Double Bass If it walks like a piezo, and quacks like a duck...

[Chris Fitzgerald]

Hmmm... To the extent it is in the voicing or the response of the amp or speaker cab, then it should be apparent regardless of the input device. That is, you should get it with a piezo, a mic, or any other transducer. I wonder how much of the objectionable sound you hear is just a function of the amp. Hard to say from here.

But what if, as I wondered aloud in my previous post, the slightly resonant frequency of the pickup/bass/setup in question matched the slightly resonant frequency of the amp setup when set flat? Wouldn't this likely produce a resonant frequency hump that might in some ways sound waterfowlesque to the imaginative listener? If so, it's not a big deal at all if the player knew which frequencies to roll back to tame it. Or maybe I should just quit while I'm behind.

EDIT: I see we cross posted, and that you've already answered this. Time for this hayseed boy to shut up and go practice!

I've never been a fan of the idea of variable input impedance. It's a parameter that alters frequency response in a way that's highly dependent on the device plugged into it. I much prefer optimization of input impedance followed by a really good tone shaping circuit (I prefer parametric ones, as on the EA amps). Optimization is what that switch does for different condenser mics that Chris mentioned. IMO, a variable impedance knob approaches (not "is," but "approaches") electronic snake-oil in many cases.

I use a Rev Solo 2, fdeck's pre-amp/HPF, an EA iamp200, and a Wizzy 12. No quacking, no honking. I can produce honking if I throw on a BassMax plugged into a less desirable rig. Just my experience.

Understood and appreciated. At the same time, you have previously said that proper loading for most piezo pickups is roughly 1 Mohm and above. Your EA combo has a 1 Mohm impedance, and your fdeck pre has a 10Mohm impedance. Does your rig sound different with the fdeck pre in the FX loop than it does with it out front? If so, might there be a sweet spot somewhere in between 1 and 10 Mohm that might sound even better? I honestly have no idea, but am just curious.

 

[drurb]

But what if, as I wondered aloud in my previous post, the slightly resonant frequency of the pickup/bass/setup in question matched the slightly resonant frequency of the amp setup when set flat? Wouldn't this likely produce a resonant frequency hump that might in some ways sound waterfowlesque to the imaginative listener? If so, it's not a big deal at all if the player knew which frequencies to roll back to tame it. Or maybe I should just quit while I'm behind.

You are not behind. You're getting really good at this. Ding! Correct again! If the resonant frequency of the amp/cabinet happened to lie in the region of the resonant frequency of the piezo, you'd get, not a quack, but a QUACK! Yes, you could try to tame it but, ideally, you'd want to create an anti-resonance (notch) that matched the shape of the peak in the piezo pickup system. If it's a really sharp resonance, meaning it only affects a very small range of frequencies (as is the case for a raw piezo with no case at all), then you'd need a notch filter. If it's a bit broader then you need something a bit more gentle than what would be considered to be a notch. That's why parametric tone shaping is so handy!

Understood and appreciated. At the same time, you have previously said that proper loading for most piezo pickups is roughly 1 Mohm and above. Your EA combo has a 1 Mohm impedance, and your fdeck pre has a 10Mohm impedance. Does your rig sound different with the fdeck pre in the FX loop than it does with it out front? If so, might there be a sweet spot somewhere in between 1 and 10 Mohm that might sound even better? I honestly have no idea, but am just curious.

I can honestly say that I never noticed much of a difference (beyond the action of the HPF) whether I use the fdeck pre-amp in front or not. Here's the thing about the impedance loading. For a piezo, once it's high enough to give you the low end response, going higher doesn't really do any better for you because once the input impedance is high enough so that the desirable low frequencies passed by the piezo are not rolled off, then that's really it. It's like once you make the door wide enough for the 350 lb guys to get in the bar, making it wider so that a 600 lb guy can fit doesn't really get you much because they rarely show up and when they do, well, you probably don't want 'em anyway!*

*The above example was used for intuitive and comical illustration and in no way was meant to disparage the super-obese.

 

[Marc Piane]

If the resonant frequency of the amp/cabinet happened to lie in the region of the resonant frequency of the piezo, you'd get, not a quack, but a QUACK!

I assume the same could be true if the room itself favors the offending frequency?

I played a room last night that I have played many times and tried several different amps in over the years and it always seems to get the Realist to quack a bit which is not an easy task IME.

 

[agedhorse]

Yes, you guys are getting this well.

There are a lot of variables, the mechanical packaging of the crystal most certainly affects the quack and resonance, as does the mounting method. As does the overall system response including the amp, the bass and the speaker. The speaker also feeds some acoustic information back into the bass, so this can affect the gain and Q ofthe entire system response (transfer function).

The input impedance of an amp can also serve to slightly damp the resonance response, just another variable in te equation. It's not generally really effective though. There is a point depending on the crustal's gepmetry where increasing the input impedance of the amp produces no additional benefit but beyond that point can interject additional problems and issues to the robustness and noise of the system. I have found this to be somewhere between 500k and 1M. There are exceptions, but they are very rare.

 

[Chris Fitzgerald]

I assume the same could be true if the room itself favors the offending frequency?

Definitely. One of the reasons I'd like to narrow in on what this frequency is is to know how to cut it in a room that brings out the ducks and geese in my setup. I have some experimenting with parametric EQ to do before I can even begin to test this.

 

[drurb]

Yes, you guys are getting this well.

Thanks, I ought to have gotten it long ago, given that these matters fall squarely within my own profession.

The input impedance of an amp can also serve to slightly damp the resonance response, just another variable in te equation. It's not generally really effective though. There is a point depending on the crustal's gepmetry where increasing the input impedance of the amp produces no additional benefit...

Yup, essentially what I said in posts #20 and #22.

 

[drurb]

Definitely. One of the reasons I'd like to narrow in on what this frequency is is to know how to cut it in a room that brings out the ducks and geese in my setup. I have some experimenting with parametric EQ to do before I can even begin to test this.

I've written a little primer on how to learn to use parametric tone shaping intuitively and effectively. I'll try to dig it up for you.

 

[Monte]

I've written a little primer on how to learn to use parametric tone shaping intuitively and effectively. I'll try to dig it up for you.

Please Post this!!

I don't do piezos, but with the Dyn B I am finding a parametric incredibly useful. However, I don't really know what I'm doing other than how to notch out the one frequency I know about. I'd like to better understand so I could maybe improve it a bit more with the other bands.

 

[UncleFluffy]

Please consider this a rough "morning coffee" sketch.

This may be a suitable buffer/filter to match impedance and EQ out the resonant peak.

It's a high-impedance input stage followed by an notch filter adjustable over 100Hz-1kHz and 0-6dB of cut. Should be buildable for $10 and an hour or two's work.

Caveat: I've only simulated it in Spice, not built it.

Any comments from the more experienced analogue electronics guys here are more than welcome.

 

[Lumpy]

I don't do piezos, but with the Dyn B I am finding a parametric incredibly useful. However, I don't really know what I'm doing other than how to notch out the one frequency I know about.

That's essentially all there is to it.

Separate the "room acoustics" from the "instrument acoustics". That is, before any instruments or voices are introduced into a room, sound gurus will use a signal generator to sweep through the entire audio spectrum and measure/record the response with a mic in the audience. Wherever there are peaks/wolves/coyotes/birdies in the freq response, they CUT (never boost) that offending frequency with their HOUSE EQ.

Now the room is flat. Theoretically, it will reproduce ALL frequencies equally.

Then separately, do each instrument the same way with it's own EQ device. ie play your bass, bagpipe, banjo, electric brake drum, whatever, through it's pickup/mic etc. Play the instrument through the entire range of it's voice. With singers, I'll ask them to sing a slow "siren" from low to high. Wherever I hear a peak, I cut the EQ at that point. I personally don't measure any more as I've gotten good at hearing the frequency and making good educated guesses at where the offending wolf is at. With your bass, play a slow multi-octave scale from lowest to highest (or hi to low) and note where (have someone else note where) the offending tones are at. Cut your EQ at that point.

On some instruments (elec guitars, et al) "flat" response is not exactly what they might be looking for. So they might cut at points which create a notable dip in the freq response.

Never boost, always cut. After each cut, increase the gain to make up for the reduced volume that happens as a result of the EQ cut.

In any of the EQ references above, substitute whatever you use, 31 band graphic, parametric, bass-mid-treb, etc. Obviously the more divisions of the audio spectrum you have, the more narrow the frequency of the cut. 31 band EQs cut ~1/3 of an octave with each slider. A typical bass/mid/treb tone stack on a guitar amp is obviously much wider and reduces the response of several octaves. The NOTCH filter in a lot of instrument preamps is usually a sort of parametric device. It cuts a given amount of dB and you have control over WHERE that cut happens as you sweep the control from one end to the other.


Lumpy

Can you play country music?
Sure. Which country would you like?

Link Removed

 

[agedhorse]

UnkleFluffy, a couple of observations regarding the filter:

1. 4.7Meg input, especially with it being part of the input bias network of U2 could be responsible for a pretty high noise floor and possibly high DC offset (above/below Vcc/2). 1M is probably a better choice. Will also be more RFI friendly.

2. The basic filter circuit is passive (but buffered in and out). This makes for a very shallow and wide (low Q) notch. Most resonant problems are better served with a higher Q notch filter, a filter that resides within a feedback loop with gain.

Here's a really good filter tutorial from TI, something that shows the depth of what can be done simply with active filters. For the ckt on 16-31, all you would need it the input buffer, the output is already buffered.

In fact, this whole TI series is a fantastic (though simplified) glimpse into the world of op amps and a good read for anybody interested in applications. You can blow off the math, just look at HOW they approach each solution. It's the equiv. of a $200 text book IMO.

[Invalid or Expired Link Removed]

 

[drurb]

UnkleFluffy, a couple of observations regarding the filter:

2. The basic filter circuit is passive (but buffered in and out). This makes for a very shallow and wide (low Q) notch. Most resonant problems are better served with a higher Q notch filter, a filter that resides within a feedback loop with gain.

You beat me to it by seconds. I was going to comment that the attenuation was too shallow and the bandwidth of the filter, perhaps, too wide. Not sure I'd use op-amps though...

 

[drurb]

That's essentially all there is to it.

Well, uh... there's more. You gave a nice description of tailoring the house sound. Using a parametric tone shape circuit when the house is not involved is another matter. There's more to using parametric tone shaping than simply attenuating peaks in the spectrum. You may also wish to know that "flattening" the response of a room almost always leads to undesirable acoustics. The audiophile and psychoacoustic community has known this for quite a umber of decades. Part of the reason has to do with the human perception of the room response as it is affected by reverberation. This is often not taken into account by steady-state measures of the type you describe. This is not to say that what you describe, on a rough scale, is not helpful. Indeed, it is quite useful to tame the large resonances.

 

[UncleFluffy]

UnkleFluffy, a couple of observations regarding the filter:

1. 4.7Meg input, especially with it being part of the input bias network of U2 could be responsible for a pretty high noise floor and possibly high DC offset (above/below Vcc/2). 1M is probably a better choice. Will also be more RFI friendly.

2. The basic filter circuit is passive (but buffered in and out). This makes for a very shallow and wide (low Q) notch. Most resonant problems are better served with a higher Q notch filter, a filter that resides within a feedback loop with gain.

[Invalid or Expired Link Removed]

I'm a digital guy revisiting the analogue world, so your wisdom is appreciated.

#1 - These are CMOS opamps that only pull 2pA or so bias current so I'm not so sure that 4.7M is that big of a problem.

#2 - Good point, thanks. I was aiming for a scrap-of-stripboard-and-an-altoids-tin design, but it shouldn't be too difficult to make it active without blowing up the component count.

Much thanks for the TI reference.

Not sure I'd use op-amps though...

Why not? Some op-amps are noisy pieces of ****, true, but not all. Is there a specific problem with using op-amps that I should be aware of, or is it a matter of taste?

 

[Marc Piane]

then take into account all the ****ed up acoustic situations we put ourselves in. With less than ideal listening environments things sound different depending on where you are sitting. I did a gig about a month back and ehochberg popped in. I felt like my sound sucked and I was tweaking all night. After the set he wanted to know what gear I was using 'cause he liked the tone. Go figure.

 

[agedhorse]

I'm a digital guy revisiting the analogue world, so your wisdom is appreciated.

#1 - These are CMOS opamps that only pull 2pA or so bias current so I'm not so sure that 4.7M is that big of a problem.
I don't recall them being CMOS, I show a limit spec. of 200pA Ib. My biggest concern would be due to noise and RFI intrusion. The cable you hang onto the input is a big antenna too.

#2 - Good point, thanks. I was aiming for a scrap-of-stripboard-and-an-altoids-tin design, but it shouldn't be too difficult to make it active without blowing up the component count.

I think you could do it with no increase in parts and maybe less. Depends on how clever you get.

I see no problem using op amps for this application. In fact, there are an awful lot of outstanding op amps on the market these days, better than the discrete op amps I have designed.

 

[agedhorse]

I'm a digital guy revisiting the analogue world, so your wisdom is appreciated.

#1 - These are CMOS opamps that only pull 2pA or so bias current so I'm not so sure that 4.7M is that big of a problem.

I don't recall the TL-072 being CMOS, I show a limit spec. of 200pA Ib. My biggest concern would be due to noise and RFI intrusion. The cable you hang onto the input is a big antenna too.

#2 - Good point, thanks. I was aiming for a scrap-of-stripboard-and-an-altoids-tin design, but it shouldn't be too difficult to make it active without blowing up the component count.

I think you could do it with no increase in parts and maybe less. Depends on how clever you get.

I see no problem using IC op amps for this application. In fact, there are an awful lot of outstanding op amps on the market these days, better than the discrete op amps I have designed.

 

[fdeck]

I don't recall the TL-072 being CMOS, I show a limit spec. of 200pA Ib. My biggest concern would be due to noise and RFI intrusion. The cable you hang onto the input is a big antenna too.

Ah, you've fallen victim to TI's lovely part numbering scheme. Fluffy called out a TLC072, which is a CMOS part.

For 9-V battery operation and moderate gain, I'm fond of the TLC226x parts, which have tolerably low noise and are rail-to-rail.

 

[UncleFluffy]

I don't recall the TL-072 being CMOS, I show a limit spec. of 200pA Ib. My biggest concern would be due to noise and RFI intrusion. The cable you hang onto the input is a big antenna too.

That's tricky - if the pickup has an output impedance of about 5M, then this needs to at least match that. (Looking around, I've seen them modelled as a voltage source in series with a 500pF cap). Realistically, this would have to sit within a foot or so of the pickup - you're right, you can't run long cable runs between high impedance endpoints.

I think you could do it with no increase in parts and maybe less. Depends on how clever you get.

3 amps is easy. 2 seems doable with a bit of pencil-and-paper work. The cost difference is pennies, it's just an elegance issue.

(I'm sorry if I've dragged the thread off-topic, I'll go away and go build this thing now)

 

[agedhorse]

I did miss the "C" in the part number. I have found CMOS op amps to be pretty noisy under these conditions. I'll have to look again, that was quite a few years ago that I tried some.

The issue I see is that the output impedance of most piezos is much lower than 5M.

Take a piezo into a non-inverting buffer and adjust the resistor to ground (use a pot) and see where you find the point of diminishing return. Listen to it.