Car Audio Explained, Part 15: Car Audio Myths He Debunks (2 of 2) (Andy Wehmeyer's Lessons)

4 days ago

26 min read

This is Part 15 of 19 in the series Andy Wehmeyer on Car Audio, Explained for Beginners. Start here · Download the full guide as a PDF

Unofficial guide. Andy Wehmeyer's points are paraphrased from his public forum posts, blog, tuning guides, videos and podcasts, with a link to each source; the plain-English explanations were written for this guide. Not written, reviewed or endorsed by Andy Wehmeyer or Audiofrog.

5.1 #39 · Myth: tweeter-less "widebander" dash mids are purer → All cars sound better with tweeters.

On axis widebanders give "clinical" pinpoint images with no space and off axis they sound dull; they work best in the dash aimed at the glass, as carmakers use them. Sources: DIYMA 2015-08-23 · DIYMA 2021-12-02

In plain English: A "widebander" (wideband or full-range driver) is a small speaker meant to cover almost the whole range of sound by itself, with no separate tweeter (the small speaker that plays the high notes). Many people believe a small widebander in the dashboard, with no tweeter, sounds "purer". He disagrees: all cars sound better with tweeters. Pointed straight at you ("on axis"), a widebander gives a "clinical" sound: each instrument seems to come from a tiny, exact spot, with no sense of space around it. Pointed away from you ("off axis"), it sounds dull, because the high notes fade off to the side. In his view widebanders work best the way carmakers use them: in the dash, aimed at the windshield glass.

5.1 #40 · Myth: a point-source or coaxial driver is the ideal → A full-range point source is impossible, and a typical coaxial has no single "phase centre".

A tweeter over a moving cone suffers intermodulation; a coax's real benefit is treble in the off-axis sound, and exact acoustic-centre alignment is "garbage" when heads move. Sources: OFCAP 19 @1:32:04 · DIYMA 2018-04-02

In plain English: A "point source" is a speaker whose whole sound comes from one single point. A "coaxial" speaker has a tweeter mounted in the middle of a bigger woofer, so both share one axis. Many people think that is the ideal design. He says a point source that plays the full range can't actually be built, and a typical coaxial doesn't have one single "phase centre" (one spot that all its sound seems to start from). Also, a tweeter sitting over a moving cone suffers "intermodulation": the cone's back-and-forth motion mixes with the treble and adds distortion. A coax's real benefit, he says, is that it puts treble into the off-axis sound, the sound that spreads sideways and bounces around the car. Lining up the exact acoustic centres is "garbage" to him, because real listeners move their heads.

5.1 #41 · Myth: special tweeters (19 mm, ring radiators, "designed for off-axis") solve dispersion → Small gains where music has little energy.

Same-diameter tweeters disperse alike; a smaller tweeter has a smaller coil in a market that gives speakers 3× their rating (he made 19 mm car tweeters at his previous company and expects big warranty costs). [RARE] Sources: DIYMA 2022-02-28 · DIYMA 2014-04-02

In plain English: "Dispersion" is how widely a speaker spreads its sound. Tweeters spread the highest notes less and less widely, so people look for special tweeters to fix that: very small 19 mm (about three-quarters of an inch) ones, "ring radiators" (a tweeter with a ring-shaped diaphragm), or ones marketed as "designed for off-axis". He says these give only small gains, and only in the very top notes, where music carries little energy. Tweeters of the same diameter spread sound alike. A smaller tweeter also has a smaller voice coil (the wire coil that drives it), so it handles less power. That matters in a market where people give speakers 3× their power rating. He made 19 mm car tweeters at his previous company, and he expects such tweeters would bring big warranty costs.

5.1 #42 · Myth: high-end SQ speakers are delicate → "Not wedding cakes".

Old high-end car speakers were untested home drivers; everyone plays loud on a Friday night, and the GB line was built for about 450 W. Sources: CMA 2023 @14:01 · OFCAP 19 @1:34:01

In plain English: "SQ" means sound quality. Many people believe expensive sound-quality speakers are fragile showpieces that must not be played loud. He says good high-end speakers are "not wedding cakes". He points out that older high-end car speakers were really home-audio drivers, never tested for life in a car. And everyone plays music loud sometimes, like on a Friday night. His company Audiofrog's high-end "GB" speaker line was built for about 450 W.

5.1 #43 · Myth: a fat, higher-Q midbass is better bass → "That higher Q IS distortion."

Door and kick locations are corners that already add about 6 dB, which is why the GB60 was voiced flat; a low-Q driver sounds thin on a board but full in the car. Sources: DIYMA 2011-08-27 · DIYMA 2017-08-17

In plain English: A "midbass" is the speaker, usually in the door, that plays the upper bass. "Q" is a number that describes how strongly a speaker resonates near its natural frequency. A higher-Q speaker has a hump of extra output there, which sounds "fat" and full. Many people take that fatness as better bass. He says that hump is the speaker adding something that isn't in the music, so it is distortion. In a car, the door and the kick panel (the panel by your feet) act like corners of a room, and corners already boost bass by about 6 dB (a clearly audible lift). That is why Audiofrog designed its GB60 midbass with a flat response. A low-Q speaker can sound thin on a store display board, yet sound full once it is in the car.

5.1 #44 · Myth: dual-voice-coil wiring changes sensitivity or Q → It changes only the load impedance;

Q depends on (BL)²/Re and both terms change together. [RARE] Source: DIYMA 2022-03-23

In plain English: Some subwoofers have two voice coils ("dual voice coil"). You can wire the two coils one after the other (series) or side by side (parallel). That changes the "impedance", the electrical load the amplifier sees, measured in ohms. Many people believe it also changes the speaker's sensitivity (how loud it plays from a given power) or its Q (how strongly it resonates). He says it changes only the impedance. Q depends on (BL)²/Re. BL is the strength of the speaker's motor, and Re is the coil's resistance. When you rewire the coils, both of these change in step, so the result, and therefore Q, stays the same.

5.1 #45 · Myth: a "subwoofer that plays midrange" exists → Efficiency, low Fs and small size don't go together

(Hoffman's Iron Law); a light-coned low-Fs driver is fragile. [RARE] Source: DIYMA 2020-08-24

In plain English: Some people hope for one speaker that plays deep sub-bass and also clean midrange. He says that mostly doesn't exist. "Fs" is a speaker's resonance frequency: the low note at which it naturally vibrates most easily. A low Fs helps a speaker play deep bass. "Efficiency" is how much sound you get per watt. "Hoffman's Iron Law" is a long-known rule of speaker design: you cannot have high efficiency, deep bass and a small size all at once. He adds that a speaker built with a light cone (good for midrange) and a low Fs (good for deep bass) ends up fragile.

5.1 #46 · Myth: speakers are instruments with their own voice, and shorting rings are shields → Listeners prefer accurate speakers; shorting rings cut coil inductance.

Pole vents exist mainly to cut noise behind the cone. Sources: DIYMA 2013-01-04 · DIYMA 2015-06-03

In plain English: This point covers three ideas. First, many people believe a speaker should be like a musical instrument, with its own pleasing "voice". He says listeners prefer accurate speakers, ones that play the recording without adding their own character. Second, a "shorting ring" is a metal ring (often copper) built into a speaker's magnet structure. Some believe it is a shield. He says its job is to cut the voice coil's "inductance", an electrical property of any coil that resists quick changes in current. Third, a "pole vent" is a hole through the centre of the magnet. He says it exists mainly to cut the noise made behind the cone.

5.1 #47 · Myth: the box in the manual is a deliberately "safe", conservative design → Audiofrog's manual boxes are the best boxes.

It designs woofers backwards from the car: in-car target first, then a convenient box size, then a driver that hits the target in that box. [RARE] Source: B1 Absolute Electronix 2023 @8:13

In plain English: A woofer's manual usually suggests a box (enclosure) to build for it. Many people believe these suggested boxes are deliberately cautious, "safe" designs rather than the best ones. He says Audiofrog's manual boxes are the best boxes for its woofers, because the company designs woofers backwards from the car. First it picks the sound it wants inside the car (the "in-car target"). Then it picks a box size that is convenient to fit. Then it builds a speaker that hits that target in that box. So for Audiofrog woofers, the box in the manual is not a compromise: it is the design.

Installation and bass (detail in sections 2 and 4)

5.1 #48 · Myth: foam rings "make midbass" and full door baffles help → Foam stops rattles;

stick rings to the bracket, never the speaker, and trim baffles down to the rain shield. Sources: OFCAP 38 @34:02 · DIYMA 2021-09-21

In plain English: "Foam rings" are foam strips placed around a door speaker to seal it against the door panel, often sold as a way to get more midbass (upper bass). "Door baffles" are foam cups that fit behind the speaker. Many people believe both improve the bass. He says the foam's real job is to stop rattles. Stick the rings to the bracket (the mounting adapter that holds the speaker), never to the speaker itself. And cut the foam baffles down so only the top part remains as a rain shield, keeping water off the speaker, rather than using them fully sealed around it.

5.1 #49 · Myth: deaden everything, and padding absorbs midbass → Deadening is rattle control.

Absorbing ~200 Hz needs about 1.5 ft of material; a 250 Hz peak with no buzz is a reflection to EQ, not a panel to deaden. Sources: OFCAP 38 @1:13:01 · DIYMA 2021-09-07

In plain English: "Deadening" means applying sticky mats or similar material to a car's metal panels to stop them vibrating. Many people believe you should cover everything, and that padding soaks up excess midbass. He says deadening is for controlling rattles. Actually absorbing sound around 200 Hz would take about 1.5 ft (about 46 cm) of material. And if you measure a peak at 250 Hz but hear no buzzing or rattling, it is a reflection: sound bouncing inside the cabin and adding up at that one frequency. You fix that with EQ (an equalizer, which turns down just that frequency), not by deadening a panel.

5.1 #50 · Myth: door midbass needs a sealed or ported door box → Not with a sub and an active high-pass,

which does the box's job; venting a box into the door throws away the port's output. [RARE] Sources: DIYMA 2016-06-17 · DIYMA 2022-01-13

In plain English: Many people believe a door speaker needs its own box built inside the door, either sealed or ported (a "ported" box has an opening, called a port, that adds bass). He says that isn't needed when the system has a subwoofer and an "active high-pass": an electronic filter, set in the processor or amplifier, that stops the deepest bass from reaching the door speaker. That filter does the job the box would do. And if you build a ported box that vents into the door, the port's sound is lost inside the door, so you hear only the speaker. In a ported box you normally hear the speaker and the port together, so this throws away the port's whole benefit.

5.1 #51 · Myth: vented boxes give "slow" bass → Group delay matters only above roughly 50 ms.

"Sloppy" vented bass means a box tuned too high, and vented can lower distortion. Sources: DIYMA 2014-11-01 · DIYMA 2022-01-14

In plain English: A "vented" box is the same as a ported box: a speaker box with an opening that adds bass. Many people believe vented boxes give "slow" or late bass, which is bad for sound quality. "Group delay" is a measure of how much later some frequencies arrive than others. He says it only matters when it is really large, above roughly 50 ms (50 thousandths of a second). When vented bass sounds "sloppy", the box is tuned too high (its port is set for too high a note). He adds that a vented box can actually lower distortion. Elsewhere he explains that in a vented box the woofer itself moves less and distorts much less, though the port's own distortion can be higher.

5.1 #52 · Myth: subs need a Qtc 0.707 box, small boxes are "tighter", and tuning below Fs blows subs → None holds in a car.

The sub is low-passed below Fc, so box Q barely shows; a too-small box gives "one-note" bass; once boxed there is no Fs, only Fc or Fb, and cabin gain changes what you hear, not what the sub does. [RARE] Sources: DIYMA 2021-02-03 · DIYMA 2021-09-04 · DIYMA 2021-05-09

In plain English: This point knocks down three common beliefs about subwoofers. First, "Qtc" is the Q (resonance strength) of a sub in a sealed box, and 0.707 is often called the ideal value. He says that hardly matters in a car, because the sub is "low-passed" (filtered so it plays only the deep notes) below "Fc", the frequency where the boxed sub resonates. So the small effect of box Q barely shows. Second, many think a smaller box gives "tighter" bass. He says a box that is too small gives "one-note" bass instead. Third, many fear that tuning a vented box below the speaker's "Fs" (its own resonance frequency, outside a box) will blow the sub. He says that once a speaker is in a box there is no Fs any more, only Fc (sealed box) or Fb (the tuning frequency of a vented box). "Cabin gain", the extra deep bass a car's small cabin adds, changes what you hear, not what the sub itself is doing.

5.1 #53 · Myth: a down-firing sub is "boundary loading" → It's an accidental bandpass box with unpredictable tuning.

Source: DIYMA 2021-10-29

In plain English: A "down-firing" sub faces the floor. Many people say this uses "boundary loading", where a speaker is placed near a surface such as a floor or wall to gain bass. He says that's not what happens. The gap between the sub and the floor, plus the opening the sound escapes through, acts like a "bandpass" box: a box whose sound comes out only through an opening, which lets through just one band of frequencies. Because nobody designed that gap as a box, its tuning is accidental and you can't predict it.

5.1 #54 · Myth: a midrange needs an enclosure, and the biggest driver that fits the pillar is best → Behind a high-pass the box barely matters, but a big driver in a small pod loses bass and peaks.

He modelled the GB25 with a 200 Hz LR4 high-pass in boxes from about 0.16 L to infinite baffle and saw only slight differences; a pillar pod is an enclosure, so size the driver to it. [RARE] Sources: DIYMA 2017-07-02 · DIYMA 2017-10-01 · DIYMA 2021-11-19

In plain English: A "midrange" is a small speaker for the middle notes, often mounted in a "pod" on the A-pillar (the post beside the windshield). Many people believe it needs a sealed box behind it, and that you should fit the biggest speaker the pillar can hold. He says neither is right. With a "high-pass" filter in front of it (a filter that blocks the low notes), the box barely matters. He modelled Audiofrog's GB25 midrange behind a 200 Hz LR4 high-pass (a common, steep crossover filter type) in boxes from about 0.16 L (roughly 10 cubic inches) up to "infinite baffle" (no box at all). The results were only slightly different. But a pillar pod is itself a box. A big speaker crammed into a small pod loses bass and develops a peak, like a subwoofer in a box that's too small. So choose the speaker to suit the pod's size.

5.1 #55 · Myth: cross door midbass at 40–50 Hz to pull bass forward → "50 Hz isn't midbass"; it's on his worst-mistakes list.

High-pass the midbass about 80 Hz and low-pass the sub about 60 Hz (both LR4) so a 9–12 dB sub lift stays in phase; told mids are crossed at 40 Hz, he "stops listening". Bass "impact" lives near the mid-to-tweeter crossover. [RARE] Sources: OFCAP 19 @1:45:06 · DIYMA 2020-12-30 · DIYMA 2021-11-02

In plain English: The subwoofer usually sits in the back of the car, but listeners want the bass to seem to come from the front. A popular fix is to make the door speakers play very low, down to 40–50 Hz, so they carry more of the bass. He says "50 Hz isn't midbass", and he lists this among the worst mistakes. When someone tells him their door speakers are crossed at 40 Hz, he "stops listening". His recipe: filter the door midbass to play from about 80 Hz up (high-pass), and the sub to play from about 60 Hz down (low-pass), both with LR4 filters (a common steep crossover type). The sub is usually turned up 9–12 dB to sound natural, and this gap between the two filters keeps the sub and the door speakers "in phase" (pushing and pulling together) even with that boost. He also says bass "impact" (the punchy attack of a drum hit) lives near the crossover between the mid and the tweeter, which sits far above the bass notes themselves.

5.1 #56 · Myth: with the windows down bass "waves can't fully develop" → Meaningless;

the cabin simply turns from a leaky sealed box into a vented one. [RARE] Source: DIYMA 2018-03-23

In plain English: Many people explain why bass changes with the windows down by saying the long bass "waves can't fully develop" inside a small car. He calls that meaningless. With the windows up, the cabin acts like a leaky sealed box around the sub. Open a window and the cabin becomes a vented box, like a speaker box with a port. That is the real reason the bass changes.

Crossovers, gain and tuning (detail in section 3)

5.1 #57 · Myth: steeper slopes protect tweeters better → LR4 protects best.

Most power is removed in the first few dB below the corner, and a 48 dB/oct filter passes more near its knee; moving a tweeter from 4 to 2 kHz roughly triples its power. [RARE] [Conflict: a round-4 note-taker could not square "LR4 beats 48 dB/oct" with simple attenuation one octave down; his knee argument is in the blog.] Sources: Audiofrog blog Oct 2021 · DIYMA 2021-07-15

In plain English: A crossover filter keeps low notes away from a tweeter. Its "slope", in dB per octave (an octave is a doubling of frequency), says how fast it cuts the sound below its set point. Many people believe a steeper slope, such as 48 dB/oct, protects a tweeter better. He says LR4 (a common 24 dB/oct filter) protects best. His reason is the "knee" (or "corner"), the bend where the cut begins. Most of the power is removed in the first few dB of cut just below that corner, and right near its knee a 48 dB/oct filter lets more through than LR4 does. Separately, moving a tweeter's crossover from 4 kHz down to 2 kHz roughly triples the power it receives. The [Conflict] tag means this: one of the researchers compiling this collection could not square his claim with simple arithmetic, since one octave below the crossover a 48 dB/oct filter cuts more than a 24 dB/oct one. His own answer, the knee argument, is set out in his blog post.

5.1 #58 · Myth: gentle 6 dB slopes "blend" better and phase shift makes crossovers bad → "Just more mythology".

In cars LR4 beats both 6 dB/oct and steeper slopes; it sums flat and in phase, and its phase shift is intended. Sources: Audiofrog blog Jan 2021 · DIYMA 2021-09-27

In plain English: A 6 dB/oct slope is the gentlest crossover filter: two neighbouring speakers overlap over a wide range of notes. Many people believe this "blends" them better. Others believe crossovers are bad because they cause "phase shift", meaning the filter slightly shifts the timing of some frequencies. He calls both ideas "just more mythology". In cars, he says, LR4 (a 24 dB/oct filter) beats both gentle 6 dB/oct and steeper slopes. Its two halves add up to a flat, even response and stay in phase (in step) with each other. Its phase shift is intended: it is part of how the two halves fit together. His reasoning is that cars use high power and awkward speaker positions, so a narrow overlap blends better and gentle slopes protect too little.

5.1 #59 · Myth: cross the tweeter (or any small driver) as low as it will go → Protect it first.

Tweeters at about 1.5–2× resonance, everything except subs at or above Fs ("resonance IS distortion"); each octave lower costs about 2× the power and 4× the excursion. Sources: OFCAP 19 @1:41:04 · DIYMA 2022-03-10 · DIYMA 2017-08-17

In plain English: The "crossover point" is the note where a speaker starts playing. Many people set each small speaker as low as it can go, to get more out of it. He says a crossover's first job is to protect the smaller speaker. Every speaker has a "resonance" frequency (Fs), the note where it naturally vibrates most easily. He crosses tweeters at about 1.5–2× their resonance (for example, a tweeter that resonates at 1,000 Hz would start at 1,500–2,000 Hz). Every speaker except a subwoofer should be crossed at or above its Fs, because "resonance IS distortion". And each octave lower (each halving of the crossover frequency) costs about 2× the power and 4× the "excursion", meaning how far the cone has to move.

5.1 #60 · Myth: active tweeters need no protection cap, or only audiophile caps → Add a cheap non-polar cap as a "parking brake".

About 68 µF for a 4 Ω tweeter guards against amp failure and DSP typos; a resistor across it cures turn-on pops, and take it out if you return to a passive crossover. [RARE] Sources: Audiofrog blog Dec 2020 · DIYMA 2021-09-22

In plain English: In an "active" system, a DSP (digital signal processor, a small computer that splits and adjusts the sound) does the crossover work, and each speaker gets its own amplifier channel. A "capacitor" (cap) is a small electronic part that blocks low notes, so a cap wired to a tweeter protects it. Many people believe active tweeters need no cap, or that only expensive "audiophile" caps will do. He says to add a cheap "non-polar" cap (one that can be wired either way round) as a "parking brake": about 68 µF (microfarads, the unit of capacitance) for a 4 Ω tweeter. It protects the tweeter if an amplifier fails or if you make a typing mistake in the DSP settings. A resistor connected across the cap (from one leg to the other) cures the "pop" when the system turns on. If you later go back to a passive crossover (a box of parts that splits the sound with no DSP), take the cap out.

5.1 #61 · Myth: set gains with a scope and 0 dB sines, and "max out" everything for resolution → Protect the DSP input first.

Set it so it can never clip at full volume (4 V radio → set 5 V); digital clipping "sounds like broken speakers" and lands in mids and tweeters. Disconnect tweeters before playing 0 dB sines. [RARE] Sources: Audiofrog blog Mar 2021 · DIYMA 2021-08-04 · DIYMA 2020-10-14

In plain English: "Gains" are the input level settings on amplifiers and processors. A common method sets them with a "scope" (oscilloscope, a tool that draws the signal's wave shape) and "0 dB sines" (pure test tones at the maximum digital level). Many people also believe you should turn every level up as high as possible for better "resolution". He says protect the DSP's input first. Set it so it can never "clip" (run out of room and chop off the tops of the waves) even at full volume. For example, if the radio puts out at most 4 V (volts), set the DSP input for 5 V. Digital clipping "sounds like broken speakers", and the harsh distortion lands in the mids and tweeters. And disconnect the tweeters before you play 0 dB test tones, because full-level tones can destroy them.

5.1 #62 · Myth: delay widens the stage or moves the centre → Delay only plays a speaker later; move the image with level.

Set delays by distance, then use balance: 6 dB moves the centre halfway to one side. Delay "tweaks" just add comb filtering, and hard-panned sound comes from its speaker (pan fully left and point). [RARE] Sources: Audiofrog blog Dec 2020 · DIYMA 2017-08-18 · DIYMA 2018-05-22

In plain English: "Delay" (time alignment) is a DSP setting that holds back the sound from the speakers nearest you, so the sound from every speaker reaches your ears together. The "stage" is the illusion of musicians spread across the dashboard, and the "centre" is where a singer seems to stand. Many people believe tweaking delays makes the stage wider or moves the centre. He says delay only makes a speaker play later. His blog's analogy: someone leaving home later doesn't move their house. Set delays from measured distances. Then move the image with level, using the balance control: a 6 dB level difference moves the centre halfway to one side. Delay "tweaks" only add "comb filtering": when the same sound arrives from two places at slightly different times, some notes cancel and others add up, leaving a row of dips like the teeth of a comb. A sound recorded in only one channel ("hard-panned") comes from that channel's speaker. To find your stage's edge, pan the balance fully left and point to where the sound is.

5.1 #63 · Myth: delays need millimetre precision from impulse peaks, lasers or cross-correlation → A tape measure is as good.

An impulse peak is not the arrival time (for subs and dash speakers it is plainly wrong); a one-sample error at 48 kHz is out of phase only at 24 kHz, and a $100 cross-correlation method lands no closer. He was accused of taking SQ "back 20 years" for this. [RARE] Sources: D&F live @15:07 · DIYMA 2021-03-26 · DIYMA 2021-09-02

In plain English: Many people believe delays must be set with great precision, using measuring software (reading the peak of an "impulse response", a recorded test click), laser distance meters or "cross-correlation" (a maths method that compares two signals to find the time gap). He says a tape measure is just as good. The peak of an impulse measurement is not the moment the sound arrives; for subwoofers and dash speakers it is plainly wrong. A DSP running at 48 kHz works in steps of one "sample" (1/48,000 of a second, about 7 mm of sound travel). Being off by one sample only puts two speakers fully out of phase (cancelling) at 24 kHz, above human hearing. And a $100 cross-correlation method gets no closer than a tape measure. Some accused him of taking sound quality "back 20 years" for saying this.

5.1 #64 · Myth: delay should also correct crossover lag, reflections or cabin gain → Delay corrects distance only.

As long as there is a direct path, the direct sound arrives first and sets the delay. [RARE] Sources: DIYMA 2021-03-26 · DIYMA 2021-03-26b

In plain English: Many people try to use delay to fix other things too: the small timing lag a crossover filter adds, sound bouncing off the glass and panels (reflections), or "cabin gain" (the extra deep bass a small car cabin adds). He says delay corrects distance only. As long as there is a straight, unblocked line from the speaker to your ear, the direct sound arrives first. Reflections always come later, so the direct sound is what sets the delay.

5.1 #65 · Myth: half a wavelength of delay plus a polarity flip brings bass forward → It fixes phase at one frequency only.

Bass won't come from the front when high-pass and low-pass phase aren't complementary; keep the acoustic crossover at 80 Hz by lowering the LR4 sub low-pass half an octave per 12 dB of sub boost. [RARE] Sources: DIYMA 2021-08-19 · DIYMA 2021-08-21

In plain English: A trick some tuners use to make bass seem to come from the front: delay the sub by half a wavelength (half of one cycle of a bass note) and flip its "polarity" (swap its plus and minus wires so the cone moves the other way). He says that only fixes the phase at one single frequency. Bass won't sound like it comes from the front when the "phase" (timing within each cycle) of the front speakers' high-pass filter and the sub's low-pass filter aren't "complementary", meaning made to fit together. His recipe: keep the "acoustic crossover", where the sub and front speakers actually meet as measured in the car, at 80 Hz. To do that, lower the sub's LR4 low-pass by half an octave for every 12 dB you turn the sub up. (Half an octave below 80 Hz is about 57 Hz.) He adds that this works if the slopes were EQ'd close to their target shapes during tuning.

5.1 #66 · Myth: flip polarity to fix imaging → A pre-DSP band-aid

that moves the door-speaker cancellation from 250 to 500 Hz; never combine it with delay. Source: Audiofrog blog Dec 2020

In plain English: "Imaging" is how well you can place each sound, like the singer in the centre. The driver sits much closer to one door speaker than the other, so their sounds reach the ear at different times, and at about 250 Hz they cancel each other. Many people "fix" imaging by flipping the polarity of one side (swapping its plus and minus wires). He calls this a band-aid from before DSPs were common. It doesn't remove the cancellation: it only moves it from 250 Hz to 500 Hz. And never combine a polarity flip with delay. In the same blog post he says about 2 ms (two thousandths of a second) of delay on the near side fixes it for one seat.

5.1 #67 · Myth: phase coherence (FIR filters, phase at the tweeter crossover) is the holy grail → "The new snake oil".

Phase at the mid-tweeter crossover can't be measured in a way that matches hearing, and FIR is not inherently "frequency-dependent delay"; what matters is sub-midbass and midbass-mid phase far from 180°. (He does use targeted all-pass filters for two seats.) Sources: DIYMA 2013-12-26 · DIYMA 2021-07-12 · DIYMA 2021-08-20

In plain English: "Phase" describes where each sound wave is in its up-and-down cycle when it reaches you. "Phase coherence" is the idea that every frequency should arrive perfectly in step. Many people treat it as the ultimate goal, chasing it with "FIR" filters (a type of digital filter) or by perfecting the phase where the mid hands over to the tweeter. He calls it "the new snake oil" (a product sold on false promises). At the mid-to-tweeter crossover, the waves are so short that phase can't be measured in a way that matches what you hear. And FIR filters are not, by their nature, "frequency-dependent delay"; he says they can adjust phase separately, but that isn't built into the filter type. What matters is the phase between sub and midbass, and between midbass and midrange: it should be far from 180° (exactly opposite, which cancels). He does use targeted "all-pass" filters, which shift phase without changing loudness, to make a tune work in two seats.

5.1 #68 · Myth: aim (or laser-aim) mids at the listener → Drivers are omnidirectional below beaming;

only tweeters above about 10 kHz respond to aiming. Source: Audiofrog blog Jun 2021

In plain English: Many people carefully aim their midrange speakers at the listener, sometimes even with a laser. He says aiming them achieves nothing. A speaker sends low and middle notes out in all directions ("omnidirectional"). Only above a certain frequency, which depends on its size, does its sound narrow into a beam ("beaming"). Below that point, aiming has no effect. In a car, he says, only tweeters above about 10 kHz respond to aiming.

5.1 #69 · Myth: one target curve is much better than another → Good car curves differ by 1–2 dB;

big claimed differences mean neither was reached. Source: Audiofrog blog Mar 2021

In plain English: A "target curve" is the planned balance of bass, midrange and treble that a tuner aims for when adjusting a system. Many people argue that one curve is far better than another. He says good car target curves differ by only 1–2 dB, which is small. If someone claims a big difference between two curves, it means neither curve was actually reached in the car.

5.1 #70 · Myth: you need an expensive mic, the right 0°/90° file and seated measurements → Calibrated mics mostly agree.

Eleven of his twelve mics agree within about 1 dB (avoid the APEX 220); a car is a diffuse field where mic orientation hardly matters; measure from outside, because your body skews the treble. [RARE] Sources: UMI-1 video @9:01 · DIYMA 2021-08-16 · DIYMA 2020-10-03

In plain English: Tuners measure a car's sound with a measurement microphone. A "calibrated" mic comes with a correction file, often in two versions: "0°" (mic pointed at the speaker) and "90°" (mic pointed up). Many people believe you need an expensive mic, the right file, and measurements taken while you sit in the seat. He says calibrated mics mostly agree: eleven of his twelve measure within about 1 dB of each other, and he says to avoid the APEX 220. A car is a "diffuse field": with so many reflections, sound reaches the mic from all directions at once, so the way the mic points hardly matters. And measure from outside the car, because your body inside it skews the treble readings.

5.1 #71 · Myth: more precision is better (hundreds of mic positions, 1/24–1/48-octave EQ, delays in hundredths of an inch) → Imported pro/home mythology.

Five or six positions around the head are enough; don't EQ the treble comb filter, which shifts with an inch of head movement. [RARE] Sources: DIYMA 2021-08-06 · DIYMA 2020-07-08 · DIYMA 2021-08-26

In plain English: Many tuners believe more precision always helps: measuring at hundreds of mic positions, EQ adjustments only 1/24 or 1/48 of an octave wide (very narrow slices of the sound), and delays set to hundredths of an inch. He calls this mythology imported from professional sound and home audio, methods that don't fit a car. Five or six mic positions around the head are enough. And don't use EQ on the "comb filter" in the treble, the dense pattern of narrow peaks and dips caused by sound arriving from several places at slightly different times. It shifts if your head moves just one inch (about 2.5 cm), and he adds that we don't really hear it.

5.1 #72 · Myth: tune with 31 sine tones and centre the image with pure tones → "A really bad idea".

Above 1 kHz a tone easily lands in a comb-filter null that differs left and right; use band-limited pink noise or music. [RARE] Sources: DIYMA 2021-07-08 · DIYMA 2020-07-08

In plain English: An old competition habit is to tune with 31 "sine tones" (pure single-pitch test tones, one per band of a 31-band equalizer) and to centre the image using pure tones. He calls this "a really bad idea". Above 1 kHz, a single tone can easily land in a "null" (a deep dip where sound cancels) of the car's comb filter. That null sits in a different place for the left and the right speaker, so the tone misleads you. Instead, use band-limited pink noise (a hiss-like test sound with equal energy in every octave, filtered to cover just the range you are checking) or music.

5.1 #73 · Myth: uncorrelated pink noise is the "better" test signal → It has no centre, so it can't check imaging.

Use mono (correlated) noise; uncorrelated noise is only for gaming an IASCA RTA lane or muting the centre in an upmixer. [RARE] Sources: DIYMA 2020-07-08 · DIYMA 2021-05-09 · DIYMA 2022-01-31

In plain English: Pink noise test signals come in two kinds. "Correlated" (mono) noise is identical in the left and right channels. "Uncorrelated" noise is different random noise in each channel. Many people think uncorrelated noise is the better test signal. He says it has nothing common to both channels, so it has no centre image, and so it can't check imaging (whether sounds appear in the right places). Use mono noise. He sees only two uses for uncorrelated noise. One is "gaming" the RTA lane of an IASCA contest (IASCA runs car-audio competitions, and an RTA, or real-time analyzer, measures the system's response there). The other is making every speaker except the centre play when tuning an "upmixer", a processor that creates centre and rear channels from stereo.

5.1 #74 · Myth: a good passenger seat must cost the driver's seat → Only when delay is your only tool.

All-pass filters, an upmixer with a real centre, or kick-panel placement (seat-to-seat path difference under about a foot) serve both seats. [RARE] Sources: SQOLOGY Pt 2 @6:02 · DIYMA 2021-06-29

In plain English: Delay can make the sound from all speakers arrive together at one seat only, usually the driver's. So many people believe that making the passenger seat sound good must make the driver's seat worse. He says that is only true when delay is your only tool. Other tools serve both seats: "all-pass" filters (which shift phase without changing loudness), an upmixer that creates a real centre channel, or placing speakers in the kick panels (the panels by the occupants' feet). Kick panels work because they keep the difference in distance between the two seats under about a foot (about 30 cm).

5.1 #75 · Myth: an L+R "centre" speaker or L−R rear fill works → An L+R centre is worse than none, and he now steers people away from L−R rear fill too.

The real centre is only the content common to both channels; for rear fill use plain stereo, high-passed at 100 Hz or more and delayed about 12–20 ms. [Changed view: in 2017 he gave DIY L−R rear-fill recipes; by 2021 he calls difference rear fill "phasy" and says he may have helped make it semi-mythical.] Sources: DIYMA 2015-08-05 · DIYMA 2021-07-25 · DIYMA 2021-12-07

In plain English: Two DIY tricks are common. One feeds a centre speaker in the dash with "L+R", the left and right channels mixed together. The other feeds the rear speakers ("rear fill", quiet rear sound for a sense of space) with "L−R", only what differs between left and right. He says an L+R centre is worse than having no centre at all. The real centre of a recording is only the content common to both channels, while a plain L+R mix also includes sounds that are only on one side. For rear fill, he now says to use plain stereo, filtered so the rears play only from 100 Hz or more upward and delayed about 12–20 ms (thousandths of a second). The [Changed view] tag means his advice changed over time. In 2017 he gave do-it-yourself recipes for L−R rear fill. By 2021 he called difference rear fill "phasy" (strange-sounding to him) and said he may have helped give it its semi-mythical status on the forum.

5.1 #76 · Myth: "cars do strange things" and "every car is different, just play with it" → Cars are predictable physics.

When tape-measured delays don't give the expected phase, something else is affecting phase; the tape measure isn't broken. Skipping basics turns tuning into a Rube Goldberg machine. [RARE] Sources: DIYMA 2021-08-21 · DIYMA 2021-12-22

In plain English: Many people explain odd results by saying "cars do strange things" or "every car is different, just play with it until it sounds better." He says cars follow predictable physics. If delays set with a tape measure don't give the phase you expected, the right conclusion is that something else is affecting the phase. The tape measure isn't broken. In his view, skipping the basics turns tuning into a "Rube Goldberg machine": a needlessly complicated contraption that does a simple job the long way round. His examples of this are delaying half a wavelength and then flipping polarity, or "playing with the delays" to widen the stage.

The full series

Download the complete guide as one PDF (351 pages, every point with its explanation).