Rebuilding the ADSR curve after Distortion

[ThermionicScott]

It IS quite an interesting topic.
But it is awkward this is completely not what most search in "touch sensitivity" and "tube dynamics" : gentle compression.
And the complete oppositie of sag. 

Exactly!  I was going to post a comment very similar to this.  Whether guitarists can articulate it or not, many of them enjoy how distortion limits the dynamic range.

This is not meant to dissuade you, Vivek.  I absolutely support you in building this effect, and in fact, I think you should go one further and apply a little volume expansion to the control signal!  That way, a distorted note or chord that is heavily-picked will stand out, while gentle picking will get really quiet.  A nightmare for the soundman or recording engineer, but fun for us.   

[iainpunk]

It IS quite an interesting topic.
But it is awkward this is completely not what most search in "touch sensitivity" and "tube dynamics" : gentle compression.
And the complete oppositie of sag. 

yes, it is, it feels really weird to play something like the double precision rectifier fuzz drive thing i described, it doesn't 'talk back to you' the way you expect from a normal drive. its feels like playing a dry, clean Direct in to the mixer type of feel, but with a distorted sound.
i do recommend trying it out for yourself tho, especially since the double precision rectifier fuzz drive thing is quite easy to breadboard and only needs minimal modification to get to work with a single supply.

well, its easy:

cheers, Iain

I gave you like number 300 !!!

thank you so much, its cool to have helped people 300 times, i hope i can help people atleast 700 times more

cheers, Iain

[iainpunk]

As interesting as the overall topic is, and not any sort of criticism of what has been mentioned by Jack, myself, and others in the thread, I think it bears noting that "envelope followers" in general, which rectify the audio signal, provide what is essentially a representation of the relative envelope.  They do not replicate it accurately.....

i have a vague memory of drunkenly searching on the web and coming across a all-waveform true RMS voltage detector claiming to be 'extremely accurate' with measuring the "area underneath'' the wave, which is directly and linearly corresponding with the perceived volume. it used resistor temperature matching to generate a DC voltage directly corresponding with the energy if the input., ill look around the web if i can find it,

cheers, Iain

edit: i can't find it. my internet history goes back only 31 days, and i guess its way longer ago.

[iainpunk]

i bet this even works with those hex-inverters? i am no expert in hex inverters but this isn't that weird to ask from those little chips, right?

cheers, Iain

[idy]

That Fuzzstain looks like it is giving you a blend of an "overdrive" and a compressor?

Interesting to think of letting opts feed forward and/or back in a staged amplifier/od.

The Expandora is the famous commercial implementation of this idea, and I have never been clear about it. Built one and was not excited (or expanded?) by it.
It confuses me: I look at the schematic and it seems like the LDR will make a lower resistance in the feedback loop, and so less gain, when the signal gets hot and the LED lights up, so it will compress?...but it's called Expandora.

[Mark Hammer]

That Fuzzstain looks like it is giving you a blend of an "overdrive" and a compressor?

Yes.  That is EXACTLY what it does.

Interesting to think of letting opts feed forward and/or back in a staged amplifier/od.

Yup.  I thought it was pretty clever, and still do.  Wish I had thought of it.

The Expandora is the famous commercial implementation of this idea, and I have never been clear about it. Built one and was not excited (or expanded?) by it.
It confuses me: I look at the schematic and it seems like the LDR will make a lower resistance in the feedback loop, and so less gain, when the signal gets hot and the LED lights up, so it will compress?...but it's called Expandora.

I haven't seen every schematic of it available, but I'm not convinced the ones I have seen are accurate.

[Vivek]

cheers, Iain

Are the time constants of coupling RC too big ?

Asymmetric clipping would make the cap build up a charge

If that is build up over 300 - 500 ms, it could have audible volume and harmonic changes for first part of note.

[iainpunk]

cheers, Iain

Are the time constants of coupling RC too big ?

Asymmetric clipping would make the cap build up a charge

If that is build up over 300 - 500 ms, it could have audible volume and harmonic changes for first part of note.

i have no idea about the time constants, i just threw this together for demonstration purposes. there is a big chance that the caps are way to big.

most schematics i post here are demonstration purposes schematics, not fully designed working effects pedals, especially when there are no power supplies drawn in the schematic. this is what i learned to do in school, especially when writing documentation for something, if you are describing a specific function, the power supply and biasing networks are a distraction rather than helpful.

cheers, Iain

[ElectricDruid]

+1 what others have said. I was also going to mention the Rocktave as an example of this type of circuit.

One probably could achieve an accurate model/representation of the original envelope, if pursued doggedly enough in the digital domain.  But let us remember here that the overarching objective of the OP, and many players in general, is simply to have a distorted sound which retains a lot of the dynamic feel and nuances of the unprocessed sound.

I actually think that could is probably impossible. Knowing the envelope of the signal involves knowing ahead of time whether the signal is going up or down. Imagine a ADSR-shaped sine wave. We full wave rectify, so we've got a series of bumps of one polarity. The peak height of each bump gives us the level of the envelope at that point. But say we just passed a peak...what happens then? The current bump will dive down towards zero volts and then start coming back up. Will it come up higher, in which case we need to aim upwards over the next few milliseconds? Or will the next peak not be as high, in which case we need to head downwards? We *can't* possibly know that ahead of time, so there is *necessarily* at least some lag in an envelope detector done in real time. And that lag gets worse as the frequencies get lower. Of course, if you've got a recorded sample and can process it in non-realtime, you can look ahead to your heart's content and do a beautiful "join the dots" job on all the signal peaks. Lucky for some.

But as you say...we're not after perfection. "Useful" is a good enough level to reach, and that's achievable even in real time, and even with some pretty basic analog tech.

[teemuk]

Another issue of (too perfect) envelope following is that it kills the sustain, which fights against one very important aspect why we started to employ distortion units in the first place. The envelope characteristics of distortion effect are just as important as the harmonics it creates or enforces.

[Vivek]

Please help me to understand this better

so we like Distortion also because of it sustain
therefore we do not mind loss of attack ?

or
It's possible to resculpt the envelope of a distorted tone to have initial Attack as well as sustaaaaaaaaaaaaain, and this is actually more desirable that attackless distortion, but more hard work inside the pedal.


In other words, are we limited to ENVELOPE FOLLOWING or we can also implement ENVELOPE PLASTIC SURGERY REBUILDING ?

[Steben]

Please help me to understand this better

so we like Distortion also because of it sustain
therefore we do not mind loss of attack ?

or
It's possible to resculpt the envelope of a distorted tone to have initial Attack as well as sustaaaaaaaaaaaaain, and this is actually more desirable that attackless distortion, but more hard work inside the pedal.



In other words, are we limited to ENVELOPE FOLLOWING or we can also implement ENVELOPE PLASTIC SURGERY REBUILDING ?

It is the idea of expanding into soft clipping and compressing afterwards
The DNA of a classic tube amp is one of cycle shifting and distorting by different attack while the power amp compresses.

[garcho]

Sustain is a percentage of total volume, yeah? If you have 100% sustain, you necessarily have 0ms hold and 0ms decay. That sounds like an e-bow. If sustain is less than 100%, hold and decay become audible; the less sustain, the more hold and decay are accentuated. I think part of the magic of guitar is its complicated envelope, you can hear distinct attack, decay and release, especially when played by a maestro. Too much sustain/compression and it sounds lifeless. The imperfect overdrive of the 60's/early 70's is a good compliment to the guitar amplitude envelope.

[iainpunk]

the original attack of a guitar signal has way more overtones than the sustain part, this results in crunchy inter-modulation distortion and a generally brighter tone at the attack, this gives the illusion of  an envelope while staying at the same volume.
This staying at the same volume is genuinely more pleasant to the ear, you don't have to readjust constantly. that's the reason that percussion alone is quite fatiguing for you hearing (hearing fatigue is real, ask any sound engineer) while a guitar is not as fatiguing, especially if its distorted or compressed to hell.

cheers, Iain

[Mark Hammer]

What "information" is conveyed via the timbre, what is conveyed via the amplitude and variations in amplitude, and what is conveyed via "note density"? 

As I've suggested previously, one of the purposes that compression is often used for is to unhook timbre from volume.  The player can vary how hard they slam the strings to get different tones out of them, without having to contend with the volume and unintended dynamics.

In the case of distortion, it's almost the reverse.  People want to be able to maintain a given timbre, but have the option to convey emotion via variations in dynamics and note density.  Unfortunately, the behaviour of diodes tends to eliminate, or at least reduce, dynamics.  The usual strategy is to raise the clipping threshold to permit some improved dynamic range, and enough gain to achieve clipping at the higher thresholds.  You can't get the tone you want unless you crank gain and pick hard.

That's the eternal challenge for guitarists.  We depend on certain sonic characteristics to "convey" what we want to convey, or at least support what a tune requires to optimally convey what it tries to convey.  But the goal of having ALL thse characteristics always available to us is by and large unachievable.  It's a kind of dialectic: the more you have of one thing, the more you have to forfeit of something else.

[iainpunk]

What "information" is conveyed via the timbre, what is conveyed via the amplitude and variations in amplitude, and what is conveyed via "note density"? 

As I've suggested previously, one of the purposes that compression is often used for is to unhook timbre from volume.  The player can vary how hard they slam the strings to get different tones out of them, without having to contend with the volume and unintended dynamics.

In the case of distortion, it's almost the reverse.  People want to be able to maintain a given timbre, but have the option to convey emotion via variations in dynamics and note density.  Unfortunately, the behaviour of diodes tends to eliminate, or at least reduce, dynamics.  The usual strategy is to raise the clipping threshold to permit some improved dynamic range, and enough gain to achieve clipping at the higher thresholds.  You can't get the tone you want unless you crank gain and pick hard.

That's the eternal challenge for guitarists.  We depend on certain sonic characteristics to "convey" what we want to convey, or at least support what a tune requires to optimally convey what it tries to convey.  But the goal of having ALL thse characteristics always available to us is by and large unachievable.  It's a kind of dialectic: the more you have of one thing, the more you have to forfeit of something else.

you are right, to a certain degree, but there are plenty cases where this isnt the case. look at how the Jesus and Mary chain used the superfuzz, very high gain bust still really dynamic. or death metal 'çhugging' where the playing style forces the distortion to have the maximum change in timbre over a short period of time, or some country players who secretly use the RAT for compression but they play finger style to convey a multitude of timbres through the distortion, or metalheads who use tube screamers in to marshalls that can, if done right, be really expressive despite being high gain.

my personal believes are that there are a few OD/DST/FZ pedals that respond really well to dynamics and OD/DST/FZ pedals that sound dynamically flat. its more a case-to-case basis than something you can generalize. just like compressor use, some players are just really sloppy and have irregular pick attack and use compressors to have less dynamics, but also finger style players who use it to decouple timbre from volume.

that wavefolder fuzz im working on and other DST/FZ i build in the past, i'm trying to design it to be as dynamic as possible, while still being hard and noisy and fuzzy when you dig in (the magic is the low gain [less than a TS9], but high harmonic generation/non-linearity). i think that concept is the holy grail of OD/DST/FZ and that is what i always want to work towards when designing a pedal.

cheers, Iain

[Steben]

I think you can add a certain view on it as if the total of guitar + FX + amp is the instrument, not the guitar. And every combination of devices summed up gives a different "instrument". Some achieve good interaction with an "instrument", others work best with a different one.

Just start with a Fuzz Face. Look at the sounds and waves produced through it by the known heroesand compare that to players belittling the thing as useless.

[iainpunk]

I think you can add a certain view on it as if the total of guitar + FX + amp is the instrument, not the guitar.

i have never seen it this way, but its absolutely genius!

And every combination of devices summed up gives a different "instrument". Some achieve good interaction with an "instrument", others work best with a different one.

Just start with a Fuzz Face. Look at the sounds and waves produced through it by the known heroesand compare that to players belittling the thing as useless.

if you don't think a FF is dynamic, you're probably better off playing something else than electric guitar, like acoustic, or banjo.

cheers, Iain

[garcho]

When native English speaker musicians say dynamics, they specifically mean amplitude dynamics. Distortion, by and large, squishes the hell out of your amplitude dynamics. I love a fuzz face for what it does, but what it doesn't do is "dynamics". A wall of J&MC guitars can be texturally, emotionally, psychologically dynamic, but to say there is some interesting amplitude envelope or dynamics going on seems absurd to me. I am a professional engineer, have been for over 20 years, acoustic guitars have an envelope not just harmonics,  drums are only fatiguing if they're loud, loud guitar is super fatiguing, nothing to do with the amplitude envelope whatsoever.

[iainpunk]

garcho, i apreciate your insights, but two questions:

When native English speaker musicians say dynamics, they specifically mean amplitude dynamics. Distortion, by and large, squishes the hell out of your amplitude dynamics. I love a fuzz face for what it does, but what it doesn’t do is “dynamics”. A wall of J&MC guitars can be texturally, emotionally, psychologically dynamic, but to say there is some interesting amplitude envelope or dynamics going on seems absurd to me.

i did not see volume dynamics as the full description, and thought timbre, feel and texture to be the main part of dynamics. what would you call those types of changes in timbre/feel/texture? that's what i meant with the fuzz face thing.

I am a professional engineer, have been for over 20 years, acoustic guitars have an envelope not just harmonics,  drums are only fatiguing if they’re loud, loud guitar is super fatiguing, nothing to do with the amplitude envelope whatsoever.

i believe you, but in my experience of working as light engineer in a gig venue and often forgetting my ear-pro, i found the strings sound check and full band quite doable, but the drum soundcheck really did a number on my ears. we'd soundcheck all other the instruments first and the drums last. only after sound checking the drums the temporary tinnitus kicks in. so the question i am left with is what the difference is between the two that makes my ears beep or not?? (im so glad the beep always went away after a good night's sleep)

cheers, Iain