Nothing Here Has a Fixed Sound

The rest of the guide

Sections of the source document that this page did not carry.

Guitars — The Source Engine

A passive, subtractive instrument whose frequency response is set by whatever it is plugged into — and whose controls are a filter stage, not just volume and tone.

Output impedance, and why it isn't a number

Ask about this section

• A pickup is an inductor with resistance, so its impedance rises with frequency — Z = √(R² + (2πfL)²). The DC resistance people quote is only the floor.
• Measured DC resistance across common pickups: Strat single coil 5.8–6.3 kΩ, Tele bridge 6.2–6.6 kΩ, Tele neck 7.6 kΩ, Gibson PAF 8.4 kΩ, P-90 8.6 kΩ, Seymour Duncan JB 16.4 kΩ.
• Inductance: Strat single coil 2.3–3.0 H, PAF humbucker 4.0–5.0 H. Higher inductance means a lower resonant peak, which is the actual mechanism behind "humbuckers are darker" — not the magnets and certainly not the wood.
• The published range for passive output impedance is 5–20 kΩ, against a typical amp or pedal input of ~1 MΩ. That's a deliberate >10:1 bridging ratio, not an impedance match — you are trying to avoid loading the pickup, not transfer maximum power.
• Which is why a mixer line input doesn't work. Sound On Sound's Hugh Robjohns: standard line inputs are 10–50 kΩ, "way too low to allow a guitar's pickups to work properly." That is the entire reason DI boxes exist.

The resonant peak — the number that actually decides how a guitar sounds

Ask about this section

• A pickup is an RLC system peaking at f₀ = 1/(2π√(LC)), and the C in that equation includes everything downstream of it.
• Measured, on a real Strat single coil: 1.58 H, 222 pF self-capacitance, self-resonance 8,500 Hz unloaded.
• Now plug it in. Guitar cable capacitance measures from ~15.8 pF/ft (low-capacitance types) to ~48.8 pF/ft, with ~30 pF/ft a common average. An 18-foot cable adds roughly 540 pF, taking total C to about 762 pF — and the peak falls to about 4.6 kHz. A 35-foot cable takes it to about 3.6 kHz.
• A true-bypass pedal adds roughly 15 pF of jack-to-jack wiring capacitance even when it's switched off. That is the real, measurable component of "tone suck" — not a mystery.
• And where the peak lands is the guitar's character: below 1 kHz reads dull and hollow; 2.0–2.5 kHz is the singing PAF humbucker region; 3.0–5.0 kHz is the bright, metallic Strat and Tele region. Healthy Q is 2–5 — below 2 sounds shallow, above 5 sounds edgy.
• So the honest statement is that a guitar does not have a frequency response until you connect it to something. Change the cable and you have measurably changed the instrument.

Two mechanisms hiding under one folk term

Ask about this section

• "Tone suck" is actually two separate effects, routinely conflated. Capacitive loading shifts the resonant peak down in frequency — that's the cable and pedal-wiring effect above. Resistive loading damps the amplitude of the peak, lowering its Q, when something with unusually low input impedance sits on the pickup.
• A buffer fixes both, and the reason is not magic. High input impedance means it doesn't resistively load the pickup. Low output impedance — a few hundred ohms to about 1 kΩ — means that any downstream RC corner, at 1/(2πRC), moves far above the audible range. The buffer doesn't add treble; it moves the filter out of the way.
• This is also why the 250 kΩ / 500 kΩ pot convention exists. A pot is a resistive load in parallel with the pickup's LC circuit. A lower value damps more treble, which a bright single coil tolerates and a darker humbucker generally doesn't. Flag: no source I found pins this to a specific Hz figure — qualitatively well established, numerically not.

Pickup position — the comb filter, worked

Ask about this section

• A pickup samples the string at one point, so it nulls any harmonic with a node underneath it: V = sin(π·X·F/(L·F_open)), with nulls at F_null = n·L·F_open/X.
• On a Stratocaster (25.5", low E open 82.41 Hz; bridge pickup 1.625", middle 3.875", neck 6.375" from the bridge), the neck pickup's first response peak lands at 165 Hz and its first null at 330 Hz. Bridge and neck sound different because of geometry — the magnets are secondary.
• Blending two pickups adds a cosine term that nulls at odd multiples rather than all of them, which is the actual origin of the Strat's position-2 and position-4 "quack." It's a different notch pattern, not a volume blend.

Scale, tension and the myth the measurement killed

Ask about this section

• T = UW × (2LF)² / 386.4. A D'Addario EXL110 set at 25.5" totals 102.5 lb; since T ∝ L², the same set at Gibson's 24.75" totals about 96.6 lb — 5.8% less. That percentage is my calculation from the verified formula and D'Addario's published numbers, not a quoted figure.
• Bending force has been measured. A plain .012 bent a full step at the 8th fret of a 25.5" scale took 1.8 lb; a wound .056 bent a step and a half took 4 lb.
• And the same experiment killed a persistent myth: the force depended only on vibrating length, gauge and target pitch. Changing the string's total length behind the bridge made no measurable difference at all.

Pedals — The Order Engine

A pedal is not an effect, it is a stage in a chain of nonlinearities — and its position determines what it actually does far more than its knobs do.

The order, and the electronic reason for each rule

Ask about this section

• The canonical source is R.G. Keen's "Effects Order," and it is a distortion-centric hierarchy rather than a pedal list: amplitude-altering effects first, then pre-distortion EQ, then distortion, then post-distortion EQ, then short time-based effects, then long ones, then reverb.
• Compressor before drive: evens out pick attack before the clipper, so the distortion works on a uniform waveform — smoother, more sustained. The cost is noise: the compressor's makeup gain lifts the noise floor, and the drive then amplifies and clips that too.
• Compressor after drive: distortion is already a brutal compressor — hard clipping pins the waveform at a fixed voltage — so a compressor downstream "adds minimal effect" dynamically. It becomes a sustain leveller and clean boost rather than a dynamics processor.
• Time-based effects after drive, and Keen gives the mechanism: phasers and delays create notches in the frequency response, and distortion placed after them "fill[s] in the response notches the time delay created." Put delay first and every repeat gets re-distorted — repeats become progressively muddy. Put delay last and it repeats an already-complete waveform, so the repeats stay distinct.
• Wah before drive is the clearest case in the whole guide. A wah is a swept bandpass; boosting one narrow band before a clipper means "the frequencies in the boosted range will be distorted most." That is literally frequency-selective clipping, and it also "cuts a lot of the harsh sounding intermodulation distortion" because the whole spectrum isn't hitting the clipper at once. Wah after drive just sweeps a filter over a spectrum whose harmonic content is already fixed — more predictable, less alive.
• Noise gates, and why serious rigs use two. A gate before the gain stage is largely useless on its own because "there's almost no noise to catch yet — the hiss and hum are mostly generated by the gain stage" downstream. A gate after the gain stage can act on the real noise, but distortion's compression means a sustaining note and the noise floor sit at nearly the same level, causing chatter. The documented two-gate solution: one gate reading the clean input for detection, a second placed after the gain block to do the muting — "detection and muting must happen on different signals for accurate tracking."
• And the disagreement is real, from named professionals. Bob Bradshaw puts distortion early, then filters and modulation. Pete Cornish puts the compressor absolutely first and warns a volume pedal before it "will defeat any compression." Dave Friedman calls order "a very subjective thing" and cites a working rig with wah last and delay first. Strymon's own guide says outright: "There is no wrong way to connect your effects." Treat the standard chain as a strong prior, not a law.

The Fuzz Face problem — the one circuit that genuinely hears what's in front of it

• Input impedance, simulated at component level: the first stage's rπ works out around 8 kΩ, but the feedback network drops it further — "in practice it will lower the input impedance to ~5 kΩ," varying between 5.2 kΩ and 8.4 kΩ with the Fuzz control's position. That is meaningfully lower than the ~10 kΩ figure usually quoted.
• It has no clipping diodes at all. Distortion is pure transistor saturation and cutoff. And it's deliberately mis-biased: the first transistor's collector sits near −1.6 V rather than the theoretically ideal −4.5 V on a 9 V supply, so small signals clip asymmetrically on one side first while larger signals push both half-cycles into saturation — "giving a nice touch sensitivity."
• Which is why a buffer in front kills it. ElectroSmash states it plainly: "the germanium transistor needs to see the inductance/impedance from the guitar pickups. If they see a buffer at the input, they tend to sound awful." The transistor's operating point is partly set by what's loading it; a buffer presents a fixed, engineered, low-impedance source the circuit was never voiced around.
• And it is why the guitar's volume knob cleans up a fuzz but not a Tube Screamer. The fuzz's input impedance is low enough to be comparable to the guitar's own volume pot resistance, so the pot isn't a clean attenuator — it interacts with the operating point. A Tube Screamer's input impedance is ~447 kΩ and a RAT's ~494 kΩ, which swamps that interaction entirely; and both have large fixed gain ahead of a fixed diode threshold (TS 12–118×, RAT up to 2305×), so even a heavily attenuated signal is still multiplied past the clipping point. Flag: the specific volume-pot-interaction chain is engineering inference from verified component values, not a single cited source.
• Important limit on the thesis, and it came from an expert pushing back rather than agreeing. A Fractal forum contributor noted that "just about any pedal that has a first gain stage that does not clip... acts as an input buffer that negates the source impedance effects." So impedance sensitivity is a property of a specific circuit category — direct-coupled transistor stages with no input buffering — not of pedals in general. Most op-amp drives genuinely do not care what's in front of them.

Clipping topologies — and where the simple rule breaks

Ask about this section

• Tube Screamer — soft clipping. Silicon diodes (~1 V forward voltage) sit inside the op-amp's feedback loop, so the stage's gain is actively reduced as the signal approaches the threshold. The waveform rounds rather than squaring. Input impedance ~447 kΩ, output ~1.2 kΩ, gain 12× to 118×.
• ProCo RAT — hard clipping. Diodes sit after the amplification stage, to ground rather than in the feedback path, so the LM308 swings toward its full gain — up to 2305×, 67 dB — before slamming into a fixed ceiling. Much squarer wave. Its tone control sweeps a low-pass from 475 Hz to 32 kHz. And the LM308's slew rate of 0.3 V/µs — about 40× slower than a TL071 — imposes a practical high-frequency ceiling near 5.3 kHz, which is itself part of the RAT's character.
• Big Muff — where the tidy rule fails. Its diodes are in the feedback loop of two cascaded transistor stages (~0.6 V each), which by the usual taxonomy makes it "soft clipping." It does not sound soft, because cascading two clipping stages matters as much as the topology of either one. Measured stage gains: 23 dB and 25 dB. Its tone stack produces a genuine notch centred at 1 kHz with about 6.5 dB of extra loss at the notch on top of ~7 dB overall stage loss, and the distortion bandwidth rolls off outside roughly 90 Hz–1.2 kHz at 60 dB/decade.
• Fuzz Face — no diodes. Clipping is the transistor's own operating point, which moves with signal amplitude and source impedance. No fixed threshold anywhere in the circuit.
• The Tube Screamer's two 720 Hz filters are worth stating precisely because they do opposite jobs. A high-pass at 720 Hz (R4 4.7 kΩ with C3) sits before the clipping so "harmonics above 720 Hz get the full gain of the distortion stage" — pre-emphasis, deciding what distorts. A separate passive low-pass at 723.4 Hz (R7 1 kΩ, C5 0.22 µF) in the output stage cuts highs after clipping. Same corner frequency, opposite side of the nonlinearity.

Compressors, and true bypass versus buffered

Ask about this section

• Ratio is arithmetic, not vibe: at 4:1, "for every 4 dB of signal that passes above threshold, the compressor allows only 1 dB through." Attack bands roughly as 0–1 ms fast (squashes the pick transient, adds perceived sustain), 1–10 ms moderate, 10–30 ms slow (lets the transient through, preserves snap). Release is what actually extends perceived sustain, by holding makeup gain into the note's decay.
• Three topologies, with real examples. OTA — MXR Dyna Comp / Ross, built on the CA3080, input impedance ~1 MΩ, timing set by C8 (10 µF) and R13 (150 kΩ). FET — Origin Effects Cali76, modelled on the studio 1176; "lightning fast... usually associated with the pumping kind of sound." Optical — an LDR whose physics does the work: "it can be turned on (gain reduced) quickly, but turns off relatively slowly. This has a natural musical property." The lag is the vintage character, not a design choice layered on top.
• True bypass versus buffered is a genuine trade, not a hierarchy. True bypass is a mechanical wire through, but a bare switch contact can pop audibly, especially into high gain. Buffered bypass removes the pop but puts active circuitry permanently in the path, and cheap implementations do degrade the signal.
• And the famous threshold is folklore. The commonly repeated "about 5 pedals before you need a buffer" comes with its own source immediately conceding "there's no magic number that fits all setups." There is no measured cliff — the real quantities are the pF/ft and pF/connector figures, and those depend entirely on your cable lengths.
• Where a buffer goes: early — right after the guitar, or after the first impedance-sensitive pedal if you want that one to see raw pickup impedance — and before any long run. And always when splitting to multiple destinations, because two amps or an amp plus a DI load each other in parallel without one.

Amps — The Load Engine

The last nonlinearity, plus the only mechanical element in the chain — and the only one whose behaviour changes with how loud you are playing it, moment to moment.

Preamp versus power amp — what actually differs

Ask about this section

• The preamp is voltage gain. Small swings amplified against the tube's grid and plate curves. Low current, high voltage, high output impedance, and the tone stack lives here — which is why preamp distortion is EQ-shapeable.
• The power amp is current delivery. A phase inverter splits the signal into two opposite-phase halves, the output tubes drive the transformer, the transformer matches impedance to the speaker. The tone stack is upstream, so there is very little EQ control over what happens here.
• The described difference: preamp distortion is "more compressed and smoother, with higher sustain at lower volume levels"; power amp distortion is "richer, more dynamic, punchier, and less compressed" — but needs volume. Hughes & Kettner frame it as dynamics: preamp compression "flattens out the dynamics," power amp compression is "very musical, with notes blooming."
• Even the phase inverter has a signature before the output tubes clip. A long-tailed pair is "the most efficient, linear, and balanced-sounding"; older paraphase designs "create smooth, early distortion with lots of compression"; split-load designs sound "sweeter and richer."
• One contested claim worth flagging rather than repeating. The popular "asymmetric clipping equals even harmonics, power amp equals even, preamp equals odd" story is disputed — one technical source argues that analysis only holds for pure sine waves, not complex guitar signals, and presents oscilloscope data on a Marshall JMP 1987 power amp showing hard clipping, not soft, arguing "the smooth tone comes from the topology of the amplifier." Informed contrarian opinion, not consensus, and not peer-reviewed.

Sag — real electronics, with numbers

Ask about this section

• Three physical causes, per amp designer Randall Aiken: rectifier internal resistance (tube rectifiers only), transformer winding resistance, and filter capacitor size.
• A worked example for the transformer term: idle current 70 mA rising to 170 mA at full power through a 200 Ω secondary gives ΔI = 100 mA, so V = 100 mA × 200 Ω = 20 V of plate-voltage drop.
• And for rectifiers, a measured figure: Steven Fryette states tube rectifiers can produce "as much as 50 volts" of drop under full load, versus a much smaller drop from solid-state diodes. Solid-state sag can be emulated by adding series resistance, "typically around 100 ohms or so."
• Sag isn't even static across a set — Fryette notes it compounds thermally: "As your power transformer heats up, the voltage it produces decreases."
• Genuine Class A amps largely avoid this, because "the current draw at full power is the same as the current draw at idle" — no current swing, no IR-drop swing.
• And too much sag is a defect, not a feature: poorly filtered supplies produce "ghost notes" — non-harmonic tones riding on the fundamental from 120 Hz ripple modulation.

Negative feedback — and why presence isn't a treble control

Ask about this section

• A small amount of output-transformer secondary signal is fed back to the phase inverter input. Effects: it "flattens and extends the frequency response," "reduces distortion generated in the stages encompassed by the feedback loop," and "reduces the effective output impedance," raising damping factor. Typical amount in guitar amps: 6–10 dB.
• Presence and Depth are frequency-selective feedback controls, not EQ. Presence works by reducing negative feedback at high frequencies, which effectively boosts treble; Depth or Resonance does the same at low frequencies.
• Which yields a hard limit almost nobody knows: the maximum boost available equals the amount of feedback in the loop. An amp with 6 dB of feedback can offer at most about 6 dB of presence boost. An amp with no global feedback loop cannot have a presence control that works the same way at all.
• Feedback also changes how the amp clips — feedback amps hold distortion low until the output stage clips and then break up abruptly, where a no-feedback design transitions gradually.

Why the same amp sounds different at different volumes

Ask about this section

• Your ear changes first. Equal-loudness contours mean identical EQ settings deliver different perceived balance at different SPLs. Concretely: at a quiet 50 phon, a 200 Hz tone must be about 9 dB louder than a 100 Hz tone to sound equally loud; at 120 phon that gap shrinks to about 2 dB.
• And a genuinely underappreciated twist: the 2003 international remeasurement found the 1956 Robinson-Dadson curves differ from modern data by 10–15 dB, especially at low frequency, "for reasons not explained" — and that the original 1933 Fletcher-Munson data agrees better with modern measurements than the 1956 revision that supposedly corrected it.
• Wattage does far less than people assume. +3 dB doubles acoustic power; +6 dB doubles sound pressure; roughly +10 dB doubles perceived loudness — and requires ten times the amplifier power. Going 15 W to 150 W buys you about a perceived doubling. The real difference between a 15 W and a 100 W amp is not volume, it's how loud it stays clean.
• The speaker changes too, in two separate ways. Mechanically: guitar drivers are deliberately non-linear with "almost no pistonic band" above ~500 Hz, and at higher excursion you get "cone cry" and "edge yowl" — though the breakup character itself is present even at microvolt levels. Thermally: heating a copper voice coil from 20 °C to 200 °C raises its DC resistance by about 72% and drops sensitivity by 4.7 dB. That's power compression, and it is measured, not folklore.
• Attenuators preserve the electrical interaction and not the mechanical one. A reactive load presents speaker-like varying impedance so the output stage behaves normally; a resistive load presents a flat impedance and, as attenuation increases, makes tone "more compressed and dark." But even a good reactive load cannot reproduce excursion-dependent breakup or thermal compression, because the cone isn't actually moving that much air. Quilter's own caveat: "simply adding reactance to a load does not result in a comprehensive guitar speaker simulation."

Impedance — the load is a curve, not a number

Ask about this section

• This is the best-documented fact in the section. A measured Eminence 8 Ω driver reads exactly 8 Ω at three frequencies — 50 Hz, 200 Hz and 500 Hz. It peaks near 70 Ω just above 100 Hz at mechanical resonance, and climbs to about 50 Ω by 20 kHz from voice-coil inductance. Nominal impedance is a mean, not a value.
• Even at DC it isn't nominal. A Celestion Vintage 30 rated 8 Ω measures 7.3 Ω DC; the 16 Ω version measures 12.9 Ω. Fs = 75 Hz, sensitivity 100 dB, rated range 70 Hz–5 kHz.
• So the output transformer and the feedback loop are interacting with a wildly variable load on every note you play. That is the thesis at hard-numbers level: there is no fixed "amp tone" because there is no fixed load.
• Why no load destroys a tube amp: the output stage behaves close to a constant-current source into the transformer primary, so as load impedance rises so does the voltage the transformer generates — remove it entirely and the voltage can go destructive. The refined mechanism is that clipping transients create fast current discontinuities, and with no secondary load to absorb them the ringing reflects back into the primary, risking flashover across insulation or sockets. Flag: that specific mechanism is well explained by practitioners but not pinned to one authoritative published source.
• Solid-state amps are usually fine because they're voltage sources with no output transformer storing inductive energy — but treat that as a strong generalisation rather than an absolute.
• The safety rule that appears consistently: mismatches within double or half the rated impedance are generally considered safe. Beyond that, risk rises sharply — too high a load stresses transformer and tubes with excess voltage, too low pulls excess current and overheats them.

Pedal platforms — and where the sourcing gets thin

Ask about this section

• Headroom is the term with real backing: "how cleanly an amplifier will respond to an increase in signal," and it operates independently at the preamp and the power amp. A pedal platform needs both — preamp headroom so a hot pedal doesn't compress the front end prematurely, and power-amp headroom so the amp doesn't add its own compression on top.
• A real interaction worth knowing: power-amp distortion "can smush reverb trails and delay repeats into a much less articulate, smeared-together howl." Which is a large part of the argument for putting time-based effects in the loop.
• Three things I'd rather flag than launder. The commonly cited 30 W floor for a pedal platform is guitar-press consensus, not a physics constant. The claim that Fender-style circuits make good platforms because of a "scooped midrange" is received wisdom, not circuit analysis. And I could not source a rigorous account of tone-stack placement and its consequences for pedal platforms — it's well-established amp-design knowledge that passive stacks sit between gain stages and therefore interact with them, but I don't have a citation strong enough to state specifics as verified.

2. FILTER FIRST, CLIP SECOND

Pedals ▸ Guitars — A guitar's tone knob can only subtract. Put that in front of a clipper and subtraction becomes selection.

The variable that moves: The passive tone control moves into the chain — as a decision about what distorts rather than what's heard.

Hinge (what already agrees): A wah is a guitar tone control made sweepable, and both are passive filters that can only remove. Neither has gain. The guitar's tone pot is a fixed low-pass with a variable corner; a wah is a bandpass with a swept centre. Same family, same limitation, same position in the chain.

Friction (what fights): But a modern board is built to be immune to exactly this. Buffered bypass, op-amp input stages, fixed impedances — every design decision on a contemporary pedalboard exists to make each box behave identically regardless of what surrounds it. That immunity is the direct opposite of the guitar's defining characteristic, and it means a filter dropped anywhere in a buffered chain simply filters, rather than interacting.

Resolution: Use position rather than impedance to get the interaction back. Put every filter before the clipping stage, and the passive subtraction becomes frequency-selective clipping: "the frequencies in the boosted range will be distorted most," while everything outside stays comparatively clean. That is the one place where a filter's effect is qualitative rather than merely tonal. And it has a second documented benefit — it "cuts a lot of the harsh sounding intermodulation distortion," because the whole spectrum isn't hitting the clipper simultaneously. Then put a second filter after the clipper for ordinary tone shaping, which is exactly what a Tube Screamer does internally: a high-pass at 720 Hz before the diodes, a low-pass at 723 Hz after them.

Recipe

Ask about this section

• Chassis: pedalboard, standard order, everything buffered except where specified
• Every filter before the drive: wah, EQ, any pre-emphasis. This is where they change what gets distorted
• The drive stage in the middle, with its gain set so the pre-filter's boosted band clips and the rest doesn't
• A second filter after the drive for conventional tone shaping — same corner frequency is fine, opposite job
• Compressor after the drive, not before, so it acts as a sustain leveller and doesn't lift the noise floor into the clipper
• Time-based effects last, so repeats stay articulate rather than being re-distorted into mush
• The audible test: sweeping the pre-filter should change the character of the distortion, not just its brightness. If it only changes brightness, it's in the wrong place
• Avoid: putting modulation or delay before the drive unless the muddied, notch-filled result is what you want.

3. FOUR CABLES

Amps ▸ Pedals — The amp is already a pedal chain. It just has the boxes soldered in a fixed order.

The variable that moves: Chain order moves into the amplifier — and discovers the amp has two clipping stages in different places.

Hinge (what already agrees): An amp is a chain of stages with the same logic as a board. Preamp gain, tone stack, phase inverter, power tubes, output transformer, speaker. And the passive tone stack sits between gain stages — so the amp is already doing merge 2 internally, filtering before one clipper and after another.

Friction (what fights): But the amp distorts in two places and only one of them is before the effects loop. A delay in front of the amp gets distorted by the preamp and the power amp; a delay in the loop is distorted only by the power amp. And that matters, because power-amp distortion "can smush reverb trails and delay repeats into a much less articulate, smeared-together howl." Meanwhile a noise gate placed at the input is nearly useless — "the hiss and hum are mostly generated by the gain stage" downstream of it.

Resolution: Four-cable method, and split the gate in two. Drives and filters go in front of the preamp, where they belong — pre-clip. Time-based and modulation effects go in the loop, after the preamp's gain, so their tails stay articulate. And the gate becomes two devices doing two jobs: one reading the clean input signal for detection, one after the gain block doing the muting, because "detection and muting must happen on different signals for accurate tracking." That's the documented two-gate approach, and it's why serious high-gain rigs carry two units that look redundant and aren't.

Recipe

Ask about this section

• Chassis: valve amp with a series effects loop, four-cable method
• In front of the preamp: filter, wah, drives, boost — everything whose job is to decide what gets clipped
• In the loop: delay, reverb, modulation — everything whose job is to process an already-shaped signal
• Gate one reads the clean guitar signal at the input, for detection only
• Gate two sits after the gain block and does the actual muting
• Set the loop level carefully: an 18 V pedal in a loop can clip a modeller's or a return stage's input — a real gain-staging issue, not folklore
• Test: a delay repeat should decay cleanly rather than fizzing or smearing. If it smears, it's in front of the preamp and should be in the loop
• Avoid: a single gate at the input on a high-gain rig. It is muting a signal that has no noise in it yet.

5. THE PICKUP IS A GAIN CONTROL

Guitars ▸ Amps — A hot humbucker and a Strat single coil into the same amp are two different amps.

The variable that moves: The amplifier's gain staging moves back into the instrument — where it was always partly decided anyway.

Hinge (what already agrees): The 1 MΩ input impedance of every guitar amp exists for one reason: to accommodate a source whose impedance is 5–20 kΩ and rises with frequency. That's a deliberate >10:1 bridging ratio. The amp was designed around the instrument's electrical characteristics, which is why a 10–50 kΩ mixer line input, per Sound On Sound, is "way too low to allow a guitar's pickups to work properly."

Friction (what fights): But the two ends have wildly different output levels and the amp has a fixed clipping threshold. A Strat single coil measures 5.8–6.3 kΩ DC with 2.3–3.0 H; a Seymour Duncan JB measures 16.4 kΩ with a PAF-class 4.0–5.0 H. Into the same amp at the same settings, one arrives well below the front end's clipping point and the other arrives well above it. The amp's "amount of gain" is not a property of the amp.

Resolution: Choose the pickup as the first gain stage and set the amp for the pickup, not the other way round. A low-output single coil into an amp set past its clipping threshold gives dynamic, touch-sensitive breakup where playing softly is genuinely clean — because the signal is crossing the threshold, not sitting above it. A high-output humbucker into the same amp is permanently past it, giving consistent saturation and no clean. Both are legitimate; they are just not the same instrument. And the resonant peak decides the character of that clipping: a 3–5 kHz single-coil peak feeds bright content into the clipper and produces an aggressive, cutting distortion; a 2.0–2.5 kHz humbucker peak produces the thick singing one.

Recipe

Ask about this section

• Chassis: valve amp set just past its clipping threshold, so the instrument decides which side of it you're on
• Low-output single coil for dynamic breakup — soft playing genuinely clean, hard playing genuinely distorted, no setting changes
• High-output humbucker for consistent saturation with no clean available. Pick the one you want; you cannot have both from one setting
• The resonant peak is the character control: 3–5 kHz feeds bright content into the clipper (aggressive, cutting), 2.0–2.5 kHz feeds warmer content (thick, singing)
• Volume knob as the fine gain control, because it moves the signal across the amp's fixed threshold
• Cable length matters more here than anywhere else — it moves the peak, which moves what gets clipped
• Do not compensate at the amp. Turning the amp down to tame a hot pickup throws away the interaction this merge is about
• Avoid: a buffer or boost in front unless you want the instrument's output level to stop mattering — which is precisely what it does.

6. TWO RESONANCES, ONE CHAIN

Amps ▸ Guitars — Both ends of the chain are resonant systems. Almost nobody chooses them as a pair.

The variable that moves: The load's resonance moves back to meet the source's — and they turn out to be the same kind of problem.

Hinge (what already agrees): Neither end of the signal chain is flat, and both are resonant for the same reason: inductance and capacitance in a real physical device. A loaded pickup peaks somewhere between 2.0 and 5.0 kHz. A speaker peaks near 70 Ω at just above 100 Hz and climbs back to about 50 Ω by 20 kHz. Neither is a designer's choice — both are consequences of the physics.

Friction (what fights): But they're specified in different units and chosen independently. The pickup's resonance is an electrical peak in the signal; the speaker's is an impedance peak that the amp reacts against, plus a mechanical rolloff at 5 kHz where the cone stops keeping up. One is voltage, one is impedance, and nothing in guitar culture puts them on the same page — so people pair a bright pickup with a bright speaker and wonder why it's shrill.

Resolution: Choose them as a pair and put the pickup's peak inside the speaker's passband. A speaker rated to 5 kHz means a 3.0–5.0 kHz single-coil peak lands right at the edge of what will be reproduced — present and cutting, and above it the cabinet is removing everything anyway, which is why distortion through a real cab never sounds like distortion through a desk. A 2.0–2.5 kHz humbucker peak sits comfortably inside the passband and reads as thick and singing, and generally wants a brighter speaker to compensate. And the speaker's impedance peak at resonance is where the amp's damping is weakest and the low end gets loosest — so the same pairing decision sets the tightness of the bottom end too. One choice, both ends.

Recipe

Ask about this section

• Chassis: amp and cabinet chosen together with the guitar, not separately
• Bright pickup (3.0–5.0 kHz peak): pairs with a darker, earlier-breaking speaker, because the peak already sits at the edge of the passband
• Warm pickup (2.0–2.5 kHz peak): pairs with a brighter speaker, because the peak sits well inside it and needs help to cut
• Cable length is the fine adjustment — it moves the pickup's peak down by roughly 4 kHz across a long run, which is a bigger change than most tone controls
• The speaker's impedance peak near 100 Hz is where damping is weakest. Expect the low end to be loosest and most resonant there, and voice the amp's low end accordingly
• Nothing above 5 kHz survives, so stop trying to fix fizz with EQ — the cabinet already removed it, and if you can hear it you are listening to a direct signal
• Presence control has a hard ceiling equal to the amp's feedback depth, typically 6–10 dB. If it isn't enough, the fix is the speaker, not the knob
• Avoid: choosing pickup and speaker independently and correcting with EQ afterwards. You are EQ-ing two resonances that could have been chosen to complement each other.

THE WHOLE CHAIN IS ONE CIRCUIT

All three at once, treated as a single electrical system rather than three products.

Three things that have no sound of their own, wired into one thing that does. The guitar sets what enters the system — a resonant peak between 2 and 5 kHz whose exact position is decided by the cable and the pedal wiring you plug it into, and a passive tone control that can only subtract. The pedals decide what gets clipped — every filter before the drive, so subtraction becomes selection; every time-based effect after it, so repeats stay articulate; the compressor after rather than before. The amp decides what survives — a power stage whose class, sag and compression all change with how hard you play, into a speaker that is 8 Ω at exactly three frequencies and removes everything above 5 kHz regardless of what you sent it. And the thing that makes it one circuit rather than three is that each stage's behaviour is set by its neighbour's impedance: the guitar by the cable, the fuzz by the guitar, the output tubes by the speaker. Change any one and you have changed all of them.

Recipe

Ask about this section

• Source: one impedance-sensitive drive first with nothing buffered in front, so the guitar's volume and tone controls remain expressive rather than decorative
• Buffer immediately after it, so cable capacitance stops moving the resonant peak for everything downstream
• Filters before the drive: wah and pre-EQ decide what distorts, not what's heard
• Drive stage in the middle, gain set so the pre-filter's boosted band clips and the rest doesn't
• Compressor after the drive, fast attack, release timed to the note decay — sustain leveller, not dynamics processor
• Time-based effects in the amp's loop, not in front, so repeats aren't re-distorted into a smear
• Two gates: one on the clean input for detection, one after the gain block for muting
• Amp set just past its clipping threshold, so the pickup and the picking hand decide which side of it you're on
• Pickup and speaker chosen as a pair: put the pickup's resonant peak inside the cabinet's passband and let the 5 kHz rolloff remove the fizz for free
• The rule that holds it together: nothing in the chain is set in isolation. Every stage is voiced against the impedance of the one next to it

Stereo pedalboard rigs — what actually works

Two things decide whether a stereo rig works: which effects are genuinely stereo, and whether the thing at the end accepts a line-level stereo input. The second one is where most rigs fail, and the marketing is not reliable about it.

• Only some effects are meaningfully stereo. Modulation (chorus, flanger, phaser — usually two LFO-driven paths offset in phase) and time-based effects (ping-pong and true-stereo delay, algorithmic reverb) genuinely produce different left and right signals. Distortion, compression and most gain stages are inherently mono processes even inside a box labelled stereo.
• "True stereo" and "ping-pong" are not the same thing and both get marketed as stereo. True stereo runs independent, decorrelated processing on each side; ping-pong takes one signal and bounces it between two outputs. Audibly different.
• Phase cancellation is the one real physical risk, and it has a documented worst case. Modulation effects deliberately produce out-of-phase left and right content — that's what creates the width. Summed to mono, "summing a stereo signal can sound like a reduction in the effect level," and a stereo tremolo running 180° out of phase "can completely erase the effect and sound like it's in bypass."
• Chase Bliss's Joel Korte states the room problem plainly: "unless one is in the sweet spot, the effect... may cancel out in the air." What sounds enormous in headphones can partially null in a room depending on where the listener stands and where the mics are.
• The two practical mitigations: keep modulation in mono before the split to stereo delay and reverb; and prefer a "split sum" (duplicate one channel to both outputs) over a naive L+R sum when collapsing, which avoids the cancellation entirely for modulation while a true sum stays fine for delay and reverb.
• Wet/dry/wet solves both problems at once. Dry signal to a centre amp for punch and mono reliability; effects split to two amps panned hard left and right. You keep a strong mono-compatible centre and still get the spatial content — and you stop the whole rig from depending on the audience standing in one spot.
• Two identical amps versus two deliberately different ones is a creative choice with no technical consensus. Korte's own wet/dry rig used mismatched amps — a high-headroom solid-state amp plus "a gnarly-little-homemade tube amp."