Advice on multi-band distortion unit

[robthequiet]

Hoping someone can help me zoom in on the current thinking about multi-band disto -- saw a demo of a box called Four Eyes by Fairfiled Circuitry which is a state-variable-filter-based distortion. I was previously thinking along the lines of using different clipping methods on each of the three bands of a SVF thus my interest in the pedal. I also found an article by Jack Orman on using different types of clipping diodes on the LPF or HPF side of a circuit which is also cool. My thinking with the SVF is that I could have a sweep control for the center frequency and tailor the distortion for each band, but would it be just as good if not better to simply have a drive stage in each band of a parametric eq? Not trying to reinvent the MT-2, I have one and imo it lacks tonal versatility (i.e. "meat") despite the mid sweeps. Any thoughts/advice would be most welcome, thx!

[aron]

I never got to play through a multi-band distortion. Wouldn't mind trying one.

[blackieNYC]

An MT2 with the ability to distort each band independently, right down to cleanish, would be awesome. But I think you might not need to go parametric at all. A three band tonestack would give you incredible variety. I've done this in two-band and I love it. A simple panner adjusts lows v highs.
Clippers in series with a pot is good. Thing is, with three bands, you'll need adjustable makeup gain for each band. Because the clippers would ... clip.  If the mids are clipped and the bass is not, that bass voltage will be much higher.
Three band parametric, or passive tonestack, you should just get out the breadboard and wing it.

[J0K3RX]

look here at the top of the list, schematic and board for download > https://soundwavedesign.wordpress.com/2014/01/01/a-small-update/

This guy has put up a lot of really good designs for download, worth checking out! Seems he dropped off the scene in 2014 no activity since then?

[robthequiet]

Thanks J0K3RX and blackieNYC -- Yes, on the MT-2, although I haven't tried any of the mods that might switch or pan between clippers, which some have tried. The Quartet is reminescent of Craig Anderton's Quadrafuzz which looks like a bit of a bear to build. But the state-variable thing not getting any votes so far. My thinking os also that bass notes need a different type of clipping than high-mids or highs, i.e. maybe do a germanium type on the lows for a finer grain and LEDs on the mids to break up with more edge, I don't know. Warming up the breadboard. All advice welcome.

[TejfolvonDanone]

Don't mind me i just bump in for update.

[Mark Hammer]

First off, Guillaume Fairfield is a local boy, and a helluva nice guy.  Doesn't come out with a lot of designs, but the ones he coes out with are thoughtful, and he rarely tries to reinvent the wheel.  Like myself, he adopts a design philosophy of most flexibility with the fewest possible controls.  And for the record, the legending on his pedals is hand-stamped, letter by letter, by a buddy of his.  Also a nice guy.  I look forward to chatting with Guillaume again at a the upcoming Ottawa GUitar Show ( http://ottawaguitarshow.com/ ).

Second, there are different sorts of objectives for multi-band distortion, and I think a person needs to ponder which of those objectives holds the highest priority for them before launching into a build or design.

• Reducing intermodulation distortion: One of the perks of clipping chunks of the spectrum independently is that the result yields less intermodulation distortion.  One of the consequences of that is that there is somewhat better note separation in the midst of all that harmonic content.  I'm just talking through my hat, here, but I imagine reducing any sideband products in the final mono mix allows us to perceptually organize all that harmonic content more easily into this-goes-with-this-note.
• Routing bands separately: A multi-band distortion could be mixed down to mono for the output, but it could also be processed as several parallel paths, with the option for routing different bands to different channels.  The original Quadrafuzz permitted that by having a pair of normalized send/return jacks for each band, ahead of the mixer stage.
• Being more selective about where you want your gronk: Splitting the signal into multiple bands allows for setting the clipping threshold individually for the bands, such that one can produce "wolf-tones" selectively, whereby a given note rings harder, with more emphasis, for dramatic effect.  To some extent, that can be easily done by situating a parametric or quasi-parametric resonant boost ahead of whatever distortion unit/s one is using so that a particular band pushes the distortion unit harder.  Billy Gibbons tones await.  In some respects it may be easier with a parametric/distortion combo than with a multi-band unit, siply because of the ease of dialing in which note or notes you want more gronk in.
• Balancing drive more effectively/evenly:  One of the things that made the TUbe Screamer so desirable and popular was the manner in which it used highpass filtering ahead of the clipping, and lowpass afer it, to deliver a tone that provided relatively even clipping across the spectrum.  Keep in mind that the biggest chunk of your signal's amplitude lives in the basement, such that it pushes the clipping circuit harder than your B and E strings do.  The TS nudged the signal amplitude of the lower strings downward and a little further from the clipping threshold so that all the strings were kept a similar distance away from or close to the clipping threshold.  A multiband distortion can, in principle, allow for a more precise and tunable version of that, by having gain and level controls for each band.  If a given band is much closer to the clipping threshold than the others, you can simply turn its individual gain down, and restore overall bandwidth by turning that band's level up.  Think of it like a TS-9, but a lot smarter.  The old Quadrafuzz had a lot of tricks, but lacked this one.  Although I suppose anyone with a bit of chops could easily install adjustable gain pots for each band if they wanted.

Aiming for any single one of these does not preclude the others.  But coming up with a feasible and usable set of controls in a reasonable package requires that one ask "Well, what do I most want this to be able to do for me?".  Form follows function.

FWIW, I came up with a 2-band distortion several years back, that allowed for either independently adjustable or reciprocally adjustable (one goes up while the other goes down) gain and level balancing of the two channels.  Could be realized with only 3 knobs.  I won't say all sounds were wonderful, but there was a lot of different sounds...if that's what you're looking for.

[robthequiet]

Thanks, Mark, I was kinda hoping you might weigh in with some perspective. The main thrust of my project is to line up, as you indicate, the "this goes with this note" so that the lows and highs cohere with each other for a stronger overall tone. I'm not locked in with using a particular form of clipping, i.e. maybe  one thing works on lows such as germanium or jfet, or another works better on highs. For all I know, a nicely balanced 4049 circuit might work everywhere but I'm looking at some lab time anyway. I have only nice things to say about the Fairfield Four Eyes, it has a lot of range.