How many distinct distortion sounds are there really out there ?

[Vivek]

Here is a YouTube of all drive pedals simulated in the Axe-Fx II.

How many distinct sounds do you feel there are ? Which pedals are the grandfathers of each family of distinct sound ?

Distinct sound of a pedal = Not possible to achieve very similar tone by taking any another pedal and adjusting EQ and gain knobs on the pedal or knobs of EQ pedal placed before and after the pedal.

[teemuk]

I think pre and post EQ are what primarily make a distortion effect sound the way it does so...

Anyway, I think you could at least have category of 1. Fuzz effects, 2. Early distortion effects and 3. More advanced distortion effects.

Earliest Fuzz effects at least were primarily based on three circuit designs: Maestro Fuzz Tone, mk II Tone Bender, and Fuzzrite (which in turn was a clone of the "Red Box" effect Orville Rhodes had built for his studio and was not sold commercially).
First wave of evolution simply tries to improve these by modest attempts to even make these circuits operate consistently. Poor biasing methods are renewed and we start to see more negative feedback to stabilise operating points.

At the time of advent of early distortion -style effects there had been so much evolution that it's hard to distinct from where designs were derived. Now we have more focus on mid-range emphasis (the effect sounds less like fuzz) and more use of diode clipping concepts popularised by e.g. Blue Clipper, Big Muff, Jordan Bosstone, OD-1, DS-1, Distortion+ and countless transistor amps such as Kustom. These are more predictable and consistent than earlier discrete solutions in several regards. Predictable and simple opamp ICs are now also inexpensive and largely obsolete discrete amplifier circuits.

In second generation the voicing trends start to mimic complete amps even more closely and features are getting more advanced overall. Circuits feature even more complexity. We start to see effects like MT-2 or Rockman. Soon some of them start bearing title of "amp emulation" or "preamp".

Meanwhile there have been various experiments with different device types to create non-linearity: FETs (late 1960's), MOSFETs and integrated CMOS (early 1970's), OTAs (mid 70's). These keep popping up in various designs and are often advertized heavily  but never generate a gigantic success.
After late 1960's there is also great deal of research on what explains "tube sound" and lot of this research is put in action: After early cumbersome attempts to explain the characteristic with just soft clipping or asymmetric clipping we start to see more focus on "amp like" voicing, circuit architectures mimicking those of tube circuits, cascaded gain stages, bias shifting and dynamically altered clipping thresholds. Some of these ideas are utilised in various amps, some find their way to effect pedals.

[Steben]

Depends on what you call "distinct" and how sharp you draw lines. There are definitely "families". For example tube screamers, super overdrives and all derivatives.... change some minor things and they become very similar.
Most diode to ground clippers are a family too. Fuzz Faces are perhaps the most defined family. Add a transistor stage before one and you have somewhat a Tone Bender mk II. One part can make a lot of difference in all those old school circuits. Tone Bender mkI was based on the Maestro fuzzes.

[teemuk]

4th unique category of distortion effects would include those that do not create sole amplitude distortion by compression or expansion but also introduce other means of generating additional harmonics. This category would include effects like octaving or ring modulation, often integrated with generic methods to distort.

Early octavers "wavefolded" or rectified the signal, which creates emphasis on second and even order harmonic distortion but also generates a unique distortion pattern of its own in comparison to plain clipping. This is one rare effect you can't mimic with EQing.

Continuation of the idea generates a lower octave signal, often very distorted one. Fundamental frequency is extracted or estimated from input signal and then fed to frequency divider circuit. Detection is fairly unreliable so we also get strange "garble" effects.
Ring modulation in turn modulates the input signal with another signal and produces a ton of intermodulation harmonics.

5th category would include circuits that divide the signal into frequency bands and then distort one or all of the bands individually. Intermodulation products of such process will be different to distortion on the entire band and therefore this is again one category of distortion which can't be replicates with mere EQing.

[teemuk]

Replace second and third stage of Big Muff with a single opamp followed by shunt clipping diodes. Is it still "Big Muff" family or Boss DS-1 Distortion? 

[Knobby]

On the fuzzier side of things, would Schmitt trigger/comparator circuits also count as a separate type?

[Vivek]

Replace second and third stage of Big Muff with a single opamp followed by shunt clipping diodes. Is it still "Big Muff" family or Boss DS-1 Distortion? 

The Ship of Theseus Paradox



Basically you ask

Is method of distortion

Or EQ tone control

The greater signature of a pedal ?

[Steben]

Replace second and third stage of Big Muff with a single opamp followed by shunt clipping diodes. Is it still "Big Muff" family or Boss DS-1 Distortion? 

The cap in series I the bug muff makes for a distinctly treble cut distortion, resulting in bassier result. Adding that cap effect to the ds1 circuit might work. Again a detail that causes distinction.

[Vivek]

5th category would include circuits that divide the signal into frequency bands and then distort one or all of the bands individually. Intermodulation products of such process will be different to distortion on the entire band and therefore this is again one category of distortion which can't be replicates with mere EQing.

Why was this idea not mainstreamed into Guitar Amps ?


We don't seem to have Amps that divide guitar signals into bands and each band being processed differently.

[Mark Hammer]

Psychologists who study perception, in any modality, be it vision, hearing, taste, touch, smell, pain, temperature, etc., like to refer to the "JND": the just noticeable difference.  A big part of what is studied is just what it requires to produce a JND by human, and occasionally animal, observers/perceivers.  We know that JNDs follow a non-linear relationship with the intensity or robustness stimulus.  That is, the less intense a stimulus is, the less change to it is required for a JND, and as it grows more and more intense, larger changes are required to be detectable.  This is more formally stated in Weber's and Fechner's Law ( https://en.wikipedia.org/wiki/Weber%E2%80%93Fechner_law ) and eventually in Stevens' power law.

I mention this for a few reasons.  One is that it informs many of our standard practices in design.  We use anti-log taper pots for speed controls and log taper pots for volume controls because of what we are more and less sensitive to in different parts of their range.  A second reason is that we have a hard time distinguishing between small amounts of coloration in sound/tone, and a much easier time identifying particular pedals when everything is on 10....or 11.  Clearly, the design topologies of a Fuzz Face, Big Muff Pi, Tube Screamer, and Boss HM-2 are different, and we would have a relatively easy time telling them apart with gain/drive maxed.  We would have a much harder time identifying them with the gain/drive at minimum, or near minimum.  I suppose the exception might be a Tube Screamer or derivative, because the bass reduction is significant enough at any gain setting to be audibly different than those other pedals.  But trim some of the bass in any of those others, and differentiating them, based on harmonic content would be difficult.

"Distortion" is the addition of harmonic content that was not in the original signal.  Variations in the harmonic content can occur in how much is added, what harmonics are added, the relative proportions of the harmonics that are added, and (to my mind, often neglected but equally important) when they are added  and disappear.  Concomitant with distortion is also some degree of change in amplitude dynamics, in the form of compression of some degree and time course, largely because distortion is generated by running out of headroom.

Teemuk is absolutely correct in asserting that pre and post EQ - what one might simply call tone-shaping, since it can come in very simple ways such as an input capacitor value - are frequently the chief determinants of something sounding "different".  Going back to Weber, Fechner, and Stevens, when there is a LOT of harmonic content, we hear "this" pedal having detectably more (a JND) of the 3rd or 4th harmonic than that pedal (as a simplified example).  At much lower gain settings, there is insufficient added harmonic content for us to detect the difference between pedals, based on whether they add this, that, or the other harmonic, especially if such added harmonics come and go very quickly.

I've been bantering on for years about the difference between overdrive, distortion, and fuzz being partly determined by how long the added harmonics remain detectable for, post pick-attack.  We tend to call something a "fuzz" when the sizzle (the added harmonic content) lasts for a long time if one holds a note or chord.

But, to answer the question that prompts the thread, distinguish between *designs* for generating harmonic content and *audible distinctions* in harmonic content.  There may be a much smaller number of the former than of the latter. 

I think many ZVex pedals illustrate this nicely (and I don't say this as any sort of criticism).  The Super Hard On is a single transistor clean booster.  The Box of Rock is several of these circuits cascaded.  So is the Box of Metal.  Not a whole lot to distinguish them in terms of design or the means of generating harmonic content, but they sound different because of how many gain stages and tone-shaping along the way.  Many of the 1990s DOD pedals also illustrate this nicely.  They are mostly op-amps feeding a pair of clipping diodes to ground, but provide different pre and post EQ, using resonant boosts here and there.

[teemuk]

A generation of Traynor amps tried that: "Tricomp" distortion circuit divided the signal into three bands and soft clipped each of them with an overdriven differential amp. Overdriven output with sinusoid input looked more like a Maya pyramid than conventional flat-topped wave.

Nope. Didn't catch on. The problem is that we are largely accustomed to wideband IMD and anything else sounds fairly unconventional in comparison. And the more we add bands the more we actually get static intermodulation products from band division. Beyond certain degree this starts to sound very unnatural and "synth-ey".

So let's talk about mainstream... Majority of people who buy amps - and therefore return investment to amp companies so they can exist and make more amps - want amps to produce traditional tones they are accustomed with, not to sound like crazy synthesizers. A company can make an amp that distorts in a genuinely unique tone, or mimic a Marshall, and the latter results to fairly good chances of sustaining one's amp manufacturing business while first idea is likely commercial flop that people forget within few years.

Though not all new ideas fail. The famous Aphex Exciter distorts just a hi-passed side chain, in some embodiments by dynamically controlled magnitude. Then the side chain is mixed with dry signal. Result is brighter and more focused tone, not overly distorted sounding one. AFAIK, some bass amps (e.g. Trace Elliot) also integrated such effect. Popular concept favoured again by bass players, "blending" distortion of a traditionally hi-passed signal with clean signal, is actually the very same concept popularised by Aphex.

[marcelomd]

I divide dirt pedals in how I would use them:
- "Mostly used with a clean amp/power amp/return input". e.g. preamps, BSIAB;
- "Used to enhance an amp's voice". TS, Klon;
- "Takes over amp". e.g. most fuzzes.

I noticed most "Used to enhance an amp's voice" tend to be simpler circuits, with one, maybe two "voicing" stages. Usually eq shaping and lower gain, even if the final effect is high gain (TS into MB Rectifier).

[teemuk]

The cap in series I the bug muff makes for a distinctly treble cut distortion, resulting in bassier result. Adding that cap effect to the ds1 circuit might work. Again a detail that causes distinction.

I would put more weight on lower coupling capacitances of DS-1. Anyway, if you break it down conceptually you see a lot of resemblence between the two: 1. Nearly identical discrete clean input booster stage, 2. Subsequent stage(s) for gain control and symmetric clipping, 3. "Big Muff" tonestack. Yet one is more a fuzz and the other a "distortion". What sets them apart tone-wise... because it surely isn't a great difference in circuit architectures/"family"? ;-)

[Steben]

The cap in series I the bug muff makes for a distinctly treble cut distortion, resulting in bassier result. Adding that cap effect to the ds1 circuit might work. Again a detail that causes distinction.

I would put more weight on lower coupling capacitances of DS-1. Anyway, if you break it down conceptually you see a lot of resemblence between the two: 1. Nearly identical discrete clean input booster stage, 2. Subsequent stage(s) for gain control and symmetric clipping, 3. "Big Muff" tonestack. Yet one is more a fuzz and the other a "distortion". What sets them apart tone-wise... because it surely isn't a great difference in circuit architectures/"family"? ;-)

No, no great difference at all. It's about the "JND".
For non-guitarists, dirt is simply dirt or "rock tone of some kind". DS-1 and big muff is peanuts.
Even a Tone Bender mk III is close to distortion while having COMPLETELY different architecture.
Having said that, the input stage of it can be an opamp without too much difference.

Psychologists who study perception, in any modality, be it vision, hearing, taste, touch, smell, pain, temperature, etc., like to refer to the "JND": the just noticeable difference.  A big part of what is studied is just what it requires to produce a JND by human, and occasionally animal, observers/perceivers.  We know that JNDs follow a non-linear relationship with the intensity or robustness stimulus.  That is, the less intense a stimulus is, the less change to it is required for a JND, and as it grows more and more intense, larger changes are required to be detectable.  This is more formally stated in Weber's and Fechner's Law ( https://en.wikipedia.org/wiki/Weber%E2%80%93Fechner_law ) and eventually in Stevens' power law.

I mention this for a few reasons.  One is that it informs many of our standard practices in design.  We use anti-log taper pots for speed controls and log taper pots for volume controls because of what we are more and less sensitive to in different parts of their range.  A second reason is that we have a hard time distinguishing between small amounts of coloration in sound/tone, and a much easier time identifying particular pedals when everything is on 10....or 11.  Clearly, the design topologies of a Fuzz Face, Big Muff Pi, Tube Screamer, and Boss HM-2 are different, and we would have a relatively easy time telling them apart with gain/drive maxed.  We would have a much harder time identifying them with the gain/drive at minimum, or near minimum.  I suppose the exception might be a Tube Screamer or derivative, because the bass reduction is significant enough at any gain setting to be audibly different than those other pedals.  But trim some of the bass in any of those others, and differentiating them, based on harmonic content would be difficult.

"Distortion" is the addition of harmonic content that was not in the original signal.  Variations in the harmonic content can occur in how much is added, what harmonics are added, the relative proportions of the harmonics that are added, and (to my mind, often neglected but equally important) when they are added  and disappear.  Concomitant with distortion is also some degree of change in amplitude dynamics, in the form of compression of some degree and time course, largely because distortion is generated by running out of headroom.

Teemuk is absolutely correct in asserting that pre and post EQ - what one might simply call tone-shaping, since it can come in very simple ways such as an input capacitor value - are frequently the chief determinants of something sounding "different".  Going back to Weber, Fechner, and Stevens, when there is a LOT of harmonic content, we hear "this" pedal having detectably more (a JND) of the 3rd or 4th harmonic than that pedal (as a simplified example).  At much lower gain settings, there is insufficient added harmonic content for us to detect the difference between pedals, based on whether they add this, that, or the other harmonic, especially if such added harmonics come and go very quickly.

I've been bantering on for years about the difference between overdrive, distortion, and fuzz being partly determined by how long the added harmonics remain detectable for, post pick-attack.  We tend to call something a "fuzz" when the sizzle (the added harmonic content) lasts for a long time if one holds a note or chord.

But, to answer the question that prompts the thread, distinguish between *designs* for generating harmonic content and *audible distinctions* in harmonic content.  There may be a much smaller number of the former than of the latter. 

I think many ZVex pedals illustrate this nicely (and I don't say this as any sort of criticism).  The Super Hard On is a single transistor clean booster.  The Box of Rock is several of these circuits cascaded.  So is the Box of Metal.  Not a whole lot to distinguish them in terms of design or the means of generating harmonic content, but they sound different because of how many gain stages and tone-shaping along the way.  Many of the 1990s DOD pedals also illustrate this nicely.  They are mostly op-amps feeding a pair of clipping diodes to ground, but provide different pre and post EQ, using resonant boosts here and there.

Oh, and all the BOSS pedals too 

I have to say as a freak - we all are here (guitarists have some clue, non-musicians simply are bigots you see) - the attack, compression and decay of a drive tone is very much into the JND territory and beyond.
A Fuzz Face is spectacular and no circuit rates higher in the "dynamic circus/simplicity" ratio. A Metal Zone in normal use is close to zero.
A D+ or DS-1 and many of that sort simply have a sizzle decay which clearly shows the limits of simple diode clippers.

In others words, a simple "instant photo" of tones compared will have pre and post EQ as prime factors.
Dynamics and decay in other words the change in character with different amplitude will have other factor coming into play. Not even talking about misbias/sputter.

[bushidov]

I think it would be an interesting project to see a "family tree of distortion, overdrives, and fuzzes". Assuming you can get correct dates of builds, you can see how certain circuit were derived from others. There might also already be one, as I haven't honestly looked.

[soggybag]

I think of distortion and fuzz in families of:

- simple diode clippers: Electra, distortion +
- complicated clippers: big muff
- transistor feedback: fuzz Face
- transistor distortion: Tonebender, Vulcan
- FET/mosfet distortion: BSIAB, Box of Rock
- Feedback clippers: Tube Screamer
- octave fuzz: Super Fuzz

[Mark Hammer]

Again, let's differentiate between how many distinct designs there are, and how many readily differentiable sounds there are.

[soggybag]

Again, let's differentiate between how many distinct designs there are, and how many readily differentiable sounds there are.

Good point Mark, you have a way of whittling the ideas to the core!

I suppose there's clippers, wave shapers, wave folders.

[CheapPedalCollector]

I would add comparators, cmos amplifiers/inverters, and op-amp saturation to the list.

[Steben]

saturated tape deck? 
grid blocking amp, cathodyne over the top ....