Parallel filter distortions...thing.

[patrick398]

VFE's DragonHound is also worth checking.

Great, thanks for that. Anybody care to explain how this is doing the filtering. I think i'm being stupid.
I thought C1, R3 and R2, C5 would provide the filters but the numbers just don't add up.

Sorry for the delay.
This effect is just 2 full range distortions in parallel with a panner at the end. Not exactly what you are doing, but I meant more for generating ideas. The idea of the DragonHound is to blend between flavors of overdrive. The Fender Pugilist works more or less the same.
R1, C1, and R2, C2, are just decoupling and limiting input. R3, R4 and R5 are bias resistors. C5 is there to cut some highs, probably RF.

The problem with multiband distortion, for me, besides level matching, is that it is not enough to filter the input. Overdrive/distortion/fuzz type effects create components with high frequencies. The bottom, mid and high end are creating high frequency content, so it may become overwhelming. It is just too much going on for me.

A single (semi)clean channel plus effect seems to be the sweet spot (for me).

Klon Centaur is a tri band effect, but as you increase overdrive in the mids, it lowers the high frequency volume. Clean bass is always present.

Thanks for that, lot's of good stuff in that dragon hound schem. When working with ICs is the filtering usually done with caps between input and output pins on the ic? The reason i ask is that i was just thinking about using an RC network before the IC input but in a lot of cases there would be no way to do this and not upset the biasing right?

My guitarist and I discussed similar topic few months ago and decide to try building it after New Year, however, there are a lots of great ideas here to study!

Let me know which direction you decide to go. I have decided i don't actually need another distortion at the moment but will likely pursue this just for educational purposes.

[Josh?]

The problem with multiband distortion, for me, besides level matching, is that it is not enough to filter the input. Overdrive/distortion/fuzz type effects create components with high frequencies. The bottom, mid and high end are creating high frequency content, so it may become overwhelming. It is just too much going on for me.

Would it make more sense to filter the outputs, rather than the inputs, then?

[Suicufnoc]

I've been tweaking a 4 band distortion pedal, and filtering both before and after the clipping circuits is magical

[marcelomd]

Thanks for that, lot's of good stuff in that dragon hound schem. When working with ICs is the filtering usually done with caps between input and output pins on the ic? The reason i ask is that i was just thinking about using an RC network before the IC input but in a lot of cases there would be no way to do this and not upset the biasing right?

Opamps can do most things "in parallel", so you can, if you want, add filtering in the feedback loop, or not =)
I would keep things simple at first, with the filtering before. If you do it right, you won't disturb the biasing. The important thing is not to have a direct path (meaning no capacitors) between the bias point and earth.

Someone mentioned VFE's Triumvirate Just use that as a base. Note the supply is bipolar, so if you want to use that, just change some grounds to the bias point.

Would it make more sense to filter the outputs, rather than the inputs, then?

I've been tweaking a 4 band distortion pedal, and filtering both before and after the clipping circuits is magical

For me, pre- and post-equalization are the magic sauce. The mechanism of distortion itself is secondary.

Mechanism of distortion is a great name =)

[Ben N]

For me, pre- and post-equalization are the magic sauce. The mechanism of distortion itself is secondary.

Mr. Keen been preaching that gospel for a long time.

[thetragichero]

I've worked on something quite close:

- A sort of crossover filter using a amz active splitter 1 in 3 out with added filter on each : a HP, a LP and BP.
- A switchable CMOS (4049) Drive and a CMOS Fuzz on each HP and BP (schematics from Tim Escobedo)
- A half clean boost on LP (4049, from Tim too i think)
- Active blender at the end

It was more like a newbie block-building game as you can see, but... after hours of work on the proto, I was quite disapointed by result.

What I learned from it :
- Before I add the active buffered mixer, it sounded like garbage.
- The stompbox was ridiculously big and with too many options
- Level matching between each drive / fuzz is really important
- Each Fuzz and Drive must have caricatural differences
- The ratio Size / Awsomeness was terrible...
- CMOS are awsome !

Also, I've been able to use a Chaise Bliss Brother and a Empress Multidrive, that was not as extreme as mine but sound close (and better)

on one of my breadboards I have something similar... lpf and hpf (I'll have to look at it/my notes to find out corner frequency and see if I remembered to put a buffer in front of it). lows go too a cmos sub octave/square wave distortion, mids/highs go to cmos muff. haven't wired ins/outs/pots, but the idea is to have a third switching jack that's for sub octave out that would make the blend control just a volume control for the cmos muff (I watched a really heavy two-piece band where the guitarist had a sub octave going to a bass amp and that got the hamster wheel spinning in my head)

[Mark Hammer]

A few considerations for multi-band distortions:

1) The amplitude of various parts of the spectrum differ in the original input signal, such that the drive or sensitivity of each band needs to be individually tailored.

2) Because distortion adds harmonics of the input, the content of the output of each path will overlaps substantially with the next band up, and potentially any one above that.....unless you insert filtering after the clipping.  How important that is will depend on the design-intent of using multiple bands.  If the goal is simply to get different-sounding tones, then it won't matter that much.  If the goal is to use an arrangement like the PAiA Quadrafuzz, where each pat has a volume control to mix the signals sort of like an EQ, then you'll want some post-clip filtering for each path, to maintain some spectral distinctness for each.