Bug? Delay LFO clicks bug [NOT A BUG]

Bodde

Bodde

Fractal Fanatic
Was trying to create some sounds of the Lexicon PCM42. I discovered a possible bug or at least some strange behavior. When using the square LFO in the Digital Mono delay you get all kinds of crazy annoying clicks and artifacts. I think this is not what it is supposed to do. Same for many other LFO types like random and trapezoid for example. Same happens in the analog delay but a bit less noticeable. Same for LFO2. Here is the preset. I kept it simple and exaggerated the settings for demonstration purposes. You can leave out the drive block and it is still the same.

You can also simply test it by changing the LFO type to square and fiddle with the depth settings. Most noticeable with low rate and high depth settings.

Why do these clicks happen? and can this be fixed? It is not usable this way.
(Axe Fx II standard, FW17.02)
 

Attachments

  • Delay LFO bug.syx
    6.3 KB · Views: 2
Last edited by a moderator:
If you re trying to modulate delay time, you will get artifacts from the digital types. Use the tape delay and modulate motor speed instead
 
ok thanks. Yes, I understand that a square wave on delay time can give some artifacts but I didn't expect it to be that much and loud. It is unusable to me this way and I am curious how people use the square wave or trapezoid on the digital and analog delay then without being bothered by clicks. Maybe it is only usable for faster rates? Can't remember that the Lexicon PCM42 or the PSP software had those clicks when using the square wave. But I am not sure anymore now.
 
FractalAudio said:
Any waveform with a discontinuity will cause this. Think about it.
Click to expand...

While I'm guessing that your imperative is intended to make people think intuitively about a speaker "jumping" in response to a discontinuity, I would be remiss if I didn't say something about the mathematics for those who might enjoy the nerdery and have the relevant background.

The Fourier transform of a periodic function (think: periodic waveform like a sine wave, square wave, etc.) measures the prevalence (amplitude) of the various harmonics (pure sine waves of various frequencies) that constitute the waveform. (It is a not-very-obvious fact that any reasonable such waveform may be decomposed into an infinite sum of such harmonics.)

Thus, for example, the Fourier transform of a pure since wave just tells you the amplitude of that wave, and gives you zero for all other harmonics. If you do this for a square wave, for example, you get nonzero constituents of arbitrarily high frequencies. That said, the amplitudes of these frequencies decrease as the reciprocal of the frequency. This phenomenon (the tapering off of high frequency amplitudes) is a general phenomenon that goes under the heading of the "Riemann-Lebesgue Lemma."

These high frequencies are responsible for the clicks you hear, and will be a feature of any waveform with a discontinuity, as Cliff notes. What's more, they will also be a feature of any waveform with a point of non-differentiability (think: corner) like a trapezoid wave, even though it is perfectly continuous.
 
DrNick said:
While I'm guessing that your imperative is intended to make people think intuitively about a speaker "jumping" in response to a discontinuity, I would be remiss if I didn't say something about the mathematics for those who might enjoy the nerdery and have the relevant background.

The Fourier transform of a periodic function (think: periodic waveform like a sine wave, square wave, etc.) measures the prevalence (amplitude) of the various harmonics (pure sine waves of various frequencies) that constitute the waveform. (It is a not-very-obvious fact that any reasonable such waveform may be decomposed into an infinite sum of such harmonics.)

Thus, for example, the Fourier transform of a pure since wave just tells you the amplitude of that wave, and gives you zero for all other harmonics. If you do this for a square wave, for example, you get nonzero constituents of arbitrarily high frequencies. That said, the amplitudes of these frequencies decrease as the reciprocal of the frequency. This phenomenon (the tapering off of high frequency amplitudes) is a general phenomenon that goes under the heading of the "Riemann-Lebesgue Lemma."

These high frequencies are responsible for the clicks you hear, and will be a feature of any waveform with a discontinuity, as Cliff notes. What's more, they will also be a feature of any waveform with a point of non-differentiability (think: corner) like a trapezoid wave, even though it is perfectly continuous.
Click to expand...

No, what I meant was that any LFO waveform that is discontinuous will cause a click or glitch if you attach that LFO to the delay time. It's analogous to instantaneously moving the playback head to a new position on a tape deck. If you move the head slowly the pitch will change. If you move it quickly the pitch will change even more. If you move it instantaneously you will create a discontinuity in the audio playback.
 
Gotcha. The mathematical explanation still applies to explain the click given this source of signal discontinuity, though, so I'll leave it for anyone who cares to muse on it.

Your level of involvement with your customers is really admirable - I don't know how you find the time for all this!
 
Good discussion guys! Did someone try my preset? Curious what you think. To me these these clicks sound unnatural and too distorted and random somehow (like something is broken). I was expecting some artifacts but not like this. Maybe I do something wrong. Just curious how you can use these waveforms (other than sine and triangle) in a nice way on the digital and analog delay.
 
I've never had cause to use a square wave (or the others you mention) in this way, but I have to say that your patch pretty much sounds like what I would expect it to given the nature of this form.

What sort of effect were you hoping/looking for?
 
DrNick said:
I've never had cause to use a square wave (or the others you mention) in this way, but I have to say that your patch pretty much sounds like what I would expect it to given the nature of this form.

What sort of effect were you hoping/looking for?
Click to expand...

Do you get the random distorted clicks too? It was not looking for a specific sound just experimenting with the possibilities/boundaries of the modulation/LFO of the delays for some weirder soundscape sounds. Was trying to get the sort of functionality that the Lexicon PCM42 has with it's delay LFO. Was not trying to copy that exact sound but just hoping the Axe can be used like the PCM42 with the LFO's .
 
i tried your patch bodde, but it wouldn't load properly

here's an example of how i would use a square wave lfo tied to motor speed in the tape delay block. you need to use the tape delay, because it's the only delay block that will allow for glitch free changes in delay time
 

Attachments

  • Square Wave Delay.syx
    6.3 KB · Views: 2
simeon said:
i tried your patch bodde, but it wouldn't load properly

here's an example of how i would use a square wave lfo tied to motor speed in the tape delay block. you need to use the tape delay, because it's the only delay block that will allow for glitch free changes in delay time
Click to expand...


Maybe it has to do with different FW (I'm on 17.04) or do others have the same problem? Will try out your patch later but I think I can not load it because you are on a later FW. Thanks for the preset!
 
Bodde said:
Do you get the random distorted clicks too? It was not looking for a specific sound just experimenting with the possibilities/boundaries of the modulation/LFO of the delays for some weirder soundscape sounds. Was trying to get the sort of functionality that the Lexicon PCM42 has with it's delay LFO. Was not trying to copy that exact sound but just hoping the Axe can be used like the PCM42 with the LFO's .
Click to expand...

I don't get random clicks - the clicks are timed in accordance with the LFO1 rate. Try increasing this rate from .1Hz to something like 1-2Hz and the regularity of the clicks will become more apparent.
 
simeon said:
no, i'm on 17.04

what you're experiencing is not a bug, or a problem. it's normal
Click to expand...

I meant the problem of not being able to load my preset...:)

Edit: I said I'm on 17.04 but I'm on 17.02. Unfortunately can't load any of your presets (also not from other topics like the Echorec). Must be a FW mismatch.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom