Voice-operated wah?

[Mark Hammer]

Now, I know we're all pretty much committed to sticking with analog, but last night I stumbled onto a deal I couldn't pass up; a Behringer DSP-1000P (rackmount multi-FX) for $50.  Given that I had my eye out for one of their digital reverb pedals that normally sells for $54.99 (+tax) around here, getting ALL those reverbs and delays and other effects (plus MIDI control and setting memories) was kinda sweet.

One of the things it has is vocoding capability.  I only managed to sneak in about 5 minutes of using it this morning (need a mic preamp...badly!) before it was time to make lunches for everyone, but those few minutes got me to thinking.  Vocoding itself sounds kinda gimmicky, but lots of folks wouldn't mind having some kind of voice control over tone.  Talkboxes are nifty but a nuisance, with pretty low "tonal return on investment" (all that stuff for one sound).  Certainly there are things like autowahs, but they also have limits to how expressive they can be, relative to foot-operated wahs.  What if there were a way to have a "mouth-operated wah"?

I thought about it some more (thank goodness for uneventful commuter bus rides), and realized that the standard wah is simply a moving bandpass filter with a known fixed minimum and maximum centre frequency.  Pull your foot back and you "ooh".  Move it forward and you get "ah".  Hold that thought for a second.

Conventional analog vocoders operate by having an array of bandpass filters tied to rectifiers (envelope followers).  This array detects how much vocal energy is concentrated across the spectrum, and then uses that information to arrange for the relative gain in each of those bands for the signal being modified.  That's how it can make a guitar sound like a voice.  of course, it has to use the information from all those bands individually because it has to construct multiple resonances at once in order to mimic a voice.

But what if you were only concerned with a single resonance?  Could you get away with using much less information about the modulating signal so as to modify the modulated signal in a simple fashion?

So here's the idea, keeping in mind I'm not astute enough to design this on my own.  Imagine that we had two fairly broad bandpass filters, centred on the upper and lower limits of where a normal wah moves in.  Each of these has its own rectifier/follower.  The two rectifier outputs (heavily smoothed with no need for superfast attack) are fed to a circuit that compares their outputs relative to each other.  When the lower filter exceeds the higher one, a more negative or lower voltage is output.  When the higher one exceeds the lower one, a more positive or higher voltage is output.  How high or low would depend on the amplitudes of the two rectifier outputs relative to each other.

At the same time, we have a third bandpass filter which is the actual wah itself and is also some form of voltage controlled bandpass filter.  It may have merely centre-frequency as its only controllable parameter, or perhaps it has both frequency and resonance available for control.  That filter, in turn, is controlled by the output from the comparison circuit.  Step up to the mic and say "ah", and the filter swings upward.  Say "ooh" and it moves downward.  say "uhhhh" and it sticks around the middle of the range.

Is this "better" than a foot operated wah?  Not necessarily, but it frees up your foot to do other things you might want to be doing at the same time (can you say "Whammy pedal"?).  If one has vocal resonance control (perhaps a third path sums the two filter rectifier outputs and derives an overall envelope amplitude value to control filter resonance such that talking louder gets you greater resonance from the signal path filter), then that would provide a little more control than a foot-wah normally does.  Finally, it almost goes without saying that the "feel" of a foot-controlled tone modifier is different than the feel of a hand or voice-operated one.  This is part of the reason why synths often have breath controller inputs.

Now, it still isn't as compact as a little autowah pedal or even a foot-operated wah because there is still the matter of the mic to plug into it, the mic stand, etc. etc.  But it certainly isn't as cumbersome as a talkbox, and not as hokey as a vocoder can be sometimes.

So, feasible?  More importantly, has this already been done somewhere by someone?

[StephenGiles]

Hi Mark, that sounds interesting. Do have a look at my post 5 here:

http://www.diystompboxes.com/smfforum/index.php?topic=57145.0

and the links to the thread over at Prodigy Pro, because the frequency and peak detection circuitry could be relevant.

[StephenGiles]

Looks like we're on our own here!!

[Mark Hammer]

Some things catch fire more slowly than others, my friend.  Patience.

[mdh]

I think it's a really interesting thought, but I don't have much to offer in terms of design thoughts.  One thing that does occur to me from Mark's first post is that it could be really interesting to have a mouthpiece that responds to air flow and lip pressure, rather than voice.  Sort of like those Akai and Yamaha synth wind instruments, but just with a couple CV's out that could be used to sweep filters, phasers, whatever.  This is perhaps ranging a little wide of your topic, but as a decent sax player who's always looking for excuses to play guitar, I can imagine that I could do a lot with something like that.  It seems like it would be a little simpler electronically, but quite a bit more mechanically complex (how to detect air flow, making suitable lip pressure sensors).

[StephenGiles]

I think it's a really interesting thought, but I don't have much to offer in terms of design thoughts.  One thing that does occur to me from Mark's first post is that it could be really interesting to have a mouthpiece that responds to air flow and lip pressure, rather than voice.  Sort of like those Akai and Yamaha synth wind instruments, but just with a couple CV's out that could be used to sweep filters, phasers, whatever.  This is perhaps ranging a little wide of your topic, but as a decent sax player who's always looking for excuses to play guitar, I can imagine that I could do a lot with something like that.  It seems like it would be a little simpler electronically, but quite a bit more mechanically complex (how to detect air flow, making suitable lip pressure sensors).

There is a site providing circuits for just that but I can't remember the name just now.

I think it was Ian Fritz' Electronic Music Site but it seems to be down at the moment.

It's here
http://home.comcast.net/~ijfritz/sos_over.htm

[Mark Hammer]

Breath controllers (to be distinguished from wind controllers like the Yamaha WX5) generally provide a kind of "sealed field" with a mouthpiece that is inserted and covers the front of the mouth as well.  Something like that is pretty easily built.

You can buy big piezo discs.  I've bought some that are somewhere between 1.5 and 2" across.  You can also buy small plastic funnels.  Mount one of those discs at the wide end of the funnel, and make sure there is a decent seal around the edges so that any air applied to the narrow mouth end results in pressure applied to the disc.  I'll bet that blowing in the end of it will likely score you sufficient change in voltage from the disc that it can be fruitfully applied.

Just note that such a module doesn't care about frequency.  All it registers is pressure from you blowing.  Of course that pressure, in turn, can be harnessed to creatively voltage-control other things like VCA attack time (slow swells à la Slow Gear without the annoying envelope-detection required), or filter frequency/resonance or even LFO speed or wet/dry mix.  The E&MM Synblo (http://experimentalistsanonymous.com/diy/Schematics/Full%20Synths%20Drum%20Synths%20and%20Misc%20Synth/Synblo.pdf) is a nifty little project that makes use of a breath controller output.

[choklitlove]

i never thought about piezos as breath controllers... that's an awesome idea!

and mark- i ordered a digital reverb from music123 for $29.99- http://www.music123.com/Behringer-Digital-Reverb-i156778.music
it's definitely worth that.

[A.S.P.]

Looks like we're on our own here!!

"You`ll Never Walk Alone",
but sometimes you got to walk yourself...

[mdh]

Ooh, thanks for those links. That Son of Stealth thing looks insane... for MIDI controller purposes, I think I'll just buy an EWI4000S when I have the cash. But it looks like some of the modules might be just the ticket for some kind of electric saxitar monstrosity.

I hope I didn't derail your topic irreparably, Mark!

[R.G.]

I think it would be far simpler to get initial results to put some kind of sensor together that measures either jaw position/tension or air pressure.

For the first, you could simply have a rubber bulb you bite. The bulb transfers air pressure to a diaphram operated pot. I'm kind of scared of running electrical wires into someone's mouth for any reason. Imagine an electrocution microphone, but with it already inside your mouth. A headset which had a lever under your jaw that operated a pot would be nice, and could be made relatively unobtrusive if you worked at it. In fact, given the range of grimacing we got from people trying out amps and effects at the Dallas Guitar Show, it would come as second nature to a guitarist to modulate with his mouth.

The air pressure one could be simpler mechanically. You could use one of the semiconductor pressure transducers connected to a tube which is near your mouth. When you want to wah, you grab said tube with your lips and blow. Sense pressure with a comparator and disable the wah if pressure is less than X.

Wasn't there an EH pedal that did mouth pressure sensing? The Soul Kiss? Or is my memory failing?

[Mark Hammer]

No, your memory is accurate.  T'was the Soul Kiss, which we had numerous discussions about over the years.  Alex Pretini has a project of it over at Topopiccione  IIRC.

The thing I like about the blow-into-it over the stick-it-in-your-mouth approach, over and above safety issues, is that you can't go "1, 2, 3, 4..." or indeed engage in ANY communication with others when you have this....thing...in your mouth.  Far easier to have something you can physically dislodge from by simply moving away.

There I go, getting all practical again....

[StephenGiles]

 

[R.G.]

The thing I like about the blow-into-it over the stick-it-in-your-mouth approach, over and above safety issues, is that you can't go "1, 2, 3, 4..." or indeed engage in ANY communication with others when you have this....thing...in your mouth.  Far easier to have something you can physically dislodge from by simply moving away.

Thank you Mark. I have to confess that you gave me the giggliest moment of the day, because the very first thing that lept into my mind was of a mike stand with a plastic tube running up it and out to the end of a bent-over gooseneck. The next image was of the Thamesmen singing "You know what want" from the movie, with Nigel and Derek playing while blowing into their microphone stands in unison.

I had to grab the arms of the office chair to keep from rolling into the floor while I laughed.

I can just see Nigel blowing into the mike stand tube while bowing his guitar with a violin.

[StephenGiles]

When the lower filter exceeds the higher one, a more negative or lower voltage is output.  When the higher one exceeds the lower one, a more positive or higher voltage is output.  How high or low would depend on the amplitudes of the two rectifier outputs relative to each other.

Didn't the What compressor have something on these lines?

http://dt.prohosting.com/hacks/whatparts.txt

I'm not so sure having read the circuit description, but trial & error and a breadboard is a wonderful thing!

Still thinking out loud, maybe there is something to be borrowed from a compressor/expander gate where the control voltages from each bundle of circuitry are juggled to produce compression when not gating and vice versa.

[Paul Perry (Frostwave)]

Remember that thing a year ago or so, where a camera looks at the shape of your mouth & operates a filter (or whatever) accordingly?
I've thought about voice control of a wah, I think that - if all you are trying to do is to sweep as you change the frequency of your voice - you could have two partly overlapping fairly low Q bandpass filters, and have a circuit to output the ratio of the envelopes from each. (using ratio, means your voice can concentrate on pitch, not level. Ratio circuits are used in absorbtion spectrometers.)