time. And after that's been achieved, i can worry about whether the curve is exponential or not.

What makes me also suspect that an exponential slew isn't the answer is the fact that the best VSTis (Serum, Avenger etc.) actually let you choose between a linear or exponential (or anywhere in between) portamento curve. And the difference is very easy to hear at all intervals.

Just some random thoughts and i'll need to experiment some more with this.

Also found this thread with an audio clip of portamento which sounds pretty good: https://www.muffwiggler.com/forum/viewtopic.php?t=15494
huffnPuff
Most chances the Doepfer sounds off because its output is unbuffered. If you have a buffer patch it between the slew limiter and the oscillator.

An exponential curve can be patched with a feedback loop, which is how a filter is implemented and what the patch described above does.
Mr D
Hmmmm, i've tried to run my pitch CV into my AJH 2412 filter, and it seems to be blocking DC, so it won't pass on the 1V /octave CV to my oscillator.

Anyone know of a cheap and cheerful filter that is good for this purpose and won't block the DC?
pieter
Mr D wrote:
Hmmmm, i've tried to run my pitch CV into my AJH 2412 filter, and it seems to be blocking DC, so it won't pass on the 1V /octave CV to my oscillator.

Anyone know of a cheap and cheerful filter that is good for this purpose and won't block the DC?


Yeah, I just ran into exactly the same problem. None of my filters is DC coupled. The trick of using an exponential slew mentioned above is probably your best bet. Maths to the rescue!
diode_destroyer
Well, the exponential setting is 'historically' accurate given that in all the old synth schematics the portamento/glide is just a first order (RC) low pass filter buffered by surrounding opamps. The reason this is "constant time" is that the solution to a first order circuit in response to stepping voltages (CV) is exponential, with an RC time constant. The RC constant depends on the potentiometer time setting (R) and a fixed capacitor value (C). I've never played with Serum but if it has linear sloped constant time glides then I should check it out.
Mr D
I think i'm getting confused here.

Obviously the CV has to increase exponentially as you go up volts / octave to get proper pitch tracking. But it's my Intellijel uMidi that's taking care of that.

So the exponential setting of the Doepfer is going to determine how it transitions from the initial CV to the new CV value, regardless of whether that interval is C3 to C#3, or from C0 to C7? So that's not going to give me that constant time that i'm looking for, right?

Yes, i may be talking nonsense as i'm not very technical.

I just tried the Doepfer again with a buffer and it definitely isn't working in the musical way i need. So big intervals take longer, short ones much less.
MUFF WIGGLER Forum Index -> Eurorack Modules Goto page 1, 2, 3  Next [all]
Page 1 of 3
Powered by phpBB © phpBB Group