From Piezo Quack to Wood Resonance: Acoustic Filtering DSP

When the Acoustic Guitar Stops Sounding Like Wood

The guitar sounds right in the living room. Warm lows, a midrange with air, treble that rings instead of bites. Then it runs through a pickup and something changes. The low end thins out, the attack turns brittle, and a hollow, plastic squawk rides on top of every strum, loudest on the plain strings.

Players call it quack. Sound engineers call it the piezo problem. Neither name explains it, which is why so many attempted fixes miss. Knowing the mechanism matters, because every working solution, from patient EQ work to dedicated filtering processors like the TC Helicon Play Acoustic, undoes the same physical distortion through the same sequence.

Why a Piezo Pickup Hears What It Hears

An under-saddle piezo pickup is not a microphone. It does not listen to the air. It measures pressure at exactly one mechanical point: the saddle, where six strings press down on a thin strip of piezoceramic crystal. Squeeze the crystal and it produces a voltage proportional to the stress. That principle, the direct piezoelectric effect, is the entire sensing chain: no membrane, no air coupling, no soundboard.

The trouble starts with what this single point misses. An acoustic guitar makes tone through a system of coupled parts: the top plate moving air, the back and sides reflecting it, the internal air cavity resonating, the braces damping selected frequencies. By the time a string's vibration becomes the sound a listener hears across a room, that mechanical filter has shaped it heavily. The saddle pickup samples the vibration before most of that shaping happens. It hears the string, not the wood.

The pickup's own frequency response makes things worse. A piezo element is stiff and lightly damped, so its output climbs toward a natural resonance, typically in the 2 to 5 kilohertz range. That happens to be where the human ear reads harshness, and where the acoustic body would normally have absorbed energy. The result is a spectral double fault: the body's natural treble roll-off is missing, and the sensor adds its own peak on top.

There is also the impedance story. A piezo element behaves like a small capacitor, with very high source impedance at low frequencies. Feed it into a standard line input and the loading rolls off bass even further. A proper instrument input, usually one megohm or more, is not a convenience; it is part of the transducer. Plug a piezo-equipped guitar into the wrong jack and the low end is already gone before any processing begins.

The reference point is worth naming. When a player says the amplified sound is wrong, the standard of correctness is the acoustic sound in a room, a signal shaped by mechanical filters that centuries of lutherie refined. The pickup skips all of it. Loudness makes the missing pieces more obvious, not less. That is why "just make it louder" never works.

Then comes the transient. A piezo crystal tracks the string's initial impulse with almost no mechanical smoothing, so the pick attack arrives carrying far more high-frequency energy than a microphone would ever hear. The microphone listens after the top plate has damped and delayed the attack. This sharp burst of upper-mid energy at the front of every note is the quack itself, and it is not a defect in the pickup, exactly; it is the sound of information the body never got to process. That distinction matters, because no static EQ can fix it: cutting the offending band removes the quack but takes clarity with it, and no EQ curve can restore resonances that were never captured.

What the Wood Would Have Done

The insight that changed acoustic amplification: you do not need to recreate the guitar's sound from nothing. You need a filter that approximates what the body would have done to the signal, applied in the reverse sequence of the damage. This is the approach behind a family of acoustic resonance processors, of which the BodyRez filter set is the most widely recognized example.

The engineering task breaks into three jobs. The first is to pull down the pickup's resonant peak with a precisely placed cut, usually a few decibels in the 2 to 5 kilohertz region. The second is to rebuild the bottom: the body of a dreadnought couples a large top to a big air cavity, producing a low-end bloom around 90 to 110 hertz that the saddle never transmits, so a shelf or bell filter in that region restores perceived warmth. The third is to reshape the attack, either by smoothing transients or by letting a dynamic stage react only when the player digs in.

That third job explains why fixed EQ keeps failing: a static curve cannot know how hard the next note arrives. Filter hard enough to tame a heavy strum and soft fingerpicking goes dull; filter lightly and the quack returns on accents. This is where adaptive tone correction enters, a level-sensing stage that applies correction in proportion to input energy. Quiet passages pass through nearly untouched, preserving detail, while hard attacks get stronger filtering automatically. The circuit behaves less like an EQ and more like an automatic gain control that operates on specific frequency regions.

Amplitude is only half the story. Every filter also shifts phase, and the piezo's instant attack, arriving before the body's bloom would have, is part of why amplified acoustics can feel stiff. A well-tuned resonance curve adjusts both at once: cutting the resonant peak softens the leading edge while the low-end boost slows the perceived onset. Players describe the result as more forgiving without knowing why; what they are hearing is the attack being pulled closer to the timing the room would have delivered.

Factory presets matter because pickup and body combinations vary enormously. An under-saddle pickup in a small parlor guitar, a soundboard transducer in a jumbo, and a magnetic soundhole pickup in a dreadnought each produce a different starting spectrum. A resonance processor typically ships with several factory models, each a complete filter recipe tuned to a body style, rather than asking the player to build curves from scratch. BodyRez ships with four such models, and choosing the right one gets a player most of the way there; the remaining fine-tuning is measured in small adjustments, not full reconstructions.

One honest caveat: no filter can fully recover information the pickup never captured. The wood's subtle decay, the air cavity's breathing, the sympathetic ring of open strings, all are absent from the source signal. What resonance filtering achieves is subtraction, not reconstruction. It removes the artifacts that announce "transducer" and lets the player's actual technique come through. In most live settings that subtraction is enough, because the room and the P.A. system add their own character anyway.

Setting Up the Signal Chain Without the Manual

A setup guide for the TC Helicon Play Acoustic, or any processor of this type, reduces to one sequence of decisions. Manuals for this class of gear tend to list features without explaining signal flow, which is why the learning curve feels steep. Work through the chain from start to finish and most of the confusion disappears.

The instrument jack comes first. The guitar needs a high-impedance input, ideally one megohm or above, or the piezo loses treble and bass together. If the unit offers a dedicated guitar input, that is the only correct destination; a shared input expecting a microphone will load the pickup incorrectly, and the damage is done before any processing begins.

Then gain structure. Set the guitar level so normal strumming peaks well below clipping, because the filters that follow need headroom. A signal already compressed or clipped gives them nothing to work with, and a signal too close to the noise floor makes the adaptive stage hunt. A strong, clean average level, with the loudest passages still 6 to 10 decibels short of the ceiling, is a sensible target.

Only after levels are stable should the body model be chosen, and it should be chosen before reverb, delay, or harmony. The filter recipe is the foundation; everything else decorates it. Match the model to the actual construction of the guitar: body size first, pickup type second. A dreadnought curve on a parlor guitar bloats the low end, and the reverse sounds thin, which is the exact complaint the processor is meant to fix.

Resonance depth gets set conservatively. Ears adapt within minutes, so what sounds better after ten minutes of continuous listening may simply be louder and brighter. Check against the bypassed tone often during the first session. If filtered and bypassed tones are hard to separate on soft passages but clearly different on hard strums, the adaptive stage is behaving as designed.

Vocal and guitar paths stay independent. A unit that takes both a microphone and an instrument runs them through separate chains before mixing, and that separation should be preserved. One reverb setting rarely flatters both a voice and a guitar; they fight for the same midrange. Independent outputs exist for the same reason: a front-of-house engineer can only balance signals that arrive on separate channels.

Save a baseline preset before experimenting. USB preset management makes a factory reset easy, but resetting also erases the last hour of work. A stored baseline costs ten seconds and keeps every later experiment reversible.

Harmony comes last. Chord-following vocal harmony measures pitch and chord quality from the guitar signal, so everything above directly determines how well it tracks. Clean levels, a matched body model, and tamed transients are not guitar polish; they are harmony setup in disguise. On units with a short looper, around 15 seconds with overdub, the loop also lands here, after the tone chain, so the captured phrase carries the corrected tone rather than the raw quack.

Where Processor Architectures Part Ways

Not every stage tool approaches this problem the same way, and the differences are architectural, not cosmetic. They show up in three places: how the guitar enters the unit, how many independent feeds leave it, and how much processing power goes to the instrument rather than the voice.

The input question sets the ceiling for guitar processing. Some designs give the instrument a dedicated high-impedance channel with its own filter bank and effects; others detect chords through a simpler auxiliary path to keep cost and size down. The BOSS VE-8 takes that second route, using the guitar mainly as a chord reference for the vocal engine, which suits performers who want harmony above all. Both approaches work; they simply prioritize different parts of the stage routine.

Output topology follows from the same thinking. A single mixed output asks the front-of-house engineer to take the whole stage sound as one signal; two outputs allow a split between vocal and guitar; three independent XLR feeds, one for the voice, one for the instrument, and one stereo mix, give the engineer full control without extra direct boxes. Which layout matters depends on the venue: a busker running into one powered speaker needs only the mixed feed, while a church setup with separate vocal and instrument channels benefits from full separation.

Units in this category cluster tightly in cost, roughly the high 200s to the mid 300s, and the differences mostly track DSP scope: how much processing goes to vocal harmony algorithms, how much to instrument filtering. None of these designs is objectively wrong. Each maps to a different stage routine. The useful question is not about ranking them; it is about which architecture matches how a given performer actually routes signal on stage.

When the Harmony Follows the Wrong Chord

Chord-following harmony is the feature most likely to fail on stage, and its failures share a common root. The engine must extract pitch and chord quality from a guitar signal in real time, a hard problem that gets harder as the signal degrades. Most troubleshooting in this category is really signal hygiene.

The first suspect is level. A guitar signal too low for the detector produces harmony that wavers or lands on wrong intervals, especially on the fifth and sixth strings. Raising the input gain so the chord reference is strong and steady fixes more tracking problems than any other single change.

The second suspect is left-hand cleanliness. Fretting squeaks and half-muted strings read as spurious transients, and a detector that follows energy instead of pitch will chase them. Playing closer to the neck produces a warmer, rounder transient that the detector reads more easily. Simplifying voicings helps too: scale templates handle common triads and sevenths with far more confidence than stacked ninths and altered extensions.

Then there is the interaction with the tone chain itself. If the resonance filter is set deep enough to carve out the upper mids, it can remove exactly the harmonic content the pitch detector relies on. When harmony misbehaves, back the filter off a notch before suspecting the harmony engine; the two systems share the same input, and the filter sits earlier in the chain.

Ground hum deserves a mention because it is common in small setups. Buskers and house-concert players often share a power strip with a mixer, a charger, and an amplifier, and ground loops announce themselves as a low buzz that follows the player everywhere. A 9-volt supply that ships separately invites whatever adapter happens to be nearby. An isolated supply, or at least a consistent one, removes a whole class of intermittent noise.

Last, keep a reset path. When a stored preset has been edited into confusion, restoring factory state is faster than untangling it. With presets backed up over USB, resetting costs nothing, and the baseline preset saved earlier becomes the recovery point.

The Filter as an Instrument

Acoustic amplification is a translation problem. The pickup converts mechanical vibration into voltage and loses the body's contribution along the way. Every stage that follows either adds artifacts or removes them. Good designs remove.

That is a strange way to think about a processor. Musicians reach for equipment that adds things: reverb, chorus, harmony, layers of sound. Yet the filter at the front of the chain earns its keep by subtraction, by deleting the harshness, the brittleness, and the plastic snap that announce an amplified guitar. When it works, nobody notices it. The audience simply hears an acoustic guitar, only louder.

Constraint shapes design here. Given a signal with a resonant peak, a missing low end, and a violent transient, the engineering response is not a bigger amplifier or a fancier effect; it is three filters and a level sensor, arranged in the right sequence: modest components, precisely placed. That pattern repeats across audio engineering: the phono preamp's RIAA curve, the microphone's presence filter, the crossover inside a loudspeaker. None of them add information. All of them correct a known, measurable distortion.

The next time a piezo pickup spits out its brittle, nasal take on a good guitar, remember what the wood would have done. Then go remove the difference.