waveshaping.
something i work with a lot for synths but for guitar it seems like there is very little information on making it work.
the first thing i am wondering if anyone has done any DIY guitar waveshapers in the last few years. the only one i have seen mentioned is TIM's triple fuzz and
that was mentioned in an OLD thread.
i have heard people say you could take a guitar signal and compress it first so that it would work,etc.
the reason i am asking this is this seems to be an untapped area of guitar fx.
any advice, articles, links,etc appreciated.
i would really like to get some new sounds going!!!
Quote from: loss1234 on August 12, 2009, 12:35:28 PM
waveshaping.
something i work with a lot for synths but for guitar it seems like there is very little information on making it work.
That's largely because synth waveshapers/folders are in essence, analog computation devices. It is easy to say what a synth waveform looks like - after all, you generated it. And synth waveforms tend to be either 5v or 10V in size.
On the other hand, it's sometimes difficult to say what a guitar waveform is. It can be quite small, 20-50mV for some older pickups, up to a couple of volts for modern humbuckers. And it already varies hugely when you get it to start processing on it - the ADSR is already part of the signal, as is the varying harmonic content of the string as it decays. This tends to put waveshaping with a guitar signal firmly in the range of computing with an unknown. It can be done, but it's much more difficult.
On the other hand...
Quotethe first thing i am wondering if anyone has done any DIY guitar waveshapers in the last few years. the only one i have seen mentioned is TIM's triple fuzz and that was mentioned in an OLD thread.
DIY guitar waveshapers are the single most used devices - they're called "distortion" and "overdrive". Admittedly, this is much duller, being primarily limiting and clipping, than complex wavefolders, but waveshaping is what's going on.
Quotei have heard people say you could take a guitar signal and compress it first so that it would work,etc.
Yep. That's an attempt to get around the low and varying amplitudes of the guitar signal. Compressing gets you a more consistent signal to work with, and then amplifying gets it up in to a voltage range where the thresholds of the typical wavefolder/shaper can work on it. This raises the issue of somehow re-impressing dynamics back onto the signal after you're done shaping it. The only pedals I'm aware of that have addressed this is Craig Anderton's "Roctave Divider" and the E&MM Harmony Generator, both of which take the original envelope dynamics and re-impress a highly-compressed output with them. Guitarists in general tend to want some compression, anyway.
Quotethe reason i am asking this is this seems to be an untapped area of guitar fx.
It is - but that's because it's both technically and user-interface-ed-ly difficulty. It's not nearly so immediately gratifying as making Yet Another Tube Screamer Clone (YATSC) and going into business, raising a buzz by announcing a six-month waiting list. :icon_lol:
The only wavefolder-style effect I'm aware of is the Hyper-Fuzz project, which I think Mark Hammer has on his page, hammer.ampage.com. This uses a three-threshold folder set up to match guitar levels inside the effect.
I've messed with wavefolders on guitar, but they were always fussy enough to set up and enough of a special case sound effect that I never developed anything for public consumption.
I've played about with bits of Ken Stone's Wave Multiplier http://www.cgs.synth.net/modules/cgs29_wave_multiplier.html (http://www.cgs.synth.net/modules/cgs29_wave_multiplier.html). Like R.G points out though these aren't designed for guitar signals and probably don't work properly with them.
The grinder section is interesting with guitar, gives a strange ringing distortion. I think I changed a few component values, to make it wilder, can't remember exactly what now though.
The bottom bit of the multiplier with the diodes and opamps makes a crazy octave up fuzz.
The Lockhart folder is cool as well, that's basically what the triple fuzz is.
The Simple Squarewave Shaper at Circuit Snippets (currently offline) is an adaptation of project/circuit that originally appeared in POLYPHONY. It is essentially a comparator that produces a square wave, which is then followed by a variable lag circuit for rising and falling edges. This lets you produce ramp, sawtooth, and triangle waveforms as well as the original square.
I have a waveshaper module that was part of the Korg PME40 series from the early 80s. It was a fairly complex circuit for what was essentially a fuzz. http://reviews.harmony-central.com/reviews/Effects/product/Korg/PME-40X/10/2
Boss makes a "wave processor" dual pedal (WP-20G) which appears to be clearance bin material in some stores. http://www.rolandus.com/products/productdetails.php?ProductId=166 It is actually a synth that requires a GK-series divided pickup to work.
I adapted Juergan Haible's Wave Multiplier B for stompbox use a few years ago, using a max 1044 for the bipolar supply, and an opamp boost and soft clipping stage before it to make it work with a guitar. It's a really unique distortion pedal.
http://www.jhaible.de/jh_wavefolder.html (http://www.jhaible.de/jh_wavefolder.html)
Some bit crushers that do bit swapping like the frostwave sonic alienator are doing the same thing as wave folding. Also the plastic and unusual Soundblox Multiwave distortion pedal that came out recently does wave folding.
So, I signed up on this forum a while ago and promptly forgot about it. Then, recently, I built a guitar adapted version of one of the wave folders on the CGS website. I was wondering if anyone else had done anything similar and this thread turned up on Google.
Here is a Youtube clip of the pedal I built:
http://www.youtube.com/watch?v=l4LTEaIYmU0
The wave folder section is preceded by a jfet mu amp, with a cap that can be switched across its output pot to cut some high end. I put a 10k linear pot between the triangle and square wave sections and ran that into a 10k audio pot for output level.
The person you really want to speak to is Anthonie "Ton" Barmentloo. You cansee many of his experiments here: http://www.youtube.com/user/puretubetechno
The, um, ones that do not have an oscilloscope in them generally have to do with beer, and not electronics.
im confused. I thought Crowther audio prunes and custard was a wavefolder style distortion. I thought tim escobedos Bronx cheer is a wavefolder. I thought synthmongers fuzzmonger is a wavefolder distortion??????
and you have to treat these circuits lie you treat Octavia or octave up by rolling back tone and possibly even volume a slight bit then play above the 9th fret. that's were the magic is with wavefolder type circuits with guitar like prunes and custard
Quote from: chemosis on January 17, 2019, 08:57:30 PM
im confused. I thought Crowther audio prunes and custard was a wavefolder style distortion. I thought tim escobedos Bronx cheer is a wavefolder. I thought synthmongers fuzzmonger is a wavefolder distortion??????
Yes, they are. A synth uses a VCA in front to vary folding "order". A guitar player achieves it with playing sensitivity.
The signal changes direction not only at highest and lowest level, but at different breakpoints above and below zero...
Prunes and custard is one of the best pedals I have ever heard and played, it is like a different kind of guitar synth, no, really, so much tonal variation. Get one! (@everyone)
Or build one :-X
Bronx cheer and Wave Tripler (?) are also nice, not so variable though, but easy to build in an hour. Bronx cheer is interesting in that it also introduces frequency dependance.
I don´t know synthmongers...
The tone god has also switch blade iirc. Sounded good on breadboard!
Quote from: noisette on January 18, 2019, 05:12:11 AMThe tone god has also switch blade iirc. Sounded good on breadboard!
Tone God Blade, or Rothwell
Switchblade ? :icon_biggrin:
:D
I meant the first...
Forget to say Prunes and Custard is especially killer on bass guitar.
Play two adjacent notes that beat against each other (like f# and g), try the upper registers, compare against deep registers, sounds of bizarre beauty---
/enthusiasm off
I built the Prunes and Custard, and i am indeed enjoying it ;D
Quote from: noisette on January 18, 2019, 05:12:11 AM
Quote from: chemosis on January 17, 2019, 08:57:30 PM
im confused. I thought Crowther audio prunes and custard was a wavefolder style distortion. I thought tim escobedos Bronx cheer is a wavefolder. I thought synthmongers fuzzmonger is a wavefolder distortion??????
Yes, they are. A synth uses a VCA in front to vary folding "order". A guitar player achieves it with playing sensitivity.
The signal changes direction not only at highest and lowest level, but at different breakpoints above and below zero...
Prunes and custard is one of the best pedals I have ever heard and played, it is like a different kind of guitar synth, no, really, so much tonal variation. Get one! (@everyone)
Or build one :-X
Bronx cheer and Wave Tripler (?) are also nice, not so variable though, but easy to build in an hour. Bronx cheer is interesting in that it also introduces frequency dependance.
I don´t know synthmongers...
The tone god has also switch blade iirc. Sounded good on breadboard!
i have had many prunes and custards and love them esp when treating them like a Octavia.such a warm smooth tone at certain settings. i like Bronx cheer alot to. if the maleeko chaos isn't a wavefolder circuit it definetly sounds like one. if you haven't heard one checkout malekkos official video of the chaos because the downer just dosent sound as good to my ears. malekko chaos! sounds different than P&C but still amazing
This is an old thread and idk why we're talking about prunes but...
I've been working on a thing in a digital format that extracts the RMS value of the input signal and uses that to set the clipping threshold after the gain stage. With certain shortish window times it works really well to keep the signal distorted by about the same amount no matter how loud or soft you play while mostly retaining the overall dynamics of the input. It's kind of unnatural but also kind of really cool, and something a lot of people have been hunting for a while. With somewhat longer window times, it does a kind of interesting subtle blooming kind of thing. At times up around a second, it turns into a special effect, but a fairly interesting one.
Somebody smarter than me would have to figure out how to do that in analog, though honestly I don't think it has to be that tough.
Quote from: ashcat_lt on March 08, 2019, 01:07:04 PM
This is an old thread and idk why we're talking about prunes but...
I've been working on a thing in a digital format that extracts the RMS value of the input signal and uses that to set the clipping threshold after the gain stage. With certain shortish window times it works really well to keep the signal distorted by about the same amount no matter how loud or soft you play while mostly retaining the overall dynamics of the input. It's kind of unnatural but also kind of really cool, and something a lot of people have been hunting for a while. With somewhat longer window times, it does a kind of interesting subtle blooming kind of thing. At times up around a second, it turns into a special effect, but a fairly interesting one.
Somebody smarter than me would have to figure out how to do that in analog, though honestly I don't think it has to be that tough.
Run clipping diodes to the output of an envelope follower instead of to ground (you'd need to run each diode to a different voltage, and place them before the coupling cap from the opamp).
OR
NE570/571. Compress, distort, expand.
The latter would probably be easier (or at-least less parts), the former would probably be closer to what you're doing digitally.
I've done compress distort expand and it didn't really work out as well. That might have been my fault. Diode biasing was what I kind of thought, and IS pretty much what I'm doing in code.
I came across this vid today, and it seems simple enough to breadboard in order to see if it can be adapted to suit guitar level input:
(https://i.postimg.cc/t12NRMV8/Screenshot-20220421-183441.png) (https://postimg.cc/t12NRMV8)
I'm going to add an Input cap and pulldown resistor, output cap and volume pot, change U1A from a buffer to a big boost, replacing the original divider gain pot with a pot in the f/b loop and probably run it off +/-12v using a 7660S charge pump.
Now for the questions: From the explanation in the vid, I get the idea that the wavefolding is caused by the clipping action of the antiparallel pairs to ground on each of the individual signal paths combined with the (crossover distortion?) effect of the antiparallel pairs that are in series with those five signal paths: If we assume the vf of the diodes is 700mV or so, I have the idea that the guitar level would need to be boosted to more than 5 x 700mV for all of the paths to have an effect.
A) Does this seem correct? B) Is a single opamp stage run on a 12v dual supply likely able to deliver this?
Quote from: nocentelli on April 21, 2022, 02:55:30 PM....I have the idea that the guitar level would need to be boosted to more than 5 x 700mV for all of the paths to have an effect.
A) Does this seem correct? B) Is a single opamp stage run on a 12v dual supply likely able to deliver this?
B) Do math. 5*700mV is 3.5V. Any 12V opamp does that easy.
A) This is not for guitar. Seems to be from Synth-World, where some systems swing +/-10V. You may need another opamp to boost. They are cheap.
you can get away with +/-7.5v rails when using 1.5v dropout op-amps (TL072, etc ...)
onset of full-swing range is about 10vPP (seen at the input side of the 100r resistor)
that's a voltage gain of x100 when starting from 100mV single-coils ...
the other challenge here is current levels drawn from input source buffer/amplifier
sims show +/-10mA at onset of full swing range and getting worse from there
I can see a design alternative yielding about 1/3 the required input for onset of full-swing range
that's at least more manageable on the gain side of the equation (x33 gain relative to 100mV input)
(https://i.postimg.cc/yg0TJhXr/BUCHLAwavebend-JCM2022.png) (https://postimg.cc/yg0TJhXr)
Waveshapers and folders strike me as being really more directed at waveforms/signals that are much more stable and consistent than what a guitar string provides, the idea being that manipulation of a wave, in real time, would add some animation to an otherwise "stagnant" signal. As well, insomuch as a great many waveshapers rely on diode arrangements to divide incoming signals into components that can be manipulated in some way, it is assumed that the signal has to be above some critical level, and not fall below that. Unfortunately, guitar strings decay quickly. One can maintain a constant high level by means of compression and overdrive, but there you've already altered the harmonic content of the signal.
I have one of the Korg PME 40X "Waveshaper" modules, and it's not really much more than a fancy fuzz. I mean, it does provide different fuzz characters, but they are not dynamically controllable, and not the sort of thing that elicits a "Wow, where has this been all my life?" reaction.
Quote from: Eb7+9 on April 21, 2022, 10:12:34 PM
you can get away with +/-7.5v rails when using 1.5v dropout op-amps (TL072, etc ...).
onset of full-swing range is about 10vPP (seen at the input side of the 100r resistor)
that's a voltage gain of x100 when starting from 100mV single-coils...
Thanks! I read somewhere that synth signals from the standard Eurorack +/-12V systems tend to be between 5-10v; I will be using humbuckers (plus other signal boosting stuff - see below) so a single opamp might work ok if it can dish out 7.5v.
Quote from: Eb7+9 on April 21, 2022, 10:12:34 PM
the other challenge here is current levels drawn from input source buffer/amplifier
sims show +/-10mA at onset of full swing range and getting worse from there
I can see a design alternative yielding about 1/3 the required input for onset of full-swing range
that's at least more manageable on the gain side of the equation (x33 gain relative to 100mV input)
OK, now I'm a bit lost at this part...
Quote from: Mark Hammer on April 22, 2022, 07:24:25 AM
Waveshapers and folders strike me as being really more directed at waveforms/signals that are much more stable and consistent than what a guitar string provides, the idea being that manipulation of a wave, in real time, would add some animation to an otherwise "stagnant" signal. As well, insomuch as a great many waveshapers rely on diode arrangements to divide incoming signals into components that can be manipulated in some way, it is assumed that the signal has to be above some critical level, and not fall below that. Unfortunately, guitar strings decay quickly. One can maintain a constant high level by means of compression and overdrive, but there you've already altered the harmonic content of the signal.
I have one of the Korg PME 40X "Waveshaper" modules, and it's not really much more than a fancy fuzz. I mean, it does provide different fuzz characters, but they are not dynamically controllable, and not the sort of thing that elicits a "Wow, where has this been all my life?" reaction.
I will be using a guitar as the "signal generator", but my chief interest is trying out a waveshaper it to add some different flavours to the EHX SuperEgo synth pedal I have: This is an upgrade of the old EHX Freeze pedal and can generate an infinitely sustaining drone from whatever note or chord is played into it. It has level controls for the synth and dry outputs, with TONNES of level available (I usually have to keep the wet level at around 3 o'clock for unity with the bypassed signal), so being able to boost it to 7.5v p-p might not be absolutely essential for this purpose..... However, I might want to also use it with direct guitar for some occasional graunchy fuzz nastiness, and I might also like to use it in the FX loop of the Superego so I can have clean guitar alongside waveshaped guitar synth drone, and I'm guessing the synth level control is post-return rather than pre-send so it probably does need to be able to deal with guitar output level signals.
Quote from: nocentelli on April 22, 2022, 08:14:19 PM
I might want to also use it with direct guitar for some occasional graunchy fuzz nastiness ...
well, uh ... yeah
what I posted above are the parameters needed to get full wavebending action
you might need up to 10vpp of signal to get three full "folds" on each signal lobe ... dun't matter if it's from your guitar or voice
you need to bring the level up to that amount of signal swing if you want what you see produced in the video happen with your source
for a signal with 100mV amplitude that translates into a gain of at least x100
that's LOTS of gain ... (you can/should try)
the other important constraint is how much current the input diode string is going to draw at THOSE signal levels
you need an op-amp that won't sneeze at 10mA output draw ... chances are many op-amps will be ok with this
and still produce an un-distorted wave shape driving the Buchla wave-bender diode network input
but I'm just pointing that out now in case someone runs into problems and wonders why
now, on the flip side, we know this pre-clipping might not necessarily be a bad thing either
but I like to separate the mechanisms in case I want to have control over each later ...
the first idea is to try a version of the circuit where there is no clipping anywhere
(ie., by using a high-gain/high-drive input amplifier)
afterwards, you can try introducing some dynamic curtailment and see what that sounds like
there's quite a bit to experiment with here
as a first try you could use an un-beefed (straight) op-amp driver ...
http://www.lynx.net/~jc/Buchla100hz200mVpp.png (http://www.lynx.net/~jc/Buchla100hz200mVpp.png)
for the front-end amplifier in the schematic I'm posting you'll want replace the 10k fixed resistor (R18) by a 10k variable one,
providing a DRIVE control
and you'll want to stick a volume pot at the output of the last op-amp,
providing a VOLUME control
holy s#it, why haven't i seen this thread before, i have been playing with a wavefolder fuzz on and off for about 5 years now.
(https://i.postimg.cc/RWyGvgG1/Screenshot-20180911-233451-2.png) (https://postimages.org/)
i found that using single-fold wave folder bits gives more flexibility and a deeper, more accurate fold. the down side is that you need to stack multiple single fold bits in series to get it to do multiples, but the amount of folds can be up to 1+2n where n is the number of stages, if designed correctly. a gain of 3 in between stages gives the most accurate folding in my experience
i find that the best sounds are had with single stage wavefolding. also, don't compress it before hand, the playing dynamics have huge effect on the sound, almost like an envelope filter sometimes, which is almost the only thing it has going for it in the conventional sense.
depending on the input amplitude and pre-gain, they can sometimes react extremely poorly to chords, but the amount of pre gain can make the clean signal poke out above the folded waves, making it a bit more like crossover distortion, in this operating mode, chords do work, but youll get a JaMC like fizz in the back ground
i currently have an experimental 2 stage folder on my breadboard, which i am in the process of enhancing for bass use, by making it even more ''''bouncy'''' shaping the EQ both up front and after the folds
ill be happy to answer more questions, or share design concepts and ideas.
cheers
i currently have an experimental 2 stage folder on my breadboard, which i am in the process of enhancing for bass use, by making it even more ''''bouncy'''' shaping the EQ both up front and after the folds
ill be happy to answer more questions, or share design concepts and ideas.
I d really like to see your bass implementation, as a bassplayer myself always keen on playing with these mutated effects. Thanks a lot
currently its just a breadboard setup, and i havent physically drawn the circuit yet.
its a simple gain of 16x opamp, followed by two wave folder stages, the first one has red leds and 4148 diodes in series for a higher forward voltage, the second stage has single 4148's. this is then followed by a bridge-T filter which has a wide, shallow scoop around the 560Hz point which basically accentuates the bass's natural low end and the ''bouncing overtones'' while not filling up to much of the mix like bass over drives often do. a cool thing about wave folding is that it retains a part of the input instrument's character. the last thing is an output buffer and volume control.
if i were doing this for a guitar, i'd put in more gain upfront, and add traditional clipping after the wave folding stages to delete some of the guitar's normal character in favour of normal distortion characteristics mixed in with the wave folding character. ow, and i wouldn't scoop the mids, maybe ill even boost the mids.
these aren't what i'd normally do for bass distortions, but it just fits the 'bouncy' nature of wave folding. its a bit of a special case. most of the overtones a wave folder generates are non-harmonic distortion, which sounds, well, different, synth-like, nasty, spunky....
hope this helps a bit
cheers