Miking a violin

My wife plays violin, and she uses a bridge pickup. To me, the sound is unnatural - I’d call it all string and no body. I’ve been faking it on the board by adding a peak around 400 Hz. This seems to give the closest sound to what I get from a regular mic at close range with no other sounds to interfere. We talked to another violinist with a bridge pickup, and she says just crank the treble way up. Funny thing is her instrument sounds pretty good that way, and it’s nearly the opposite of what we’re doing. Using the regular mic is not feasible - it picks up too much other sound. Would a contact mic solve the problem? We’re on a tight budget right now, so price/performance tradeoff is probably unavoidable.

2nd question: What in the bowing technique affects sub-harmonic generation? Does the pitch drop an octave in this mode, or is it entirely a timbre change?

Contact mics are horrible on violins.
Most bridge pick-ups are piezo and require a very high load impedance - somewhere around 1 MegaOhm.
If a piezo pickup goes straight into a low impedance input you loose all of the bottom end and it’ll sound thin and scratchy.
If she’s not using a pre-amp with a high impedance input, and you’re on a tight budget, get a graphic Eq pedal. They’re pretty cheap and the input impedance is usually fairly high, plus you have an Eq pedal.

The bowing technique for creating subharmonics is mostly about bow pressure but also requires very controlled smooth movement across the strings. It’s possible to hit the G bellow the bottom string, which puts it in the same range as a 'Cello. If the pressure is too great or the movement not smooth enough it’ll sound like a frog farting.

We’re using a passive direct box that’s supposed to have a hi-Z input, but it doesn’t say how high. She has a volume pedal with a 250k pot, which she used for a while and then took out. If the load imp were low (dynamic mic range), all the control range would be squeezed into the top end of the pot, which it wasn’t (that was my test so I remember it). For a while, she was using an instrument amp with a hi-Z input just so she could use an earphone for a personal monitor. She found the combination cumbersome and not too effective for the purpose, so gave up on that. The amp has a line out XLR, so connection to the system was electrical. I had to add an isolation transformer since we have phantom power. I don’t even remember what the violin sounded like back then, but it would be worth an experiment. Will try it.

The pickup is a Piezo, and I will measure its capacitance so I can determine the minimum load resistance required. I’ll also devise a means of measuring the D-box input impedance with a 600 Ohm (any way to insert symbols on this forum?) load on the XLR end. Y’know, if this is the culprit I betcha I can build a phantom-powered isolation amp that will fix it economically. People are always forgetting to turn off battery-powered stuff, and Lisa leaves her instrument plugged in so a jack switch wouldn’t help. No EQ unless we need it there - controls are always the most expensive parts. Although the design project would be more fun if it included EQ :sunglasses: .

Thanks for shooting down my contact mic idea before I spent $$ on one.

Would that be a bullfrog or a leaf peeper?

My guess is “not very”. I’d be interested to hear what it is.
If I remember correctly my (cheap) graphic pedal is about 750k.

I wish they wouldn’t use that term :imp: hiZ can mean anything from 2k to 10M (some violin pick-ups recommend as high as 10M :open_mouth: )

As far as I know, only like this :

Very true, which is why I built my pre-amp to have very low current drain. I’m not too bad at unplugging it, but I probably use it up to 8 hours a week and Duracell PP3 lasts over a year in it.

With enough vigour it could be a decent impression of a cane toad farting.

I measured the d-box input z. It’s 21.8k resistive in parallel with about 360 pF, which must be the winding capacitance. I measured it from 100 Hz to 20 kHz. It probably goes inductive at low enough frequency, but that shouldn’t affect the violin. With the pickup+cable capacitance of about 3 nF, this puts the LF -3dB point around 2 kHz :open_mouth: . So my fudge in the channel EQ is making sense. I could add a capacitor across the input to lower the LF pole. Lowering the pole by a decade would cost me 20dB in signal, but it’s a reasonable temporary fix and the board has enough gain. Since the pickup is a charge source, adding a cap across it wouldn’t add a HF pole.

The instrument amp input measures 180k resistive. Not at all impressive, but it beats the dbox by about an order of magnitude. I’m gonna make a phantom-powered DI box just for the violin. I’ll call it the “Fiddle D Box” :laughing: .

Those measurements are around what I expected. Thanks for posting your results.

Somewhere in the order of 200k is about average for instrument inputs on mixing desks (way too low for most piezo fiddle pick-ups). For directly connecting to a mixer, a series resistor can be used between the pick-up and a microphone input. I’ve used this approach successfully by making up a lead with two 500k resistors (one on each of pins 2 and 3 of an XLR plug). It’s not an ideal solution as it is dependent on the characteristics of the input, the pick-up, and the pick-up lead (in some cases it will buzz like crazy) but it often works and can be a very quick and cheap solution if you always use the same pick-up and desk.

My next experiment will be just to add a capacitor across the input to the direct box. Pickup + cable capacitance is around 3 nF and I’d like to get the corner frequency at least down to 100 Hz (an octave below the G string). Still using the dbox, increasing the total C to about 60 nF should do it (in exchange for 26 dB of signal). Using a shunt capacitor has two advantages over using a series resistor:

1: It theoretically doesn’t affect the high frequency response. Since the source impedance is essentially pure capacitive reactance, if the load were purely capacitive it would form a flat voltage divider. No matter where you put the resistor, there’s capacitance downstream of it. Let’s say I add 500k in series with the line, and I put it right at the dbox to minimize noise pickup and keep the cable capacitance on the upstream side. The dbox has about 360 pF of capacitance across its input per my measurements (based on impedance phase angle vs frequency). The resistor sees this in series with the pickup and cable, so the effective capacitance is about 320 pF. 500k x 320 pF = 160 us, for a HF breakpoint of about a kilohertz! YMMV

2: It actually reduces the noise by lowering the overall impedance, although since it also reduces the signal (as does a series resistor) it doesn’t improve the SNR. Both methods will degrade the SNR because of the front-end noise in the mixer. The series resistor increases the impedance, making the dbox input circuit a soft target for noise, and besides that there’s the thermal (Johnson) noise which increases per the square root of the total resistance.

I tried the shunt capacitor experiment, using 39 nF which is less than I originally intended. I happened to have one that fit nicely in the strain relief of one of those old genderless microphone connectors for which we also had a 1/4" adapter of the same vintage. I made from this a removable capacitive load that can be plugged into the “output” jack of the direct box, which is actually in parallel with the “input” and “ground lift” jacks. This avoids having one dbox that’s different from the rest.

I had my wife bow each open string in sequence, followed by the highest note she could manage. I think I heard one of those frog farts (not a cane toad). We did this with and without the added capacitor. Resulting spectra were a surprise. The 39 nF cap has most of its effect below 500 Hz. Finally discovered the reason: I measured the dbox input impedance with a 600 Ohm load on the XLR connector. Our board (Peavey RQ 2318) has 2.2k input resistance on the mic inputs. So forget the 21.8k measurement. It’s probably closer to 80k and I’m wondering if the transformer is properly loaded. I will repeat the dbox measurements with a 2.2k load.

There is some HF rolloff starting around 10-13 kHz with the capacitor added. This could be due to some equivalent series resistance in the pickup. The overall sound, however, is much improved (at least to my ears, which don’t register much above 14k). I will change the capacitor after I redo the dbox test.

Hello, I’m a violinist and I came across your post. I’m not sure what sort of bridge pickup she’s using, but the general characteristics of what she’s experiencing are expected. I use the “Violinissimo” pickup, which is a type of piezo pickup that I absolutely adore, at least for the type of sound I’m doing. I don’t miss anything and the direct sound feels very close and full and even quite raw, not “breathy” and subdued like I’ve experienced with a microphone. The thing is, this pickup has three contacts – two for under the bridge feet and a third for tonality (I guess) that you can place in a bridge slot, either on the low end or on the high end, depending on what sort of sound you’re going for. I can’t stand very bright sound most of the time, it just hurts my ears, and so I take the third contact and place it in the bridge slot under my G string. This warms up the sound.

As far as equalization for my recordings, I’ve really enjoyed the RCA Victor 1947 setting in Audacity. It’s beaten all of the custom work I’ve done, although I’m sure I could find a better setting.

Pick-ups under the feet provide a much “warmer” sound than pick-ups mounted into the bridge as they pick up more sound from the body of the violin whereas a pick-up in the bridge picks up a greater proportion of sound from the strings. The trade-off is that pick-ups under the feet are more susceptible to feedback due to sound from the speakers causing the body to resonate, which is then picked up by the transducers and sent through the amplifier to the speakers and round and round.

The third transducer can be left detached from the body, leaving the feet pick-ups to do the work. This is often a good arrangement for a fairly natural sound at relatively low volumes. The feet pick-ups pick up quite a lot of sound from the body of the violin

Placing the third transducer into one of the bridge wing slots will pick up a lot more sound directly from the strings with a relatively small proportion of sound picked up from the body by this transducer. This can give a more “immediate” sound that is brighter and tighter than using the feet transducers alone. Placing it on the G string side will (not surprisingly) pick up more from the lower strings whereas placing it on the E string side will pick-up more from the upper strings. The maximum “volume-before-feedback” level is likely to be substantially improved by use of the in-bridge transducer.

One or two in-bridge pick-ups without the feet pick-ups will provide maximum “volume before feedback”, but these generally require quite a lot of carefully tuned equalisation to prevent the sound from being overly bright. Some 3-transducer pick-ups allow the “feet” transducers to be inserted into the wing slots and the third transducer can then either be left unattached or used under the bridge.

Violin pick-ups are an inexact science - arguably as much an art as a science and there is no such thing as “the perfect” violin pick-up. The only way to get a “true” and “natural” violin sound is from a microphone, but the lively acoustic construction of a violin makes this horribly susceptible to feedback and is only viable for very gentle sound reinforcement. Experimentation can produce dramatic differences to the sound produced.

I’m now using 15 nF across the pickup. With the 2.2 k load of the board on the dbox, which probably expects a 600 Ohm load, the input impedance is in the 50-70 k range over most of the spectrum. It’s capacitive above, and inductive below, 2 kHz. The 15 nF cap results in a 2-pole response with -3dB point at 100 Hz and has very little effect on the high frequencies. It’s pretty close to a Butterworth response curve, at least in simulation. The experiments with the violin are not consistent enough to compare spectra with any real precision. Now the downside: I sometimes hear a kind of nautical creak like the rigging of a sailboat but much more rapid. Turning down the bass on the channel eliminates most of it, but that defeats the purpose of adding the capacitor. I’ve never noticed this sound with a microphone. The pickup is at the E string end of the bridge, and the sound occurs mainly with the high strings.

Hi there!

Why don’t u use programed violins? They are so much easier to mix than real ones and the sound is so realistic nowadays…

Have you ever heard of live music?

@Steve: My wife says “Thanks”!

@Emmah: This was her great-great-grandfather’s violin. 'nuf said?

@Paxtonbt: The multipoint pickup sounds like the way to go. Also sounds expensive (we’re on a zero-income budget). Do the 3 transducers have separate cables for external mixing? I don’t have a good mental picture of the vibration modes of the body under the bridge, although I’m reasonably sure the transmission from strings to body would be of opposite phase in the two feet. Are the foot elements connected in phase opposition? If so, I would think this would tend to cancel some of the acoustic feedback. The sound post makes the body asymmetric, so the cancellation would be compromised.

Our situation is church service, so using Audacity to eq the violin wouldn’t help during the service. I also don’t have a separate violin channel in the recording, so can’t do it later unless she plays a solo or a duet and I put her in one channel. Might try it electronically when I get around to building that phantom-powered DI box - looks simple enough. Thanks!

General question: Could the rapid “sailboat rigging” noise I mentioned in my Feb 28 post be a matter of bowing technique? Lisa has never had lessons, although she’s been playing now for about 12 years.

Just tried the RCA 1947 eq on my test recordings. It really boosts the “rigging noise” although it does restore the missing bass in the recordings without the added capacitor. Sound is too muffled though. I agree one could come up with a better setting.

Bow pressure, during normal playing, will cause mechanical distortions in the bridge. These are normally inaudible, but bridge pick-ups can pick them up. Depending on how rapidly the bass rolls off, sub-sonic frequencies will be at a very low level, but the better the bass response of the pick-up, the more low frequency thumping, creaking and other weird sounds will become apparent. A steep low-cut filter can often reduce these effects, but it’s a careful balance to avoid making the sound “thin” again. When playing with other instruments these unwanted sounds will usually be masked by other sounds, so it is generally only a problem with solo playing. Barbera pick-ups get round this problem by using an almost solid and very rigid bridge that flexes very little, but the downside is that the violin will then be very quiet with a heavily “muted” sound (like playing with a bridge mute) so they are only really suitable for an instrument that is dedicated to being plugged in. Barbera pick-ups are also rather expensive. Violin Bridge, Violin Pickups, Violin Pickup, Professional Violin Pickups, Acoustic Violin Pickups, Electric Violin Pickups.

A gold mine of information here. Many thanks! I think my DI box design will include a tunable 2-pole HPF. The rest of the spectrum I can take care of satisfactorily with the channel eq on the board.

That is quite a dilemma.