Quack tamed by proper impedance matching - why?

[Villamarzia]

A bad impedance matching can enhance piezo quack, as far as I know and tested, but can someone explain why? And if you have a piezo into a 1 Mohm input with an improved quack over a 200k ohm input, will it improve even more with a 10Mohm input or is there a limit above which quack remains? And more in general and scientifically, what produces the infamous quack?

[martingitdave]

From the Interwebs:

The impedance of capacitors and inductors in a circuit depend on the frequency of the electric signal. The impedance of an inductor is directly proportional to frequency, while the impedance of a capacitor is inversely proportional to frequency.

By varying the input impedance of the amplifier, you are changing the frequency response of the whole system. The rule of thumb is that each stage of amplification should should have a 10X impedance value.

The common wisdom in acoustic guitar amplification is that piezos generally sound more natural with higher impedance. What defines "higher" is different for each pickup and amplifier, and listeners ear. Many under saddle piezos sound best with a 1M Ohm or higher input impedance. Most preamps designed for under saddle piezos have values of 5M Ohm to 10M Ohm. Some sound board transducers like the K&K sound best with 500K Ohm to 1M Ohm. But, there is no hard and fast rule, and you can try and find what suits your ears.

If the quack is gone for you at 1M Ohm, and you like the sound, you may not find any improvement by going higher.

[Cuki79]

Originally Posted by Villamarzia

A bad impedance matching can enhance piezo quack, as far as I know and tested, but can someone explain why? And if you have a piezo into a 1 Mohm input with an improved quack over a 200k ohm input, will it improve even more with a 10Mohm input or is there a limit above which quack remains? And more in general and scientifically, what produces the infamous quack?

Your tone posses different frequencies. The first stage of your amplification system (PA speaker + mixer, amp...) is called a preamplifier. Depending on its input impedance, a low cut frequency will be defined.


Most piezo needs about 1 MOhm impedance input to provde a low-cut frequency compatible with the frequency response of a guitar. If you use a too low impedance input, the low cut frequency might be 500 Hz or 800 Hz.

The result will be a huge lack of low end.

Now keep in mind that sound preception is about balance. A lack of low end might be perceived as a giant high mid boost.

I would say the piezo quack artifact is situated around 3-6 kHz. So a giant mid boost will "enhance" the piezo quack perception.

Some acoustic amps, are overly bassy or woofy to "mask" the piezo quack, but the best way in my opinion is to use IR convolution (Tonedexter, Fishman Aura, homemade...)

my 2 cents,
Cuki

PS: If using a 1 Mohm input impedance enables you to get all the audible low end frequencies, there is no use going to 10 Mohm. There is nothing to valuable to hear under 40Hz for example. The lowest guitar frequency is around 80Hz (low E open string).

[Brent Hahn]

FWIW, if you have an instrument with a piezo and an onboard preamp, which most piezo-equipped guitars have, the whole impedance issue is out of your hands. If you have quack, it's baked in.

As an FOH mixer (or at least I used to be), I'll generally treat that by yanking down the 2k-ish range, sometimes severely. This will, of course, take some of the "real acoustic guitar" frequencies down along with the quack, but it'll still fill up the speakers and reinforce the vocal nicely and give the audience a satisfying musical performance.

[James May]

I and the team at Audio Sprockets have studied this topic, here are the main points of our understanding:

The thing that most people refer to as quack is a result of under saddle piezo transducers picking up a lot more direct string sound than body sound, due to their location. I would also describe this as a boingy, rubber band sound. It also applies to in-saddle transducers like the Baggs LB6 and other designs with individual piezo elements for each string.

It is not the sound of the piezo crystals—they are in effect perfectly linear transducers. It is the direct sound of the strings. Contrary to popular belief, they do not get more quacky when you strum harder due to some mythical non-linearity. You just notice it more because you are giving the strings a more full range excitation.

These crystals are typically small and have a low capacitance which means that if used passively, you'll need to work into at least 1Mohm in order to avoid losing low end. If you do lose low end, the perceived tonal balance will shift toward the highs and it will sound more brittle, and perhaps more quacky.

As has been mentioned, if your pickup system is active on board, the sound is baked in and no reasonable difference in impedance that you plug into will matter.

You can tweak the mids with an EQ, it can help a bit, but won't get around the inherent issue which is the direct sound of the strings dominating.

Using a ToneDexter (or other properly made IR) can effectively eliminate almost all the quack.

So, the premise of this thread that quack can be tamed by proper impedance matching is mostly false.

[Villamarzia]

Originally Posted by James May

It is not the sound of the piezo crystals—they are in effect perfectly linear transducers. It is the direct sound of the strings. Contrary to popular belief, they do not get more quacky when you strum harder due to some mythical non-linearity. You just notice it more because you are giving the strings a more full range excitation.

Thanks all and James, this makes lot of sense to me. Let me tell my situation:

I have a classical guitar, which I play with nails and strum from time to time (including some flamenco techniques, rasqueado and such). It's a luthier high-end guitar. I've installed a Barbera pickup which I prefer over K&K or Schatten HFN-C both in tone and feedback resistance. However, I notice a tiny bit of quackiness to it. My rig is simple: barbera into a MarkBass acoustic head into a Daedalus cabinet. The input on the head is 1Mohm.

How to erase also that last tiny quackiness, provided I love how it sounds (just like my guitar) and that I've been using IRs but here I want a pure analog sound, maybe just for romantic reasons.. In the past I had good results, but with different sets, using Lehle sunday driver, BBE maximizer and Origin compressors. But still, I don't want to alter my sound or color it. I was thinking to add a Rupert Neve DI between the guitar and the amp, using the mic input of the amplifier. This way I would have a 2.2Mohm input and a good DI for situations where I don't want/need to bring the amp with me. But, based on your advices, adding a DI and increasing the input impedance would not result in anything audible.. Any other analog advice???

PS: I'd like to avoid 2-channel solutions, eg adding a mic (been there) because of loud volume gigs..

[James May]

Originally Posted by Villamarzia

Thanks all and James, this makes lot of sense to me. Let me tell my situation:

I have a classical guitar, which I play with nails and strum from time to time (including some flamenco techniques, rasqueado and such). It's a luthier high-end guitar. I've installed a Barbera pickup which I prefer over K&K or Schatten HFN-C both in tone and feedback resistance. However, I notice a tiny bit of quackiness to it. My rig is simple: barbera into a MarkBass acoustic head into a Daedalus cabinet. The input on the head is 1Mohm.

How to erase also that last tiny quackiness, provided I love how it sounds (just like my guitar) and that I've been using IRs but here I want a pure analog sound, maybe just for romantic reasons.. In the past I had good results, but with different sets, using Lehle sunday driver, BBE maximizer and Origin compressors. But still, I don't want to alter my sound or color it. I was thinking to add a Rupert Neve DI between the guitar and the amp, using the mic input of the amplifier. This way I would have a 2.2Mohm input and a good DI for situations where I don't want/need to bring the amp with me. But, based on your advices, adding a DI and increasing the input impedance would not result in anything audible.. Any other analog advice???

PS: I'd like to avoid 2-channel solutions, eg adding a mic (been there) because of loud volume gigs..

I think 1 - 2.2M impedance will be fine and higher won't help your issue. My only advice, if you want to stay analog, would be a parametric EQ to dip the mids a little, as Brent Hahn suggested, as well as perhaps a bit of high frequency rolloff to soften the overall sound.

One thing worth noting is that playing through a speaker/amp in a room with its own bit of natural reverb will do a lot to hide the quack. The ear just doesn't notice it as much. (This is one of the reasons most pickup makers demo their pickups this way.)

[Villamarzia]

Originally Posted by James May

I think 1 - 2.2M impedance will be fine and higher won't help your issue.

James, just to be sure I get your advice right: do you mean that having 2.2Mohm instead of 1Mohm will probably result in some improvement but going higher than 2.2 won’t make any difference? Regarding the EQ, I have some gear I can use to try it, thanks!

[James May]

I think most passive piezo pickups are just as good at 1Mohm as at something higher. By just as good, I mean you'll have all the note fundamentals and plenty of lows below that to make percussive attacks sound full. (Remember, a sharp attack with a pick has an impulsive nature which contains a whole spectrum of frequencies - you can thank Mssr. Fourier for the theory explaining this.)

There are a couple of very low capacitance pickups that would benefit from 2.2M or even 5M or 10M. Some that come to mind are the PUTW and Open to Source sensors, perhaps the McIntyre Feather as well. But to be clear, the higher impedance won't reduce quack, it will just extend the low frequency response.

[Brent Hahn]

Originally Posted by James May

... to be clear, the higher impedance won't reduce quack, it will just extend the low frequency response.

This can actually make a huge difference with upright bass pickups -- there are lots of passive piezos on basses out in the world, way more than guitars.

On the other hand, with a passive piezo on a mandolin or a ukulele, a very-hi-z preamp can actually cause problems. You'll get a massive thump with each plucked note that you have to HPF away.

[Cuki79]

Originally Posted by James May

The thing that most people refer to as quack is a result of under saddle piezo transducers picking up a lot more direct string sound than body sound, due to their location.

I agree with James. My little experience with making SBT pickups is the further you place them from the saddle line, the less quack you get.

For example, I think the Schatten HFN post is a way to move the sensor away from the saddle and still be placed under the saddle line in order to achieve natural balance.

I have to be careful with the use of "piezo" quack term now.

However James, and I know we disagree on this, I feel somehow the piezo disc keeps a bit of the brass disc resonance once installed. I once computed with finite element method those resonances and the main mode was exactly in the 3-5 kHz range. Of course your argument about the glue making a perfect coupling to the top makes sens. But I feel that the velocity mismatch between the top (wood) and the disc (brass) would still be strong enough to maintain a bit of brass disc resonance. I have no measurement to prove it thought. So my question is: Since you think the piezo quack is only due to position of the sensor, and since the Ultratonic pickup can avoid feedback by destructive interference, why don't you elaborate a method to locate the optimal position of the piezo discs and avoid the necessity of Tonedexter?


Thanks James,
Cuki

[Villamarzia]

Originally Posted by Cuki79

why don't you elaborate a method to locate the optimal position of the piezo discs and avoid the necessity of Tonedexter?


Thanks James,
Cuki

if you were not mentioning the ceasing necessity of Tonedexter, you might had some chances to convince him...

[Cuki79]

Originally Posted by Villamarzia

if you were not mentioning the ceasing necessity of Tonedexter, you might had some chances to convince him...

James would sell more pickups and less pedals!

[jonfields45]

Cuki, it is good to see you back.

Both magnetic and piezo pickups have a series parasitic, inductance for a magnetic, and capacitance for a piezo. To sound good you want your load to be much higher than the parasitic's impedance. 1M ohm is a great choice for both types of passive pickups (irrelevant if already buffered with a preamp). As maybe the only person who's RainSong CO-DR1100N2 with a K&K sounded bad into 10M ohm, maybe that might not be a good choice for some. In any case too low an input impedance rolls off the highs for a mag and the lows for a piezo. Plug a passive pickup into a mixer mic input to hear it first hand.

Quack, I thought, was slang to refer to that classic UST sound which is not always a bad thing. They work well in rock bands as an example. They can be perfected with an IR but the load impedance is fixing a different problem.

IMO, if your guitar has a factory pickup and it isn't good enough, buy a ToneDexter. It is easy, well built, has all the preamp features you probably need anyway, and sounds great. If you like to tinker try my IR generator or Cuki's.

If your guitar does not have a pickup, then good luck getting advice on the AGF :~).

I like some Soundboard Transducers better than others and all better than a UST, but I find little broad agreement on SBTs with and between my AGF peers.

A second source mic is great if you can get away with it. I am amazed Doug Young uses an IR and a Mic. That requires some very discerning ears and/or an interest in tinkering.

[martingitdave]

Originally Posted by Cuki79

I agree with James. My little experience with making SBT pickups is the further you place them from the saddle line, the less quack you get.

For example, I think the Schatten HFN post is a way to move the sensor away from the saddle and still be placed under the saddle line in order to achieve natural balance.

I have to be careful with the use of "piezo" quack term now.

However James, and I know we disagree on this, I feel somehow the piezo disc keeps a bit of the brass disc resonance once installed. I once computed with finite element method those resonances and the main mode was exactly in the 3-5 kHz range. Of course your argument about the glue making a perfect coupling to the top makes sens. But I feel that the velocity mismatch between the top (wood) and the disc (brass) would still be strong enough to maintain a bit of brass disc resonance. I have no measurement to prove it thought. So my question is: Since you think the piezo quack is only due to position of the sensor, and since the Ultratonic pickup can avoid feedback by destructive interference, why don't you elaborate a method to locate the optimal position of the piezo discs and avoid the necessity of Tonedexter?


Thanks James,
Cuki

I might suggest that no pickup will sound like a mic without processing. It’s inside the guitar and only picking up vibrations on the bridge. Regardless if the pickup is brass, plastic, or an exotic material, it will never capture the sound of the air leaving the instrument, no? Better, more elaborate pickups with structures like yours, Dazzo, HFN, etc. seem to have the best chance of capturing sound that’s more complex, but none sound the a mic in front of a guitar. Just “mic like” in characteristics.