How to Tune a Car Audio System: Method and Measurement
Tuning is the set-up you do on a DSP (digital signal processor, the box that controls each speaker's volume, timing and tone) after the speakers are installed. It decides whether they sound like one system or a pile of separate parts. Peter Steinbacher of Pssound treats it as the most important part of any system, and teaches a fixed method that uses one microphone, free software and, at the end, your ears.
The short version
In Peter's experience, the tune, not the hardware, holds back well-built systems about 90% of the time.
Tuning corrects what the cabin does to the sound; it cannot fix a poor install or a bad speaker location.
His method is a fixed sequence of about 60–70 steps: each driver alone, left matched to right, then pairs, then everything.
One microphone and free software are enough, given a consistent technique and a mic you have checked.
A curve can look right and still sound wrong, so the last adjustments are made by ear.
An automatic tune gets about 30% of what a system can give; a proper tune aims for 90%.
A three-way system with subs takes Peter about two hours, and it needs a recheck after the speakers run in.
Why the tune matters, and what it cannot fix
After about 10 years of competing, Peter reckons the tune is what holds back an otherwise well-built car about 90% of the time (watch). His advice: get the car tuned before you change anything, and ask whether it was tuned at all before you blame the speakers.
His key point is that you tune the cabin, not the speakers. A big peak in a tweeter measurement is the car at your listening seat, not a fault in the tweeter (watch). EQ (turning chosen frequencies up or down) and time alignment (small delays so every speaker's sound reaches you together) can correct that, but they cannot cure a bad location or a poor install.
Peter says peaks are easy to cut, but a deep cancellation (a dip where sound waves cancel each other out) usually cannot be fixed with EQ, so you look for a different speaker location.
He finds that speakers closer to you get less help from the cabin at low frequencies, so the driver's side shows less bass than the passenger side, and the tune must balance them.
A DSP is not a tune (see our DSP guide). A six-figure car with expensive kit, simply dropped in, sounded worse than a factory system in a Ford Focus; Peter says to pay a tuner with a proven record.
Automatic tuning delivers about 30% of what a system should give, he says, while he wants 90% (watch). He thinks remote tuning can reach a good 75–80%, but probably never the last 20%, so he tunes on site.
Get the car ready before you measure
Peter says setting up a system is not just jumping in with the microphone.
First power-up: listen for turn-on pop and noise, confirm each speaker is in the right place, then play music through basic crossovers (settings that decide which frequencies each speaker plays). There should already be life in the system.
Check polarity (whether all the speakers push out and pull in together). Trainees who skip it find that delaying left and right shows no result (watch). Even Peter's own S-Class once had a mid inverted.
Switch off bass EQs and boosts on amps and head units, and set amp filters to full range where there is a switch, so the DSP does the filtering. One French customer's amp had its bass EQ at +12 dB, which made a hump near 28–30 Hz and buried the tweeter level (watch).
Save the old preset before you touch anything, and save often.
Tune somewhere quiet. Traffic, or even a car playing music nearby, can show up in bass measurements.
Peter's order: each driver alone, then pairs
Peter's tuning is a fixed sequence of about 60–70 steps for an experienced tuner who makes no mistakes. Each step builds on the last, so one wrong step spoils the rest, and nobody can easily rescue someone else's half-finished tune. Below is the common thread; numbers vary by car.
Measure every driver on its own first, with no EQ and only a protective filter, before choosing any crossover (watch). What you find guides the crossover points.
For safe first measurements, put a tweeter behind a 1 kHz high-pass (a filter that cuts everything below it) and a midrange behind a 100 Hz one, both at 24 dB per octave (a steep slope). Play pink noise (a steady test hiss) at a moderate 85–90 dB. The final tweeter high-pass should be at least 2 times its resonant frequency (Fs), ideally 3 times. Measure subs full range, or crossed deliberately high.
In a three-way front (tweeter, midrange, midbass), start with the midrange, because everything ties to it. With only a midbass and a tweeter, start with the midbass.
For each pair, EQ left and right separately until their curves match, level-match them, time-align them, then measure them as a pair and treat them as a pair from then on (watch).
Do tweeters and midbass the same way (he has done either first), then check them with the mids for cancellations, using very fine time-alignment steps. Last come the whole front, the subs and the rear fill.
His statements on timing differ: in August 2020 he started with time alignment, but in May 2020, 2021 and 2025 it came after left and right were EQ-matched. His latest session follows the second order.
Measuring without fooling yourself
Peter says how you measure matters as much as the steps themselves. Measure wrongly and everything after it rests on a false picture. He uses Room EQ Wizard (REW, free measurement software) with a UMIK-1 microphone, and says "the magic is not in the mic" (watch), but in the knowledge.
Move the mic one inch and you get a completely different picture. So measure your midrange three times back to back: if the results differ, your technique is wrong and you should not trust it (watch).
For before/after comparisons, fix the mic in place and keep the measurement level the same.
Set the graph's vertical scale to 5 dB or tighter. Zoom out to 20 or 50 dB and anything looks linear (watch).
Smoothing averages the graph and blurs detail. Peter shows 1/12-octave graphs and tells beginners to use 1/6. He dropped 1/48 in 2020: it makes you chase every tiny peak and dip.
A calibration file does not make a mic trustworthy. When his masterclass compared mics, results were pretty wide, his own included (watch). Mics also drift: he replaced a two-year-old one. One mic can need a different house curve (target response) than another.
Measurement and ears work together
A graph can look right and still sound wrong. One Skoda measured correctly but sounded cold, and a door midbass can measure perfectly yet not kick; Peter puts that down to rigid mounting. In a car, an RTA (real-time analyser, the live graph of loudness by frequency) also cannot separate direct sound from reflections. Peter's rule: "Measurement is not enough. You have to use your ears." (watch)
The last stage is by ear: fine-adjusting imaging (where voices and instruments appear) and spectral balance (the mix of bass, mids and treble). He calls it the easy part, but few people get that far.
Top-end check: push about 10 dB into the system above 8 kHz. If it still does not sparkle, even with your head beside the tweeter, something is wrong (watch).
Cancellation check: play notes from one side only, through all its speakers. An even level is good; a drop means a cancellation. Repeat with windows up, windows open, then doors open.
Finish with music you know really well (our guide to judging a system covers test tracks).
Target curve and worked numbers
Peter says flat is not the goal in the bass. Because of how hearing works, you need more level the lower you go. In May 2020 Peter started the rise around 200 Hz at no more than 3 dB per octave for pure sound quality. In his 2025 session the response is linear above about 250–300 Hz and rises at least 3 dB per octave below that, and 4–5 dB in sports cars (watch). So 3 dB per octave has gone from an upper limit to a lower limit.
Peaks and dips of about 10 dB colour the sound. The front in his SQ in Cars Part 9 video ended within about 1–2 dB.
Fully tuned does not mean 4–5 EQ bands (separate adjustments). He means at least 15–20 per channel on average, with the midrange almost fully used, and 10–15 for a sub (watch).
Example crossovers from his 2025 three-way session, not recipes: midbass to mid at about 330 Hz, mid to tweeter at about 3.2 kHz, front sub up to about 100 Hz, rear sub up to about 53 Hz. Our crossover guide goes deeper.
How long a tune takes, and when to retune
Peter says a proper tune takes longer than people expect. Speed comes from a ready car, software you know and a practised process. Do the baseline once, properly; the final fine tweaking can wait a week or a month.
His rule of thumb is about 2 hours for a three-way front plus front sub, rear sub and rear fill, and at most 2.5 hours if he takes his time (watch). His record is 53 minutes for a three-way plus sub on a Helix DSP in Singapore, explaining every click to two trainees.
Retune once the speakers have run in: bass drivers loosen up and their transient response changes, so timing needs small adjustments (watch). One C43's final tweak came after about 100 hours, and a Mini Clubman's competition tune had drifted a year on (still okay, but a retune brought it back).
Retune after hardware changes too: new tweeters and a new amp needed almost a full retune.
Re-check the next day, especially if the temperature changes. True infinite-baffle subs (a sub using the boot or outside air as its box) are highly affected by it: at one event a car measured about 3 dB down in the bass after the venue cooled, and re-measuring fixed it in about three minutes.
Common mistakes to avoid
Buying new speakers or amplifiers before the current system has had a proper tune.
Trusting an autotune, or the mic supplied with an automated set-up, to do the whole job.
Measuring before you check polarity, wiring and amp bass boosts, or without saving the original preset.
Trusting one measurement: a moved mic, a changed level, a zoomed-out graph or heavy smoothing can mislead you.
Assuming a calibration file makes a mic accurate, or copying someone else's target curve.
Trying to EQ away a deep cancellation, or expecting tuning to rescue a poor install.
Never retuning after run-in or hardware changes.
Learning tuning only from videos. Peter says it takes practice with someone correcting your mistakes, then 10, 20 or even 50 more tunes.
Watch Peter explain it
This is an independent, plain-English summary of Peter Steinbacher's public YouTube videos, written by VA3 Car Studio. It is not affiliated with or endorsed by Pssound. The links jump to the moment in each video where he makes the point.
Part of our series: Car audio lessons from Pssound, start here
Comments