How does a chorus work exactly?

[YouAre]

I used to think chorus was just essentially just delay without Feedback/repeats and the delay time was "wiggled" to create pitch shifting. But looking at the Little Angel Chorus, it seems that it's modulating the reference voltage. So can someone clear up how chorus is achieved on a basic modular level? Looking at the Small Clone Chorus, I see a basic splitter (with filtering) splitting into clean and wet, where the wet goes through a BBD. I understand that the mn3007 BBD is a delay chip, and the LFO drives a multivibrator, but what is that doing to the mn3007? That's what I'd like to understand.



Also, is flanger just chorus, but with the wet signal fed back on itself? If that's the case, then why aren't there repeats? Because It's my understanding that feedback in delay is what causes multiple repeats...


Thanks for the help!

[geertjacobs]

For your flanger question:
Here's a GEO article on "The technology of phase shifters and flangers".
http://www.geofex.com/article_folders/phasers/phase.html

[anchovie]

I used to think chorus was just essentially just delay without Feedback/repeats and the delay time was "wiggled" to create pitch shifting. But looking at the Little Angel Chorus, it seems that it's modulating the reference voltage. So can someone clear up how chorus is achieved on a basic modular level? Looking at the Small Clone Chorus, I see a basic splitter (with filtering) splitting into clean and wet, where the wet goes through a BBD. I understand that the mn3007 BBD is a delay chip, and the LFO drives a multivibrator, but what is that doing to the mn3007? That's what I'd like to understand.

Your abstract definition of chorus is correct. With the Little Angel, Rick discovered that you could wobble the delay time by connecting an LFO to the Vref pin of the PT2399. In the Small Clone (and many other BBD-based choruses), the LFO modulates the clock driving the BBD (i.e. th delay time of the wet signal).

[merlinb]

Also, is flanger just chorus, but with the wet signal fed back on itself? If that's the case, then why aren't there repeats? Because It's my understanding that feedback in delay is what causes multiple repeats...

Chorus is indeed just a very short delay time (less than 30mS usually) that is modulated and mixed with the dry signal.
A flanger is essentially the same, but the time delay is usually shorter (less than 10mS) and the LFO is *generally* run at a slower rate (less than 1Hz). Also, a flanger may also use feedback. The delay time is too short to be heard as discrete delays; it creates deeper frequency notches and peaks instead.
A phaser is the same as a flanger, except the delay time gets smaller for higher frequencies so the spacing of the notches increases with frequency, meaning you get the same number of notches per octave, unlike a flanger where you get an increasing number of notches per octave as you go up in frequency. Feedback is also optional with a phaser.

[merlinb]

Incidentally, I recently started messing about with PureData. One thing this is handy for is to set up a precise delay line and fiddle about with delay times, modulation, feedback, filters etc, while playing a sound file through it. This really helps when trying to characterize sounds and get a feel for the real numbers that are aimed for when designing a 'real life' effect.

[YouAre]

For your flanger question:
Here's a GEO article on "The technology of phase shifters and flangers".
http://www.geofex.com/article_folders/phasers/phase.html

Much thanks! How sad is it that I've learned as much from geofex articles as I did from 5 years of engineering courses?

Your abstract definition of chorus is correct. With the Little Angel, Rick discovered that you could wobble the delay time by connecting an LFO to the Vref pin of the PT2399. In the Small Clone (and many other BBD-based choruses), the LFO modulates the clock driving the BBD (i.e. th delay time of the wet signal).

So the delay time control on certain choruses shifts the range of how much inherent delay there is, and the lfo modultes that slightly?

Chorus is indeed just a very short delay time (less than 30mS usually) that is modulated and mixed with the dry signal.
A flanger is essentially the same, but the time delay is usually shorter (less than 10mS) and the LFO is *generally* run at a slower rate (less than 1Hz). Also, a flanger may also use feedback. The delay time is too short to be heard as discrete delays; it creates deeper frequency notches and peaks instead.
A phaser is the same as a flanger, except the delay time gets smaller for higher frequencies so the spacing of the notches increases with frequency, meaning you get the same number of notches per octave, unlike a flanger where you get an increasing number of notches per octave as you go up in frequency. Feedback is also optional with a phaser.

Soooo theoretically, I can turn a small clone into a flanger with a different LFO, lowering the delay time, and adding some feedback? Just trying to get an understanding for the difference. Can you suggest a simple flanger (simple as in not toooo crazy, something as complex as a small clone or phase 90...)

Incidentally, I recently started messing about with PureData. One thing this is handy for is to set up a precise delay line and fiddle about with delay times, modulation, feedback, filters etc, while playing a sound file through it. This really helps when trying to characterize sounds and get a feel for the real numbers that are aimed for when designing a 'real life' effect.

::confused scooby doo sound::

PureData? What's that? Sounds interesting...

[Mark Hammer]

1)  When several or more people are attempting to play or sing the exact same sequence of notes in synchrony, a few things happen.  One is that there temporal synchrony is imperfect.  One player may be juuuuuuusst a little ahead of the beat and another juuuuussst a little behind.  Another is that their pitch may be aimed at being identical, but ends up not being so, with one or more juuuuusssttt a little sharp, and one or more juuuuuussstt a little flat.  This is what a chorus pedal attempts to mimic.

2) The BBD essentially samples the incoming signal, passing those analog snapshots along the series of little voltage-storage cells until reaching the output.  When the samples are played back faster than they were originally sampled, this is identical to playing a recording at faster speed, and the pitch goes up.  When the samples are played back slower than originally captured, the pitch goes downward.

3) A chorus pedal, because it uses a time delay chip that cannot reach absolute zero-delay, is always a tiny bit behind the original signal.  The LFO slows down the clocking of the BBD, and then speeds it up so as to catch up, in time, to the extent that it can (remember, it's always just a little late).  The slowing down and catching up results in the samples being played back slower and faster than they were originally obtained, yielding pitch fluctuation.  Since that is combined with a clean signal whose pitch has not changed at all, it creates the illusion of a co-musician who is sometimes flat, sometimes sharp, and always juuuuuuussssttt a little behind.

4) More sophisticated, and respected, chorus units are often referred to as "multi-phase", and may use 2 or even 3 BBDs, clocked differently, such that the sound mixed in with the clean is still always a tiny bit behind, but not exclusively flat or sharp, or even behind by the exact same amount.  You'll often find them in string synthesizers from the late 70's or early 80's, where the multiple delays were important to create the sense of more than two instruments playing at once. These are much easier to listen to because the pitch wobble doesn't stick out quite so much.

5) Yes, the primary difference between chorus and flanging effects is the amount of delay used, but there is more to it than that.  The number and location of notches produced is a function of the delay time.  The longer the delay, the more notches created, and the further down into the audible spectrum they go.  The delay range used by most chorus pedals results in notches created throughout much of the spectrum.  They move around, but I suspect because there are so many, they fail to capture our attention.  Instead, our attention is drawn to the pitch relationships, and the - at that amount of delay - audible stagger between signals.  Because there are notches all over the place, adding regeneration into the equation does nothing, since whatever emphasis to notches and peaks they might produce is so spread out, you'd never hear it as a swept effect but simply as something that just sort of pulse or blurred.

6) Flanging is created by a delay short enough that at one extreme end of sweep (shortest delay), very little of the signal has notches imposed.  As the sweep continues, the notches begin to multiply and descend into the audible spectrum.  Because they go from being only in one corner of the signal to everywhere, the location of where the notches begin captures our attention.  Regeneration focuses our attention on where the notches begin even more.  And of course, because it is a slow sweep, at a fairly short delay, we tend not to notice any pitch change at all, even though it IS being created.  Certainly, if you crank the speed, and leave the sweep width wide, pitch change/wobble will become apparent, but not at the sorts of speeds we normally prefer to run when flanging.

7) So, yes, flanging and chorus are produced via the exact same "engine" (LFO swept BBD), but because of the properties of what happens within the audible spectrum by virtue of the delay produced, we hear different attributes.

[merlinb]

So the delay time control on certain choruses shifts the range of how much inherent delay there is, and the lfo modultes that slightly?

Exactly.

Soooo theoretically, I can turn a small clone into a flanger with a different LFO, lowering the delay time, and adding some feedback?

Yes I suppose so (feedback is optional).

PureData? What's that? Sounds interesting...

It's an open source program for doing music synthesis, mostly. I find it handy as an aid to circuit design, since you can build up audio "circuits" in it and actually hear what happens to sound. It's handy for getting a feel for where one effect becomes another as you make things more extreme. For example, how much delay modulation do you need before a chorus turns into a spacey pitch shifter? Or what freq's and resonance make a really good wah sound? Once you figure these things out with PD, you can design a real-life circuit that meets the conditions you set, and you get your perfect wah, or whatever  
http://www.youtube.com/watch?v=NzYzxY1WaQU

[Earthscum]

6) Flanging is created by a delay short enough that at one extreme end of sweep (shortest delay), very little of the signal has notches imposed.  As the sweep continues, the notches begin to multiply and descend into the audible spectrum.  Because they go from being only in one corner of the signal to everywhere, the location of where the notches begin captures our attention.  Regeneration focuses our attention on where the notches begin even more.  And of course, because it is a slow sweep, at a fairly short delay, we tend not to notice any pitch change at all, even though it IS being created.  Certainly, if you crank the speed, and leave the sweep width wide, pitch change/wobble will become apparent, but not at the sorts of speeds we normally prefer to run when flanging.

So, to make an ideal Flanger, you should run 2 BBD's and 2 clocks, one set to shortest delay, and one set to be swept from shortest to (depth/rate), and mix the stable delay as "dry" and swept delay as "wet".    I wonder if you could do some kind of flanging with a pair of PT's like that. The only thing would be the inherently long delay from pick to sound.

[frequencycentral]

So, to make an ideal Flanger, you should run 2 BBD's and 2 clocks, one set to shortest delay, and one set to be swept from shortest to (depth/rate), and mix the stable delay as "dry" and swept delay as "wet". 

TZF

I wonder if you could do some kind of flanging with a pair of PT's like that. The only thing would be the inherently long delay from pick to sound.

Currently experimenting with just such a device. 

[YouAre]

Hmmm multi voice Little Angel? 3 pt2399's with different delay times all in parallel? Yes please!!!

[YouAre]

Mark, you cleared up every question I had, Thank you! I was basically confused as to how Chorus pedals basically resemble delay pedals, but without feed back (example: echo base vs Little Angel). So the introduction of feedback (which i now know is not necessary for flangers...) would leave me to believe that there should be repeats. But now I realize that delay time is too short on flangers for the repeats to do anything but smear together into swooshy goodness, right?

So in general, what are the ballpark delay times for chorus and flange? I know most delay pedals are like minimum 30ms, so that's a good reference.

[Mark Hammer]

The most dramatic flanging will sweep over a range of delay times that is probably at least 30:1.  The Boss BF-2 sweeps from 1ms to 13ms; not stellar.  The A/DA sweeps from 0.33ms to 14ms, at 42:1, much better.  The PAiA Hyperflange sweeps from 0.17ms to 15msec, which is an 88:1 sweep.  Here are some interesting flanger construction articles to read.

http://hammer.ampage.org/files/Anderflange1.PDF
http://hammer.ampage.org/files/PAiA_Phlanger.PDF
http://hammer.ampage.org/files/Hyperflange.PDF

Chorusing doesn't require the broad sweep range/ratio since the goal is not to have the notches commence way u in the spectrum and then go all the way down to the bottom.  Simply going sharp and flat is enough.  So, for example, the DOD FX-60 sweeps from 5-8ms.  Chorus pedals are generally found sweeping in the vicinity of 3-4ms out to maybe 20msec.

Doubling and double-tracking start to kick in when the delay is out mast 15ms or so.  Taking a delay line set for a relatively short delay (usually a bit shorter than what might serve as slapback echo) and modulating it ever so slightly, like 20-24ms, will achieve a rather ethereal kind of sound that is somewhat different.  Think Pat Metheny rather than Andy Summers.

[YouAre]

The most dramatic flanging will sweep over a range of delay times that is probably at least 30:1.  The Boss BF-2 sweeps from 1ms to 13ms; not stellar.  The A/DA sweeps from 0.33ms to 14ms, at 42:1, much better.  The PAiA Hyperflange sweeps from 0.17ms to 15msec, which is an 88:1 sweep.  Here are some interesting flanger construction articles to read.

http://hammer.ampage.org/files/Anderflange1.PDF
http://hammer.ampage.org/files/PAiA_Phlanger.PDF
http://hammer.ampage.org/files/Hyperflange.PDF

Chorusing doesn't require the broad sweep range/ratio since the goal is not to have the notches commence way u in the spectrum and then go all the way down to the bottom.  Simply going sharp and flat is enough.  So, for example, the DOD FX-60 sweeps from 5-8ms.  Chorus pedals are generally found sweeping in the vicinity of 3-4ms out to maybe 20msec.

Doubling and double-tracking start to kick in when the delay is out mast 15ms or so.  Taking a delay line set for a relatively short delay (usually a bit shorter than what might serve as slapback echo) and modulating it ever so slightly, like 20-24ms, will achieve a rather ethereal kind of sound that is somewhat different.  Think Pat Metheny rather than Andy Summers.

Ah, so flangers have a much wider range of modulation, at a much lower delay time. Like chorus "wiggles" between +/-2-3ms centered somewhere around 20-25 ms, whereas flanger gets more of a +/-7ms range centers at 8ms. That seem right? Sorry if it's just repeating what you just said, I just want to make sure my understanding is solid.


And I'm guessing a cool way of "flangifying" the small clone is to replace the "delay time" resistor with an LDR that's modulated by a nice and slow LFO. What do you think about that?

I'm guessing a more practical way would be to set the delay time fairly low, and lower the speed/increase the amplitude of the original LFO? (Increasing the amplitude would effectively increase the modulation from stock 3-4ms to 5-7ms, eh?)

Thanks!

[merlinb]

Based on my own experiments, I propose the following as a rough guide. It doesn't show LFO speed or feedback of course, which blur the subjective boundaries rather.

[YouAre]

Based on my own experiments, I propose the following as a rough guide. It doesn't show LFO speed or feedback of course, which blur the subjective boundaries rather.

Whaaaaaat!! Too cool! Thank you!

[Mark Hammer]

Based on my own experiments, I propose the following as a rough guide. It doesn't show LFO speed or feedback of course, which blur the subjective boundaries rather.

Nice chart.  I think there is more overlap between the major categories than is shown, but the gist is certainly sitting in the ballpark (inside the baselines, rather than in the outfield or out in the parking lot).
Just note that normal human perceptual properties apply, such that the greater the rate of sweep, the narrower the sweep should be.  A change of 10msec delay, accomplished over 5 seconds, sounds smooth.  A similar change within less than a second is perceived as disruptive and jarring.  This is, of course, one of the reasons why the wider delay ratio (i.e., 30:1, 40:1, etc.) matters more in flanging than it does in chorusing; because the sweep rate is generally much slower.

[puretube]

Disagreed...

for me, Chorus is more about pitch than delay (though a little time variance won`t hurt...) 

[YouAre]

Disagreed...

for me, Chorus is more about pitch than delay (though a little time variance won`t hurt...) 

Well yeah, that's the desired affect. But how do you achieve that apparent pitch bending with analog components?

[mr clack]

Ive got a synth book by Barry klien called "musical electronic circuits" there is a good section in there that describes the differences in a more mechanical sense, I wouln't call it definitive but maybe gives an idea about how different elements of the effects are created. I imagine most circuits exploit more than one of these properties thus the blurring of what a chorus / Flanger is. I have described them.

Vibrato = Low frequency sine oscillator modifying the delay speed no feedback or wet/dry summing
frequency shift = Saw / Ramp modifying the delay speed no feedback or wet/dry summing
Flanging = Feedback from the output to input (regeneration) no wet/dry summing
Chorus = Sine wave moifying pitch with wet/dry summing
Echo = Both feedback and wet/dry summing big delay