Car Audio Explained, Part 16: Reading Specs and Reviews, and Planning a System (Andy Wehmeyer's Lessons)

4 days ago

27 min read

This is Part 16 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.

5.2 Reading specs, measurements, reviews and claims

What this part is about: Reading the numbers, graphs, reviews and sales claims you meet when shopping for car audio: power ratings, amplifier specs, speaker data sheets and measurements. His general message is to ask what a number really measures, how it was tested, and what the physical mechanism behind a claim is.

5.2 #1 · A speaker's power rating means nothing without its test filter.

Ratings use pink noise with a high-pass for tweeters and mids, so a "100 W tweeter" handles 100 W of high-passed signal (Audiofrog's GB10 sheet pairs 100 W with a 2.5 kHz, 24 dB/oct high-pass). Sources: Audiofrog blog Dec 2020 · GB10OE sheet (COMPANY)

In plain English: A watt rating on a speaker describes a test, not a promise that the speaker can take that much of any signal. Pink noise is a hiss-like test signal containing all frequencies (all notes), with more energy in the low notes. A high-pass filter blocks low notes and lets high notes through. Tweeters (small speakers for high notes) and mids are normally rated with such a filter in front of them, so much of the pink noise's energy never reaches them. His point is that a "100 W tweeter" handles 100 W of that filtered signal. Audiofrog's GB10 tweeter sheet, for example, gives 100 W together with a 2.5 kHz (2,500 cycles per second), 24 dB per octave high-pass, a steep filter. So the watt number only makes sense alongside the filter it was tested with.

5.2 #2 · Read "RMS" as continuous average power plus a separate peak rating, and treat both as limits, not requirements.

Speakers die from heat or mechanical peaks and have no minimum power; Audiofrog's 100-hour test uses shaped noise matching both ratings, and mechanical limits need their own test (Don Keele's ramped-sine "boink" test). [RARE] Sources: OFCAP 19 @26:02 · OFCAP 19 @39:01 · DIYMA 2021-09-19

In plain English: "RMS" is the power figure most people look at on a speaker. He says to read it as the continuous average power the speaker can take, and to look for a separate peak figure for short bursts. Both are ceilings, not amounts you must supply: a speaker has no minimum power. Many people believe a speaker "needs" its rated watts to work properly; he disagrees. A speaker dies in one of two ways. It overheats (its voice coil, the wire coil that drives the cone, gets too hot), or a big peak pushes it past its mechanical limits. Audiofrog's own test runs for 100 hours with a noise signal shaped so that its average matches the RMS rating and its peaks match the peak rating. Mechanical limits need a test of their own. He names Don Keele's "boink" test, which uses a ramped sine (a single pure tone that steadily gets bigger) to test the mechanics without heating the coil.

5.2 #3 · Sanity-check amp power claims against the fuse and the price.

A 20 A fuse on a "2000 W" amp, or 2000 W for $99, can't be true; car amps are roughly constant-voltage, so "doubling into 2 Ω" is theory, not a given. Sources: DIYMA 2011-12-16 · DIYMA 2013-04-23

In plain English: A fuse is a safety part that blows if an amp draws too much current from the car's electrical system. An amp can't put out more power than it takes in, so a small fuse and a huge watt claim don't fit together. His examples: a 20 A fuse on a "2000 W" amp, or 2000 W for $99, can't be true. He also explains a common spec trick. Ohms (Ω) measure a speaker's electrical load; lower ohms lets more current flow. Car amps behave roughly like a fixed-voltage source, and in theory the same voltage into half the load (2 Ω instead of 4 Ω) gives double the power. He says that doubling is only theory, not something an amp is guaranteed to deliver. So in practice: check the claimed watts against the fuse size and the price.

5.2 #4 · A heavy heatsink signals sustained power; small amps often roll back.

Some amps cut output within minutes at rated power, so "150 W" behaves like 75 W; many tiny amps now cut rail voltage when hot instead of using aluminium. [RARE] Sources: CMA 2022 @1:09:04 · DIYMA 2022-01-14

In plain English: Amplifiers turn part of their power into heat. A heatsink is the finned metal body, usually aluminium, that carries the heat away. His point is that a big, heavy heatsink is a sign the amp can keep up its power for a long time. Many small amps can't. They "roll back", meaning they cut their output when they get hot. He says some do it within minutes at their rated power, so an amp sold as "150 W" behaves like a 75 W one. "Rail voltage" is the internal supply voltage an amp uses to make its output; lowering it lowers the power available. Many tiny amps now protect themselves this way when hot, instead of carrying lots of aluminium. Many people take a small amp's watt rating at face value; he says to look at how it deals with heat.

5.2 #5 · Compare amps' distortion and noise at about 1 W, not at rated power.

A tweeter lives near a watt, some amps distort more at very low power, and hiss matters more in a car because you sit close to tweeters. [RARE] Sources: DIYMA 2021-05-18 · CMA 2023 @22:05

In plain English: Amp spec sheets usually quote distortion (unwanted changes the amp adds to the signal) and noise at full rated power. He says to compare amps at about 1 watt instead, because that is close to what a tweeter actually receives with music. Some amps distort more at very low power than at high power, and a full-power figure hides that. Noise matters more in a car too. Hiss is the steady "shhh" that comes from the speakers even with no music playing, and in a car you sit very close to the tweeters, so you hear it. Many people compare amps by their full-power numbers; his point is that the figures near 1 W tell you more about how the amp will sound.

5.2 #6 · Class-D treble response depends on the speaker load.

A class-D output filter is designed for one load like a passive crossover, so 2 Ω vs 4 Ω gives a roll-off or a peak; Audiofrog moved its filter to 40 kHz. Real, he says, but not what separates great cars from bad ones. [RARE] Sources: CMA 2022 @35:07 · B1 @29:58

In plain English: Class-D is the efficient amplifier design used in most small car amps. It works by switching on and off very fast, and it needs a filter on its output to remove that switching signal. A passive crossover is the set of coils and capacitors that splits sound between speakers, and it only works as designed with the speaker it was designed for. He says a class-D output filter is the same: it is designed for one load. Connect a different load, such as 2 Ω instead of 4 Ω, and the top of the treble either falls away (a roll-off) or rises (a peak). Audiofrog moved its filter up to 40 kHz (40,000 cycles per second, well above the roughly 20 kHz upper limit of human hearing) so the effect lands outside what you can hear. He says the effect is real, but it is not what separates great-sounding cars from bad ones.

5.2 #7 · Small class-D amps: ask about FM interference, hiss and layout.

Tiny amps hiss or whine because the supply transformer's field sits near the preamp; look for shielding, four-layer boards and careful transformer placement. Nothing on the box warns you. [RARE] Sources: CMA 2023 @26:04 · B1 @26:01

In plain English: Very small amps pack everything tightly together. He says that is why many hiss or whine: the magnetic field from the power-supply transformer (the part that converts the car's 12 volts into what the amp needs) sits close to the preamp, the sensitive first stage that handles the weak incoming signal. Class-D switching can also interfere with FM radio reception. What helps is shielding (metal that blocks the interference), four-layer circuit boards (boards with extra internal layers, which can include a ground layer acting as a shield) and careful placement of the transformer. Nothing on the box warns you about any of this. So in practice: before buying a small class-D amp, ask about FM interference, hiss and how it is laid out.

5.2 #8 · Amps with built-in DSP are convenience products with a gain-structure trap.

Big inputs are attenuated to the converter's limit, and the output DAC can't exceed 0 dBFS with no analog gain after it, so after tuning cuts a "1,000 W" amp may deliver about 10 W. For high-end systems use a separate DSP and amps with analog input controls. [RARE] Sources: B1 @33:35 · B1 @36:46

In plain English: A DSP (digital signal processor) is the small computer that handles crossovers, delays and EQ (tone adjustments) in a modern system. Some amps have one built in. He calls these convenience products and warns of a "gain-structure" trap. Gain structure means how the signal level is set at each stage of the chain. The trap, step by step. First, the input: the converter that turns the incoming signal into numbers can only accept a limited voltage, so a big input signal has to be turned down to fit. Second, the output: the DAC (digital-to-analog converter) that turns the numbers back into a signal cannot go above 0 dBFS, the digital maximum ("full scale"). In these amps there is usually no analog volume control after that converter, so nothing can raise the level later. Third, tuning: tuning mostly cuts levels. With nowhere to get the level back, he says, a "1,000 W" amp may end up delivering about 10 W. So in practice: for high-end systems he recommends a separate DSP plus amps with analog input (gain) controls.

5.2 #9 · Know the sensitivity arithmetic.

Compare at 2.83 V (car amps are voltage sources); 80→125 W is only +1.9 dB, and about 99% of amp power becomes heat. A sub's sensitivity is a midband number that doesn't predict in-car level, and GB tweeters are 6 dB hotter on purpose so you can turn gains down and cut hiss. [RARE] Sources: DIYMA 2015-11-26 · DIYMA 2022-01-26 · DIYMA 2017-10-01

In plain English: Sensitivity is how loud a speaker plays for a given input. A dB (decibel) is the unit for differences in loudness. He says to compare speakers at 2.83 volts rather than "1 watt", because car amps act as voltage sources: they deliver a set voltage, and the watts that flow depend on the speaker's ohms. (2.83 V gives 1 W into 8 Ω or 2 W into 4 Ω.) Some arithmetic he gives: going from 80 W to 125 W adds only 1.9 dB, a small step in loudness. And a speaker turns only about 1% of the amp's power into sound; about 99% becomes heat. A subwoofer's sensitivity figure is measured in its middle range and doesn't predict how loud it will be in a car. Finally, Audiofrog's GB tweeters are made 6 dB more sensitive on purpose, so you can turn the amp's gain down, which also cuts hiss. Many people chase more watts; his numbers show how little extra watts buy.

5.2 #10 · Efficiency vs power: context, not reversal.

In 2008–2013 "efficiency is overrated, power is cheap" argued for active crossovers; in 2021 he stressed efficiency for very loud systems, where heavy coils lower efficiency further. Sources: DIYMA 2013-11-28 · OFCAP 19 @33:07

In plain English: Efficiency is how much sound a speaker makes from a given amount of power. Two statements of his can look like opposites, and this point explains that they answer different questions. Between 2008 and 2013 he said "efficiency is overrated, power is cheap". That was an argument for active crossovers: giving each speaker its own amp channel and setting levels electronically, instead of juggling resistors and parts in a passive crossover to balance the speakers' levels. In 2021 he stressed efficiency for very loud systems. There, the heavy voice coils needed to take huge power make the speakers less efficient still, so efficiency matters more.

5.2 #11 · Thiele/Small parameters tell box fit and low-frequency roll-off, not quality.

"Low-Q woofers are more accurate" is mostly wrong; BL alone means nothing; and Qts = Qes·Qms/(Qes+Qms), not a product. Sources: DIYMA 2009-10-23 · DIYMA 2021-05-09

In plain English: Thiele/Small (T/S) parameters are a standard set of numbers describing how a woofer behaves at low frequencies, such as its resonance and its "Q" values. He says they tell you what box suits the driver and how its bass rolls off (fades out) at the bottom. They don't tell you whether it is a good speaker. Q describes how strongly the speaker's resonance is held in check. Many people believe "low-Q woofers are more accurate"; he says that is mostly wrong. BL, the strength of the motor (magnet and coil), means nothing on its own. He also corrects a formula people get wrong: total Q (Qts) is Qes × Qms divided by (Qes + Qms), the same maths as two resistors in parallel, not simply Qes × Qms.

5.2 #12 · Complete speaker data is enough to judge a driver better than ears.

T/S parameters, on-axis plus 15/30/45/60° responses and 2nd/3rd-harmonic distortion; check that off-axis angles match before comparing sheets. Klippel data mainly saves designers time, and specs reveal a driver's intended job. [RARE] Sources: DIYMA 2016-01-29 · DIYMA 2021-12-24 · DIYMA 2022-01-24

In plain English: A driver is a single speaker unit (a woofer, mid or tweeter). He says a complete set of data lets you judge a driver better than your ears can. Complete means: the T/S parameters; the frequency response (a graph of how loud it plays each note) straight ahead, called on-axis, plus at 15, 30, 45 and 60 degrees to the side; and its second- and third-harmonic distortion (unwanted extra tones at two and three times the note being played). Before comparing two data sheets, check that both measured at the same off-axis angles. Klippel is a specialised speaker-measurement system; he says its data mainly saves designers time. Specs also reveal what job a driver was made for. Many people believe only listening can tell you whether a speaker is good; he disagrees.

5.2 #13 · Red flags on a spec sheet.

Missing curves, 1/3-octave smoothing, impossible wiggles where cones are still pistonic (near 80 Hz on a 6.5"), sub curves flat to the top; about 10% tolerance on suspension parameters is normal. Sources: DIYMA 2015-09-30 · DIYMA 2015-05-31

In plain English: A spec sheet is the maker's data page for a speaker. He lists warning signs. Graphs (curves) that are missing. Curves shown with 1/3-octave smoothing, which blurs away detail the sheet should show. Wiggles (peaks and dips) where they are physically impossible: a cone is "pistonic" when it moves all in one piece, like a piston, and in that range its response is smooth, so wiggles near 80 Hz on a 6.5-inch speaker are a red flag. And subwoofer curves shown flat all the way to the top of the graph; the sources list this as a red flag without spelling out the reasoning. On the other hand, about 10% variation from unit to unit in suspension parameters (those set by the springy parts that hold the cone) is normal, not a defect.

5.2 #14 · Read distortion by type and shape, not just level.

Even-order comes from asymmetry (often the spider) and masks detail; odd-order comes from the coil leaving the gap and sounds mechanical. Convert with 20·log; near-field measurements miss cone breakup, so measure far enough away. [RARE] Sources: DIYMA 2013-11-29 · DIYMA 2021-12-14 · DIYMA 2017-07-30

In plain English: Distortion is sound a speaker adds that wasn't in the music, usually extra tones at multiples of the note being played. He says don't only look at how much there is; look at its type and its shape. Even-order distortion comes from something lopsided in the speaker, often the spider (the springy fabric ring that holds the voice coil centred), and he says it masks (covers up) detail. Odd-order distortion comes from the voice coil moving out of the magnet's gap on big movements, and he says it sounds mechanical. Distortion given as a percentage converts to decibels with 20 × log; he gives 0.2% as about −54 dB. Finally, close-up (near-field) measurements miss cone breakup, the point where the cone stops moving in one piece and starts to flex, so measure from far enough away.

5.2 #15 · In cars off-axis behaviour matters more than on-axis.

You always listen off axis near reflecting surfaces, so judge drivers at the angle you'll hear them; off-axis holes cost midrange definition. [RARE] Sources: Audiofrog blog Dec 2020 · Crutchfield

In plain English: "On-axis" means straight in front of a speaker, where it is usually measured. "Off-axis" means at an angle. In a car you almost never sit straight in front of a speaker, and the speakers sit near surfaces such as glass and the dashboard that reflect sound. So he says to judge a driver by how it performs at the angle you'll actually hear it from. An off-axis "hole" is a dip, a band of notes that goes quiet at certain angles; he says such holes cost the midrange its definition (clarity). Many people judge speakers by their straight-ahead graph; in a car, he says, the angled response matters more.

5.2 #16 · An in-car graph shows the car, not the speaker.

A speaker outputs what its manual graph shows; every change at the seat is the car or a filter you added, so single-point in-car RTAs make every speaker look bad. Read them spatially averaged or smoothed. Sources: DIYMA 2021-08-06 · DIYMA 2015-12-22

In plain English: An RTA (real-time analyser) is a microphone plus software that shows how loud each frequency is where the microphone sits. His point: a speaker puts out what the graph in its manual shows. Any change you see at the seat comes from the car (its reflections and shape) or from a filter you added, not from the speaker. One microphone in one spot picks up a jagged pattern of peaks and dips from reflections, so it makes every speaker look bad. To read in-car graphs fairly, he says to use spatial averaging (measuring at several spots around the listening position and averaging them) or smoothing (a setting that evens out the fine jaggedness).

5.2 #17 · A measurement is invalid only if you misread it.

A calibrated mic is "correct" but the screen can mislead (comb filters, no distortion shown, no centre in stereo noise); the test is whether it correlates with hearing, like a scale that doesn't "feel" weight yet is still useful. [RARE] Sources: DIYMA 2017-06-28 · DIYMA 2017-10-28 · DIYMA 2021-07-09

In plain English: A calibrated measurement microphone (one whose own errors have been measured and corrected for) reports the sound at its tip correctly. He says the problem is how people read the screen. His examples: comb filters, a row of narrow dips caused by a sound and its reflection arriving at slightly different times; distortion, which the usual graph doesn't show; and stereo test noise, which has no centre, so it can't show the centre image. The real test is whether a measurement correlates with (lines up with) what you hear. His comparison is a weighing scale: it doesn't "feel" weight the way you do, yet it is still useful. Many people say measurements miss what ears hear; he says a measurement is only invalid if you misread it.

5.2 #18 · Store sound boards and demo walls mislead.

A board hurts every speaker mounted on it, a low-Q midbass sounds thin on a board but full in a door, and a mistuned board (wrong port tuning) hurts every brand on it. Sources: DIYMA 2015-11-15 · B2 @57:53

In plain English: Shops often show speakers on a sound board or demo wall: a panel of speakers you can switch between. He says these mislead. A board hurts every speaker mounted on it. A low-Q midbass (a woofer designed with less of a bass bump) sounds thin on a board but full once it is in a car door, because a door mounting adds bass that the board doesn't. And if the board itself is built wrong, for example a port (the tube in a bass box) tuned to the wrong frequency, every brand on it suffers. So a store demo tells you little about how a speaker will sound in your car.

5.2 #19 · Liking is not proof: ask what was compared, and demand quantified claims.

"I did all this and I like it" doesn't mean correct; unquantified workarounds become myths ("play with delays for width"). Test claimed DSP differences with a null test (align, invert, mix). [RARE] Sources: DIYMA 2015-11-14 · DIYMA 2021-08-14 · DIYMA 2021-01-04

In plain English: He says enjoying a result doesn't prove it is right. When someone says "I did all this and I like it", ask what they compared it with. "Quantified" means backed by numbers or measurements. Tips without numbers, such as "play with the delays to get more width", get repeated until they become myths. For claims that one DSP sounds different from another, he describes a null test, step by step: record the same music through both; line the two recordings up exactly in time; flip one upside down (invert its polarity); then mix them together. Whatever is identical cancels out, and only the real difference remains. He adds that working out what causes that difference is the harder part.

5.2 #20 · Independent reviews and maker specs differ for honest reasons.

Environments and techniques differ, so "gotcha" reviewing is silly, though some makers over-claim; he found an independent review useful for catching manual errors. Source: DIYMA 2017-08-09

In plain English: When an independent reviewer measures a speaker, the numbers often don't match the maker's spec sheet. He says there are honest reasons: different test environments and different measuring techniques give different results. So he calls the "gotcha" attitude, treating any mismatch as proof the maker lied, silly. He also admits that some makers do over-claim. And he found one independent review of his own product useful: it caught small mistakes in the owner's manual that he would fix.

5.2 #21 · "Quality" has two meanings, and there is no perfect speaker.

Engineers mean meeting spec consistently, buyers mean meeting expectations; change one thing in a speaker and "a thousand other things" change, so each is right only for an application. Sources: SQOLOGY Pt 1 @49:05 · CMA 2022 @7:02

In plain English: He says the word "quality" means different things to different people. To an engineer, a quality product is one that meets its specification every time, unit after unit. To a buyer, quality means it does what they expected. He also says there is no perfect speaker. Change one thing in a speaker's design and "a thousand other things" change with it, so every design is a set of trade-offs. That means a speaker is only right for a particular application: a particular job, location and use.

5.2 #22 · Car drivers need car-grade qualification.

Audiofrog's 40-page standard includes a 200-hour heat-humidity cycle (90 °C to −35 °C), vibration, thermal shock and UV; home drivers aren't built for this. Sources: DIYMA 2015-07-27 · CMA 2021 @46:02

In plain English: A car is a harsh place for a speaker: it bakes, freezes, gets damp and shakes. He says car speakers must pass car-industry testing. Audiofrog's standard runs to 40 pages. It includes a 200-hour cycle of heat and humidity between 90 °C (194 °F) and −35 °C (−31 °F), plus vibration, thermal shock (sudden temperature changes) and UV (sunlight) exposure. He says drivers made for home speakers aren't built for this.

5.2 #23 · Buy performance, not features; ask a brand what makes it better.

Optical inputs, "super-duper" DACs and gold RCAs are features whose importance is "rarely substantiated"; a DSP's software is the hard part, and "12 channels" or "passion" is not an answer. [RARE] Sources: DIYMA 2008-06-11 · OFCAP 19 @1:30:03

In plain English: A feature is something a product has; performance is how well it does its job. He names features whose importance, he says, is "rarely substantiated" (rarely shown to matter): optical inputs (a digital connection by light cable), "super-duper" DACs (the chips that turn digital audio back into a signal) and gold-plated RCA plugs (the usual audio cable connectors). For a DSP, he says the software, the part you use to set it up and tune it, is the hard part to get right. So in practice: ask a brand what makes its product better. Answers like "12 channels" or "passion" don't answer that question. Many buyers compare products by their feature lists; he says to compare what they actually do.

5.2 #24 · Run claims through the "Truth-O-Meter": what is the mechanism?

Metaphor is not engineering ("effortless", "blacker blacks", "chocolaty" bass, tweeters "made of weasel noses"); he says audio has much "purposeful confusion". Debunking without a working alternative is no better than a falsehood. [RARE] Sources: DIYMA 2011-03-26 · DIYMA 2021-09-15 · B2 @17:53

In plain English: His "Truth-O-Meter" is a simple test for any audio claim: what is the mechanism, the actual physical reason it would work? He says metaphor is not engineering. He gives examples of language he rejects: "effortless" sound, "blacker blacks", "chocolaty" bass, and (mocking marketing) tweeters "made of weasel noses". He says audio has a lot of "purposeful confusion", meaning some of the confusion is created on purpose. He also warns debunkers: knocking down a false claim without offering something that actually works is no better than the false claim itself.

5.2 #25 · A subjective impression is an experience, not data, but "unmeasured" isn't "imaginary".

Only someone who knows every measuring tool can say which; trained ears can be an objective tool. Source: SQOLOGY Pt 1 @45:03

In plain English: A subjective impression is what a person hears and feels while listening. He says it is an experience, not data. But he warns against the opposite mistake too: something that hasn't been measured isn't automatically imaginary. Telling which is which, he says, takes someone who knows every measuring tool. He adds that trained ears can themselves be an objective tool.

5.2 #26 · Read the books, not the forums.

Toole's Sound Reproduction first, then Colloms and Darlington's High Performance Loudspeakers, Geddes and Voishvillo on nonlinearity, Griesinger's site and the ARTA near-field note; Facebook and YouTube "expertise" often treats car acoustics as unknowable. Sources: DIYMA 2008-10-01 · DIYMA 2021-12-24 · DIYMA 2021-07-26

In plain English: He sends learners to books and researchers rather than forums. His list: Toole's book Sound Reproduction first; then High Performance Loudspeakers by Colloms and Darlington; Geddes and Voishvillo on nonlinearity (the ways speakers stop behaving perfectly, which is what produces distortion); the website of Griesinger, an acoustics researcher; and the ARTA near-field note (a technical note on measuring speakers up close, for the ARTA measurement software). He says Facebook and YouTube "expertise" often treats car acoustics as unknowable.

5.2 #27 · Forum culture: "the loudest idiot is the authority".

Non-experts should qualify claims ("I think"), and answers boiled down too far stop being true. Sources: DIYMA 2008-07-27 · OFCAP 19 @17:03

In plain English: He describes forum culture as "the loudest idiot is the authority": whoever shouts loudest gets treated as the expert. He wants people who aren't experts to qualify their claims, with words like "I think". He also warns that answers boiled down too far, to make them short and simple, stop being true.

5.2 #28 · Ignore claims of custom MS-8 firmware.

It was an encrypted binary, only one tiny update ever shipped, and the only real mods were hardware (op-amps). [RARE] Source: DIYMA 2021-09-28

In plain English: This point is about one product: the JBL MS-8, a car-audio sound processor he often discusses. Firmware is the software built into a device. Some people claimed to have custom MS-8 firmware; he says to ignore those claims. The firmware went to the factory as an encrypted binary (scrambled, compiled code that can't simply be read or edited), and only one tiny update was ever released. The only real modifications people made were to the hardware, such as swapping op-amps (small amplifier chips in the signal path).

5.3 Planning and buying a system

What this part is about: Planning a system and spending money wisely: where money matters most (installation and tuning), how to upgrade a factory system in stages, which parts are worth paying for, and what to avoid buying.

5.3 #1 · Money saved on installation is the worst money you'll ever save.

Cars need far more fabrication, disassembly and troubleshooting than home audio; some customers have paid six stores. Sources: OFCAP 19 @2:28:05 · SQOLOGY Pt 2 @38:02

In plain English: Installation is the work of fitting the system into the car. He says money saved there is "the worst money you'll ever save". Setting up home audio mostly means placing speakers and plugging things in. A car needs far more: fabrication (building custom mounts, panels and boxes), taking the car's interior apart, and troubleshooting. He says some customers have been to six different stores and paid six times.

5.3 #2 · A correct tune matters more than $200 vs $2,000 speakers; priorities are processing, installation, then brand.

EQ is "THE ONLY THING" that overcomes the car's response; at a China dealer training every bad car's fault was installation, adjustment or design, never the gear. [RARE] Sources: OFCAP 19 @1:54:03 · DIYMA 2008-06-21 · DIYMA 2010-06-06

In plain English: "Tuning" means setting the processor's EQ, delays, crossovers and levels for your particular car. He says a correct tune matters more than whether the speakers cost $200 or $2,000. His order of priority is: signal processing first, installation second, the brand of gear third. EQ (equalisation: turning individual frequencies up or down) is, in his words, "THE ONLY THING" that overcomes the car's own effect on the sound, because even the flattest speaker stops being flat once it is mounted in a car. His evidence: at a dealer training in China, half the dealers planned to replace all their equipment because their cars sounded bad, but in every bad car the fault was installation, adjustment or system design, never the gear. Many people believe better speakers are the main route to better sound; he ranks the brand of gear last.

5.3 #3 · Speakers still matter for output and distortion: a smaller budget should cost loudness, not accuracy.

No DSP removes a bad speaker's distortion; his GS line keeps accuracy and gives up maximum output versus GB. Sources: DIYMA 2016-03-21 · OFCAP 19 @1:36:02

In plain English: This balances the previous point. Speakers still matter for how loud a system can play and how much it distorts. No DSP can remove the distortion a bad speaker makes. His rule for a smaller budget: it should cost you loudness, not accuracy. Audiofrog's GS line is built that way: compared with the GB line, it keeps the accuracy and gives up maximum output.

5.3 #4 · Baseline first: full bandwidth with a sub, enough output, no distortion or rattles; then width and a stable centre.

Width is his personal top priority; thick midbass is his pet peeve. Sources: OFCAP 38 @35:03 · OFCAP 19 @1:35:05

In plain English: He sets a baseline every system must reach before anything else. Full bandwidth with a sub: the system plays the whole range of notes, with a subwoofer for the deepest bass. Enough output: it plays loud enough. No distortion and no rattles. Only after that come the refinements: width (how far the sound spreads across the front of the car) and a stable centre (a singer who stays put in the middle). Width is his own top priority. "Thick midbass", too much of the upper-bass range, is his pet peeve.

5.3 #5 · Decide early whether the system serves one seat or every seat.

He would take an immersive GS system (rears plus upmixer) over a two-channel GB one; a big-budget system that makes the passenger ask "you paid how much?" is a lost sale. [RARE] Sources: OFCAP 38 @45:00 · CES 2016 @1:02

In plain English: Many high-end car systems are tuned for the driver's seat only. They use delays (holding back the nearer speakers) so that, from that one seat, each sound seems to come from the right place on an imaginary stage, at the expense of the other seats. He says to decide at the start whether the system is for one seat or for every seat. He would choose an immersive system built from Audiofrog's GS speakers, with rear speakers plus an upmixer (a processor that spreads two-channel stereo across more speakers), over a two-channel system built from the higher-end GB speakers. And if a big-budget system makes the passenger ask "you paid how much?", he counts that as a lost sale. Many people believe sound quality means optimising for the driver alone; he disagrees.

5.3 #6 · Factory-system upgrade path: speakers first on the factory amp and EQ, then power, then sub, rears and DSP.

Only about 10–20% of a factory EQ compensates for the stock speakers; example tiers on a Bose truck ran about $2,000 / $3,500 / $4,500 / $7,500. [RARE] Sources: CMA 2023 @11:00 · CMA 2023 @49:01

In plain English: Many cars come with a factory sound system that has its own amplifier and EQ. He describes upgrading it in stages. First, replace the speakers but keep the factory amp and EQ. Next, add power (aftermarket amps). Then add a subwoofer, rear speakers and a DSP. Many people believe new speakers are pointless without new amps and a DSP, because the factory EQ was tuned around the cheap stock speakers. He says only about 10–20% of a factory EQ makes up for the stock speakers; the rest corrects for the car's cabin, so it stays useful after a speaker swap. His example tiers on a Bose-equipped truck cost about $2,000, $3,500, $4,500 and $7,500.

5.3 #7 · Keep good factory processing; don't "sum it back to flat".

Factory all-pass filters can't be un-EQ'd, so on Logic 7 / Mark Levinson-type systems "DO NOT TRY TO SUM" the car back; in upmixing factory systems wire each factory channel one-to-one to new amps, fit stronger speakers, then just re-EQ and set levels. [RARE] Sources: CMA 2021 @35:00 · DIYMA 2021-09-26 · B1 @58:34

In plain English: Many premium factory systems split and process the sound heavily before it reaches each speaker. A common aftermarket approach is to "sum" (recombine) the factory channels back into a plain, flat stereo signal, then process it again with a new DSP. He says don't do this with good factory processing. One reason: factory systems use all-pass filters, which shift timing (phase) without changing loudness, and those can't be undone with EQ. So on Logic 7 or Mark Levinson-type systems, which use an upmixer (a processor that spreads stereo across many speakers), his words are "DO NOT TRY TO SUM" the car back. Instead, in an upmixing factory system, connect each factory channel one-to-one to a new amp channel, fit stronger speakers, then just re-EQ and set the levels. Many installers believe the first step is always to recover a flat signal; he disagrees for these systems.

5.3 #8 · The DSP is your installer's tool: let them choose it.

If you tune yourself, buy the one whose features you can learn; any amp "that doesn't suck" works, but for high-end systems prefer separate DSP and amps (5.2 #8). Sources: OFCAP 19 @2:46:08 · B1 @38:23

In plain English: He sees the DSP as the installer's tool, so let the installer choose the one they know, the way a carpenter picks their own tools. If you will tune the system yourself, buy the one whose features you can learn to use. For amplifiers, he says any amp "that doesn't suck" works. For high-end systems, though, he prefers a separate DSP and separate amps, for the gain reasons explained in 5.2 #8.

5.3 #9 · Pay only for parts that buy something audible.

With a sub you don't need an 8" midbass; tweeters close to the woofer can stay passive; swapping one good 6.5" for another at similar power changes little; most "voicing" you dislike is fixable with EQ. Don't reuse a passive crossover on a different speaker. Sources: CMA 2023 @7:01 · DIYMA 2021-12-03 · DIYMA 2021-12-21

In plain English: He says to spend money only where you'll hear the difference. His examples. With a subwoofer handling the deep bass, you don't need an 8-inch midbass (the woofer in the door). Tweeters mounted close to the woofer can stay on a passive crossover (the coils and capacitors supplied with the speakers) instead of each getting its own amp channel. Swapping one good 6.5-inch speaker for another at similar power changes little. "Voicing" is a speaker's built-in tonal balance; most voicing you dislike can be fixed with EQ. And don't reuse a passive crossover with a different speaker from the one it was designed for.

5.3 #10 · Skip showpiece fabrication if sound is the goal.

A beautiful trunk with cheap amps and speakers sounds poor; use factory locations when they work. Sources: OFCAP 38 @32:04 · OFCAP 38 @1:05:01

In plain English: Showpiece fabrication means show-car work, such as a beautifully built trunk display. He says that if good sound is your goal, skip it. A gorgeous trunk with cheap amps and speakers still sounds poor. And where the car's factory speaker locations work, use them.

5.3 #11 · The core upgrade is still "better front plus real bass".

Carmakers can't add a 40 lb sub box and a 60 A amp (his OEM limits: about 30 A and 12 lb), and factory head units can beat aftermarket ones (balanced outputs). Sources: OFCAP 19 @58:06 · DIYMA 2008-06-21

In plain English: The aftermarket means parts added after the car leaves the factory. He says the core thing it offers is still "better front plus real bass": better front speakers and a proper subwoofer. Carmakers can't add a 40 lb (about 18 kg) subwoofer box and an amp drawing 60 A. The limits he gives for factory (OEM) systems are about 30 A and 12 lb (about 5.4 kg). He also says a factory head unit (the radio unit in the dash) can be better than an aftermarket one; his example is balanced outputs, a type of signal connection that rejects electrical noise.

5.3 #12 · No adjustable system is "entry-level".

Two front speakers plus a DSP amp, installed, costs at least about US$2,500, so that buyer expects great sound and the shop must deliver it. Source: B2 @25:03

In plain English: "Entry-level" means the cheapest, most basic tier, where buyers expect little. He says no adjustable system (one with a DSP that can be tuned) is entry-level. Even two front speakers plus an amp with a built-in DSP, installed, costs at least about US$2,500. Someone spending that much expects great sound, so the shop has to deliver it.

5.3 #13 · Cheap fixes outrank exotic ones.

A clean windshield "probably" matters more than rear delay settings, quieter tyres were his best midbass upgrade, and 5.1 vs 7.1 is splitting hairs. [RARE] Sources: DIYMA 2012-02-02 · DIYMA 2016-01-25

In plain English: He says simple, cheap fixes often beat exotic ones. A clean windshield "probably" makes a bigger difference than the delay settings for the rear speakers. Quieter tyres were his best midbass upgrade. And arguing over 5.1 versus 7.1 (surround-sound layouts with five or seven main channels plus a sub) is splitting hairs. The source notes don't spell out his reasoning for the windshield or the tyres. Many people chase advanced settings and formats first; he points to the basics.

5.3 #14 · Never buy used speakers, amps or DSPs online except from the maker.

It's worse than used tyres; Audiofrog drivers can't be reconed, so "maybe you can have it fixed" means walk away, and pay with recourse if you must. [RARE] Source: DIYMA 2020-07-01

In plain English: He advises never buying used speakers, amplifiers or DSPs online unless they come from the manufacturer. He calls it worse than buying used tyres. Reconing means repairing a blown speaker by replacing its cone, coil and suspension, and Audiofrog drivers can't be reconed. So if a seller says "maybe you can have it fixed", walk away. If you must buy used anyway, pay with recourse, meaning a payment method that lets you get your money back if the item turns out to be faulty. Many people see used gear as a smart bargain; he disagrees.

The full series

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