Guitar-like distortion techniques
-
- KVRist
- 66 posts since 29 Jun, 2006 from Bay area, CA
Hi, I've been wondering if there is a way to get that deep amp/cab combo sound out of straightforward digital distortion techniques. That is, just basic filters like biquads, waveshapers, and EQs.
What I've tried so far is this: highpass -> offset-soft-clip (so that even harmonics are emphasized) -> lowpass/EQ. The trouble though is that to get that all-out distortion sound, the gain going into the effect has to be so high that the sound loses all characteristics and it basically becomes a hard clip. Lower gains don't really sound "distorty" enough.
So, I was wondering if there are other tricks I could do. Has anyone here had any success with guitar-like distortion short of modeling analog electronics? Maybe tricks like distorting transients or various frequency bands differently? Also, what is meant when people say "3-stage distortion"?
Cheers,
Louis
What I've tried so far is this: highpass -> offset-soft-clip (so that even harmonics are emphasized) -> lowpass/EQ. The trouble though is that to get that all-out distortion sound, the gain going into the effect has to be so high that the sound loses all characteristics and it basically becomes a hard clip. Lower gains don't really sound "distorty" enough.
So, I was wondering if there are other tricks I could do. Has anyone here had any success with guitar-like distortion short of modeling analog electronics? Maybe tricks like distorting transients or various frequency bands differently? Also, what is meant when people say "3-stage distortion"?
Cheers,
Louis
OPL2 is forever
-
MeldaProduction MeldaProduction https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=176122
- KVRAF
- 14325 posts since 15 Mar, 2008 from Czech republic
Well, first there is no reason to have the saturator "clip" the signal. Just imagine atan function which devs like so much
. It goes to PI/2 or in the infinity, but it raises all the time, so even with very high gain it does not hard-clip the signal.
I'm not sure about multi-stage distortion, but it seems you simply apply lower gain to avoid harder saturation, and rather perform the same distortion multiple times. Everytime it creates some harmonics, so at the end you can theoretically get more harmonics with less gain.
Another thing are impulse responses. AFAIK NI usually uses truncated IRs to emulate combos, microphones etc. Just try the GuitarRig. When you disable the microphone emulation, it usually sounds like complete shit, just ordinary distortion you have done. Well, most guitarists say GuitarRig sounds like shit anyway
, but at least it proves the IRs are pretty important.
I'm not sure about multi-stage distortion, but it seems you simply apply lower gain to avoid harder saturation, and rather perform the same distortion multiple times. Everytime it creates some harmonics, so at the end you can theoretically get more harmonics with less gain.
Another thing are impulse responses. AFAIK NI usually uses truncated IRs to emulate combos, microphones etc. Just try the GuitarRig. When you disable the microphone emulation, it usually sounds like complete shit, just ordinary distortion you have done. Well, most guitarists say GuitarRig sounds like shit anyway
Last edited by MeldaProduction on Tue Jun 08, 2010 2:34 pm, edited 1 time in total.
- Beware the Quoth
- 35433 posts since 4 Sep, 2001 from R'lyeh Oceanic Amusement Park and Funfair
http://lmgtfy.com/?q=biquad+filterocmtime wrote:what is a BIQUAD ?
An idiot on Set Theory:
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
"In some cases there is an object called red that contains everything that is red. In much the same way a pot is a plate."
-
- KVRist
- 444 posts since 16 Dec, 2003 from here and now
don't take your pills never
take all your pills always
distortion? your ears first, loudy pression, too loud, blop!
then use what you please to divert any signal
here's the tip, sry there 's no dmg for that
distortion? your ears first, loudy pression, too loud, blop!
then use what you please to divert any signal
here's the tip, sry there 's no dmg for that
-
- KVRist
- 88 posts since 20 Nov, 2009
In general guitar-like distortion is highpass->several distorting stages->tonestack->speaker emulation.
To avoid aliasing this should be done with oversampling.
The distortion stages should only generate low order harmonics and each stage not to much. It is sufficient to have only one late stage to add more even order harmonics (though some 2nd order can be added in every stage). This can be done quite easy with polinomial waveshaping.
For the (quite important) tonestack you can either simulate standard designs or simply cut somewhere between 300 and 1000 Hz.
Simple speaker emulation consists of 2nd order highpass at about 100 Hz and a 4th (maybe 2nd) order lowpass at about 5 kHz.
Chris
To avoid aliasing this should be done with oversampling.
The distortion stages should only generate low order harmonics and each stage not to much. It is sufficient to have only one late stage to add more even order harmonics (though some 2nd order can be added in every stage). This can be done quite easy with polinomial waveshaping.
For the (quite important) tonestack you can either simulate standard designs or simply cut somewhere between 300 and 1000 Hz.
Simple speaker emulation consists of 2nd order highpass at about 100 Hz and a 4th (maybe 2nd) order lowpass at about 5 kHz.
Chris
-
- KVRist
- 213 posts since 30 Dec, 2006 from Darmstadt, Germany
For multiple distortion stages to be different from a single one, you must apply filtering in between the stages.mahaya wrote:In general guitar-like distortion is highpass->several distorting stages->tonestack->speaker emulation.
To avoid aliasing this should be done with oversampling.
The distortion stages should only generate low order harmonics and each stage not to much. It is sufficient to have only one late stage to add more even order harmonics (though some 2nd order can be added in every stage). This can be done quite easy with polinomial waveshaping.
For the (quite important) tonestack you can either simulate standard designs or simply cut somewhere between 300 and 1000 Hz.
Simple speaker emulation consists of 2nd order highpass at about 100 Hz and a 4th (maybe 2nd) order lowpass at about 5 kHz.
Chris
-
- KVRist
- Topic Starter
- 66 posts since 29 Jun, 2006 from Bay area, CA
Wow, this is easier than I thought it'd be.. I got a decent guitar sound out of Pure Data. The filters at the end of the chain need quite a bit more work, but it's coming along now: http://www.extentofthejam.com/attempt.wav .. maybe some resonances between the distortion stages would sound interesting
Thanks guys!
Thanks guys!
OPL2 is forever
-
- KVRian
- 657 posts since 9 Jul, 2006
what is pure data ? the software for audio processing that google finds when I enter "pure data" ?
very nice effect ,did you use a real guitar or string samples before the distortion and how many stages of eq and distortion did you use.
and what vsts ?
very nice effect ,did you use a real guitar or string samples before the distortion and how many stages of eq and distortion did you use.
and what vsts ?
-
- KVRist
- Topic Starter
- 66 posts since 29 Jun, 2006 from Bay area, CA
Pure Data is a freebie Max/MSP clone. It's very good for trying out new synth tricks before going to the trouble of coding them. For input, I used Slayer in clean mode (no fx, no amp, no cab). There were 6 stages of distortion, each with a lowpass and a highpass. Between each stage, I multiply the signal by the same amount. I assume this will result in the most crunch for the first stage because the subsequent stage's amplitudes cannot exceed 2/3 amplitude. At the end I have 3 bandpasses with varying degrees of bandwidth to shape the tone. Here's a sound sample using just a Karplus-Strong generator for input: www.extentofthejam.com/karplus.wav (and different tone settings).ocmtime wrote:what is pure data ? the software for audio processing that google finds when I enter "pure data" ?
very nice effect ,did you use a real guitar or string samples before the distortion and how many stages of eq and distortion did you use.
and what vsts ?
Hmm, I'm wondering if peaking filters would give me nicer results than bandpasses.
OPL2 is forever
-
- KVRist
- 156 posts since 8 May, 2007 from California
And, for completeness, Miller Puckette, who developed PD, also developed the original Max while at IRCAM.
-
- KVRian
- 657 posts since 9 Jul, 2006
are these three bandpasses affection three copies of the fignal .LouisG wrote: At the end I have 3 bandpasses with varying degrees of bandwidth to shape the tone.
are they used in a series of filters or in parallel ?
-
- KVRist
- Topic Starter
- 66 posts since 29 Jun, 2006 from Bay area, CA
They are in parallel, so what you get are 3 different bands at the end. I should improve my patch and then post itocmtime wrote:are these three bandpasses affection three copies of the fignal .LouisG wrote: At the end I have 3 bandpasses with varying degrees of bandwidth to shape the tone.
are they used in a series of filters or in parallel ?
OPL2 is forever
