Comparison analog <-> digital distortion

[carrejans]

For a school project (a subtask of a thesis), I have to write a popularizing paper. (so not too scientific)
The subject is the comparison of analog and recent digital distortion fx for guitar.

What would be interesting to compare is probably the clipping graphs and frequency responses for showing the harmonics.
Do you guys know any articles or links of comparisons of commercial digital distortion pedals?

What are all the problems of the recent commercial digital distortion fx?
- more intermodulation distortion (is this correct?)
- lower quality because of sampling and quantisation (is this still a problem?)
-

And what is your opinion about the recent dig distortion fx?


Thanks

[Toney]

 No matter how many times I have played digital drive and distortion effects, they eventually end up giving me the sh#ts.

I 'went' Boss GT5 for a while... about a month. Couldn't freaking wait to hand it back. I spent all my time with my head virtually in the speaker box tweaking those parameters 'looking' for my sound, always feeling there was this 'missing' element I couldn't quite quantify, only to wake up the next day, plug in and immediately decided I had tweaked it into some weird tone corner and that it sounded like  warm puss. I remember staying up late nearly every night I had the thing, trying to get the drive/distortion to sound decent.
As I took more and more patches out of the chain and replaced them with my actual pedals. It just never stopped sounding and feeling wrong/unfamiliar/responding weird/dry harsh and non-intuitively. Now don't get me wrong, that GT would have been hellava useful at the time, I just couldn't dial in anything I could use and feel happy with. The final realization was the 'wrong-ness' of plugging in my good old fav drive alone, with no other effects or patches, and working out that it too was responding oddly. I realized that the second I took it out of the GT loop and went straight in to the amp, all was well with the world again.
The times I have borrowed Pods, I found more I played the Pod the more 'Pod-like' it ends up sounding to me... I usually then notice Poddy sounding recordings on TV adverts radio jingles etc.  for the next few days.
Those experiences, especially the expense and ultimate disappointment with the GT series made me the analog lover I am today.
I do use a Pandora for practice and ideas but there is usually a time limit before it tires the ears out.
The way I figure it, imitation raspberry cordial is still IMITATION, no matter how well it is done.
Hope those experiences are of some use to you.

[MarcoMike]

I don't have much experience about digitals... but I totally agree on the "weird feeling-response". the owner of my band's rehearsal room has a T.C.electronics NASA style pedal and I played with it a bit... only bad vibes from it... the only cool thing is storing presets... but that's not what this digital vs analog is about..

[Sir H C]

The biggest issue I would say with digital distortion is that distortion is a very non-linear and not necessarily time invariant effect.  So trying to model it with DSPs that really like linear and time invariant systems, it gets "false" sounding. 

[km-r]

not to mention most cheap and affordable digital distortions are made from low-bit DSPs which never sound good...
2c

[caress]

not to sound like the devil's advocate here, but there are a fair amount of software distortion/drive emulation plugins that do a decent job at getting a good "stompbox" distortion.  not only that, but they can generally go further into sound design territory (if that's what you're looking for), something most stompboxes cannot do.  they're just different tools, imo...

some decent ones for include izotope trash, NI guitar rig, bluetubes drive, a handful of smaller "boutique" plugins that i can't think of the names of.

i agree that digital stompbox/rack distortions generally sound horrible, though...

[petemoore]

  Digital Distortion is a fine aquaintance.
  I became very aquanted with a couple digital devices, have seen, tried...
  And noticed every one of them lacks:
  The live dynamic.
  The reason for it:
1  There's a live dynamic in tube amp / speaker / guitar wood / pickup / strings / air.
2  Digital is 'told' where to go and it says oh, I have a place 'you want to go here'?
  Substituting a tube for SS output to speaker is one thing, this can work quite well,  taking the output tube distortion characteristics and replacing it with...analog~ !
  Ramming a guitar output into a signal decoder, then a 'decider' of tone, then a signal uncoder is another deal entirely, however some question the ability of extremely powerful processing to replicate to some degree the influece that sympathetic vibration has on wood/air/strings and other metals, really it is a rediculous notion that such interactions could be pre-decided.
  There is always a predestination with digital. Seemingly endless varieties of replication until the novelty wears off, and the realization sets in that the tones offered are superb and enourmously tweekable, or that the tones offered sound only like a 'live rig' in some ways.
 
   

[tranceracer]

I use and have used multiple digital distortions and amp simulators from the Roland COSM to the J-Station, V-AMP Pro, POD, and Digitech Stomp Pedals.

They have their place and are very flexable units if used properly.  One thing most digital Amp simulations and digital distortions lack is dynamics.  Amp sims can and digital distortions can be a little anemic.  But what they lack in dynamics you gain in flexibility you can have a multitude of tones and effects at a press of a foot switch.  The band I was in played "dance music" everything from Santana, ZZ, Tabares, Heatwave, Lighthouse, Billy Ocean, Kool and the Gang etc. so I needed to have a wide variety of tones.  If I were playing only a certain genre i.e. vintage rock (which I love) then hands down it would be a tube rig.

Here are some interesting things I've found fiddling with digital amp simulators.

The most important and difficult thing about digital amp sims (or any digital device) is to NOT overdrive the input into digital distortion.  Most digital amp sims don't have an input clip light so the only way you know when the input is being overdriven is when it sounds shrill.  Many guitarists don't realize this and run their pedals cranked into the amp sim, like they do with a tube amps or SS to overdrive the amplifier's preamp stage.  This is a no-no for digital amp sims / digital devices.  You can still run your overdrives, effect and booster pedals before the digital hardware but have to be aware that the output levels of the last pedal is not overdriving the the amp sim input stage into digital distortion.

Another thing is, to get good tones from an amp sim, you MUST "dial it in" or tweak the settings to YOUR axe.  This can take hours.  Most amp sims don't come with good factory settings and this can take some time.  I plugged my friend's Strat into one of my favorite amp sim settings which was set for humbuckers and it sounded like crap.  we had to fiddle with the tone controls and find the right amp/cab emulator that worked.  But even all this tweaking sometimes, like with some amps, an acceptable tone just cant be dialed in.

Good luck on your thesis!

[carrejans]

The biggest issue I would say with digital distortion is that distortion is a very non-linear and not necessarily time invariant effect.  So trying to model it with DSPs that really like linear and time invariant systems, it gets "false" sounding. 

But you can make a look-up table of a tube response. Only problem is that it's static; I think.


I think the biggest reason why guitarist don't like digital distortion is that they heard digital distortions with cheap processors. Because of this the ADC and DAC is not good enough; and the processing power is not enough for complicated algorithms.

Do you think that with good enough processors, it is possible to make decent digital distortion?


Thank you all for your replies. Keep em coming.

[carrejans]

  Digital Distortion is a fine aquaintance.
  I became very aquanted with a couple digital devices, have seen, tried...
  And noticed every one of them lacks:
  The live dynamic.
  The reason for it:
1  There's a live dynamic in tube amp / speaker / guitar wood / pickup / strings / air.
2  Digital is 'told' where to go and it says oh, I have a place 'you want to go here'?
  Substituting a tube for SS output to speaker is one thing, this can work quite well,  taking the output tube distortion characteristics and replacing it with...analog~ !
  Ramming a guitar output into a signal decoder, then a 'decider' of tone, then a signal uncoder is another deal entirely, however some question the ability of extremely powerful processing to replicate to some degree the influece that sympathetic vibration has on wood/air/strings and other metals, really it is a rediculous notion that such interactions could be pre-decided.
  There is always a predestination with digital. Seemingly endless varieties of replication until the novelty wears off, and the realization sets in that the tones offered are superb and enourmously tweekable, or that the tones offered sound only like a 'live rig' in some ways.

You say: "the influece that sympathetic vibration has on wood/air/strings and other metals, really it is a rediculous notion that such interactions could be pre-decided."

I don't get this. There is still a guitar that you are using; so still wood/air/strings. No?
But if you were referring to the modelling fx; I understand.

But I don't want to discuss the modelling stuff. Just a plain distortion. (although that's some kind of modelling too; because it models a tube in a way)

[Toney]

 For me, part of the beauty of a singing tube amp is the interaction between edgy pickups and a driven power tubes. Breakup that is just 'so' with a bend. And most importantly, interaction from the proximity of the player, amp and the guitar. Riding all of that properly is a real joy. Kind of like surfing.
So, do you think proximity dynamics can be 'modeled'?

[carrejans]

I use and have used multiple digital distortions and amp simulators from the Roland COSM to the J-Station, V-AMP Pro, POD, and Digitech Stomp Pedals.

They have their place and are very flexable units if used properly.  One thing most digital Amp simulations and digital distortions lack is dynamics.  Amp sims can and digital distortions can be a little anemic.  But what they lack in dynamics you gain in flexibility you can have a multitude of tones and effects at a press of a foot switch.  The band I was in played "dance music" everything from Santana, ZZ, Tabares, Heatwave, Lighthouse, Billy Ocean, Kool and the Gang etc. so I needed to have a wide variety of tones.  If I were playing only a certain genre i.e. vintage rock (which I love) then hands down it would be a tube rig.

Here are some interesting things I've found fiddling with digital amp simulators.

The most important and difficult thing about digital amp sims (or any digital device) is to NOT overdrive the input into digital distortion.  Most digital amp sims don't have an input clip light so the only way you know when the input is being overdriven is when it sounds shrill.  Many guitarists don't realize this and run their pedals cranked into the amp sim, like they do with a tube amps or SS to overdrive the amplifier's preamp stage.  This is a no-no for digital amp sims / digital devices.  You can still run your overdrives, effect and booster pedals before the digital hardware but have to be aware that the output levels of the last pedal is not overdriving the the amp sim input stage into digital distortion.

Another thing is, to get good tones from an amp sim, you MUST "dial it in" or tweak the settings to YOUR axe.  This can take hours.  Most amp sims don't come with good factory settings and this can take some time.  I plugged my friend's Strat into one of my favorite amp sim settings which was set for humbuckers and it sounded like crap.  we had to fiddle with the tone controls and find the right amp/cab emulator that worked.  But even all this tweaking sometimes, like with some amps, an acceptable tone just cant be dialed in.

Good luck on your thesis!

Do you know why the lack dynamics and sound anaemic?

[Toney]

 
  Ramming a guitar output into a signal decoder, then a 'decider' of tone, then a signal uncoder is another deal entirely, however some question the ability of extremely powerful processing to replicate to some degree the influece that sympathetic vibration has on wood/air/strings and other metals, really it is a rediculous notion that such interactions could be pre-decided.

   

Exactly right Pete.
I know where your coming from.
That's the 'suck factor' that ruined even my favorite old analog drive when plugged in to the GT with no other patches running. The signal got analog digital converted - then again at the out put. Best mystery tone/dynamics kill yet! You see it sounded familiar but it lost all of it's life dynamics and sweetness. Much like an Mp3 with a low bitrate of a famliar song, but of course more than that too. A guitarists signal is interactive it's a work in progress, a two way street with the signal chain.
All the digital stuff I have used eventually feel 2 dimensional in a 3D world.
Maybe if I was in a wedding covers band or something I wouldn't care.

[Sir H C]

The biggest issue I would say with digital distortion is that distortion is a very non-linear and not necessarily time invariant effect.  So trying to model it with DSPs that really like linear and time invariant systems, it gets "false" sounding. 

But you can make a look-up table of a tube response. Only problem is that it's static; I think.


I think the biggest reason why guitarist don't like digital distortion is that they heard digital distortions with cheap processors. Because of this the ADC and DAC is not good enough; and the processing power is not enough for complicated algorithms.

Do you think that with good enough processors, it is possible to make decent digital distortion?


Thank you all for your replies. Keep em coming.

Lookup tables are static, also think about the dynamic range, how big a signal before the ADC starts clipping, or do you compress here thereby altering the sound/response of the circuit vs. an analog variant.  I love to slam some fuzzes with insane voltages from other pedals, and not sure you can do that with an digital version.

With enough power yes a great one could be made, the problem is that you have to control latency, work hard on the front end to make it robust but not intrusive, and work harder on the digital side.

In the end, it is probably still cheaper for Danelectro to do an analog circuit for this than digital.

[carrejans]

 
  Ramming a guitar output into a signal decoder, then a 'decider' of tone, then a signal uncoder is another deal entirely, however some question the ability of extremely powerful processing to replicate to some degree the influece that sympathetic vibration has on wood/air/strings and other metals, really it is a rediculous notion that such interactions could be pre-decided.

   

Exactly right Pete.
I know where your coming from.
That's the 'suck factor' that ruined even my favorite old analog drive when plugged in to the GT with no other patches running. The signal got analog digital converted - then again at the out put. Best mystery tone/dynamics kill yet! You see it sounded familiar but it lost all of it's life dynamics and sweetness. Much like an Mp3 with a low bitrate of a famliar song, but of course more than that too. A guitarists signal is interactive it's a work in progress, a two way street with the signal chain.
All the digital stuff I have used eventually feel 2 dimensional in a 3D world.
Maybe if I was in a wedding covers band or something I wouldn't care.

I doubt that it has anything to do with the ADC and DAC. The sampling and quantisation is not audible, I think. The rates are very high.

But keep the discussion going.  


And I'm still looking for an articles or so of comparisons of commercial digital distortion pedals? Plots or charts.

[eleanor296]

With good enough A/D and D/A converters, as well as a lot of processing power and good algorithms, you can get some really good sounds.
You know the Axe-FX by Fractal Audio, right?
All the rave right now in the rack world.  And MAN, that thing can sound righteous.  Go check out their clips.
As pointed out earlier, digital distortion "sucks" because what we're used to are low-line products that are sold to the masses by advertising a lot of DIFFERENT sounds, not necessarily a lot of GOOD sounds.

Andy

[carrejans]

With good enough A/D and D/A converters, as well as a lot of processing power and good algorithms, you can get some really good sounds.
You know the Axe-FX by Fractal Audio, right?
All the rave right now in the rack world.  And MAN, that thing can sound righteous.  Go check out their clips.
As pointed out earlier, digital distortion "sucks" because what we're used to are low-line products that are sold to the masses by advertising a lot of DIFFERENT sounds, not necessarily a lot of GOOD sounds.

Andy

No, never heard of the Axe-fx; but checking it out now.

I also believe that the big problem is that we are used to those low-cost digital distortions.

[carrejans]

Does anyone knows which methods or algorithms the commercially available digital distortions use?

I read interesting articles of Julius O. Smith about "numerical solutions of nonlinear ordinary differential equations". Has this been done in commercial pedals?



For those who are interested; I will post the titles of some real interesting papers about implementing digital distortion. They are all very recent publications. I can mail them to you, if you want.

- "Discrete-time models for nonlinear audio systems" - Schattschneider & Zölzer (1999) [-> Volterra]

- "Simulation of the diode limiter in guitar distortion circuits by numerical solution of ordinary differential equations" - Yeh & Abel & Smith (2007) [-> Numerical]

- "Numerical methods for simulation of guitar distortion circuits" - Yeh & Abel & Vladimirescu & Smith (2008) [-> numerical]

- "Simplified, physically-informed models of distortion and overdrive guitar effects pedals" - Yeh & Abel & Smith (2007) [-> simplification]

- "Wave digital simulation of a vacuum-tube amplifier" - Karjalainen & Pakarinen (2006) [-> WDF]

- "Non-linear digital implementation of a parametric analog tube ground cathode amplifier" - Santagata & Sarti & Tubaro (2007) [-> phenomenological model]

[puretube]

Digital may/might possibly have some future...

Analogue definitely has a past/immortality and present, and future...

[Processaurus]

This topic reminded me of a tidbit on digital distortion pedal design, I reverse engineered the analog in/out section of the Line 6 AM-4 amp modeler (same board as DM-4 fuzzbox modeler ) a while back and was interested to find that despite being a mono pedal, it uses a stereo processor.  What they did was there is normal gain going to one channel, and the other channel was fed with 47x the gain (a huge ~5v P-P signal hitting the ADC, from a moderate guitar level input), and was clipped by some diodes in the opamp's feedback loop pre processing.  The hi gain sounds, it's reasonable to assume, choose the high gain, analog-ily pre-distorted channel. 

Thought that was interesting as a peek into the designer's thinking, that if you were to apply hi gain digitally to a small signal, you'd lose resolution exponentially, as if you're heavily clipping something, you are hearing just a small window around the signal's zero crossing, and if that gain happens in the digital domain, poof, there goes all the dynamic resolution, and with it a measure of the instrument's character and liveliness