2-Way vs 3-Way and Full-Range versus Sub+Sat

so @kimkardashian did NOT ask for this explainer but he is going to get it anyway

before we start i have already written about how subwoofers and midranges are engineered differently here:

and here:

i don’t remember what i wrote there and won’t re-read what i wrote so i may repeat myself or may miss something twice - just ask if you have any questions.

here i want to focus on how different types of drivers can be used SYNERGISTICALLY to complete the system.

my preference is obviously for Sub+Sat system and for a 3-Way + Sub system at that.

a technically optimal / perfectly engineered setup would be a JBL VTX F35

paired to a distributed array of multiple 21" subwoofers based on the Tetracoil driver:


and using large vented enclosures, possibly with variable tuning ( multiple ports that can be individually plugged ). being able to tune the ports from about 16 to 32 hz would be ideal. if you have four ports ( one in each corner of the baffle ) and can run 1, 2, 3 or 4 of those ports you would have this type of flexibility to tune the subs either for music or movies, either for home or PA use and so on. 16 hz would be ideal for home theater for example while 32 hz would be ideal for PA use.

you can’t do better than this - you might be able to save money by DIYing the equivalent of VTX F35 instead of buying it. the F35 is about $7,000 each or about $5,000 on sale and it only has about $2,000 worth of components in it, so there is potential for savings there …

but the real expense in running F35s isn’t the speaker itself - and not even the amplification - but the DSP required …

or rather it must be powered by a Crown AMP that has built in DSP presets for it, it must be Quad-Amped and those amps cost an arm and a leg …

thus the real $$$ saving for DIYer comes from being able to design and use your own DSP filter and perhaps buying used amplifiers on Ebay. this combined with making your own cabinets will save you something like $10,000 at which point it may actually justify the effort so long as you enjoy the process.

it wouldn’t be easy, i have never done anything on that level and that isn’t what this thread is about.

this thread is merely to explain WHY this is the type of setup that is optimal …

ok so why is this ideal ?

well basically experiments have found that the lower the frequency gets the harder it is for the brain to detect distortion …

so while 1% distortion is audible in the vocal range even 10% distortion is not audible in subwoofer range

and of course at the very top end distortion again becomes in audible, but we’re talking the VERY top end like the Hi-Hat frequency range …

so let’s say:

10+ khz - harmonic distortion totally inaudible even at 100% distortion, only Intermodulation Distortion and Sub-Harmonic distortion audible. Sub-Harmonic distortion is a strange type of distortion when spurious tones are produced at frequencies BELOW the fundamental. Beyma did a paper on this - basically has to do with suspension design of compression drivers if i remember correctly. IMD is when two or more tones combine to form a third tone and this can be an issue in Compression Drivers and the amount of IMD will be somewhat proportional to THD so even if THD itself is inaudible over 10 khz it is not irrelevant, because it is potentially an indicator of IMD, which is itself rarely measured ( due to measurement complexity ) but usually proportional to THD.

6 - 10 khz: harmonic distortion somewhat audible, but may not be audible as distortion as such. instead may sound as “hash” or lack of clarity. that is the harmonics may be audible but we may not be able to detect their pitch to recognize them as harmonics. IMD would of course also be an issue in this range.

500 hz - 6 khz: this range has to be absolutely pristine. any distortion in this range makes your speaker not worth the plywood it is made out of.

100 hz - 500 hz: low order harmonics probably not a huge concern here as our sensitivity to distortion is getting lower towards bass, but higher order harmonics would be projected into critical vocal range.

below 100 hz: THD rather irrelevant here and so is IMD. most harmonics won’t reach vocal range and thus are rather irrelevant if audible at all. instead the thing to watch out for here is spurious noises like port chuffing or any sort of weird suspension noises like leads slapping the spider or whatever that will be projected into the vocal range and above …

now i somewhat made up the actual numbers above but the general idea is that different frequency ranges have different amounts and type of distortion that can be tolerated. you can do your own research to find out exactly how much and what type of distortion is audible at what frequencies. what i proved above is sort of like a general overview - not meant to be a definitive reference.

so when we design our 3-Way + Subwoofer system we are really balancing physics tradeoffs of driver, cabinet and horn design against psychoacoustics characteristics of ear-brain mechanism at different frequencies …

i have to say that while very few people are intelligent enough to logically analyze a problem this complex the way i do it - many PROs with decades of experience in the game have an intuitive understanding of these things simply from listening to hundreds of speakers of various designs over the years. they found through experience what kind of performance can be expected from what kind of design. a great company like JBL will have both people with that sort of intuitive understanding as well as people with analytical design / engineering skills as well as a full suite of measurement tools and testing capabilities to optimize their designs.

the only problem is that once they put all that knowledge into a product like JBL VTX F35 they aren’t going to let you have it for cheap. they correctly figure that their expertise is worth something - namely your money and thus you will have to pay significantly more for the speaker than the sum total cost of the components and materials going into it.

DIYers on other hand mistakenly believe that they are saving money by building speakers at the cost of components and materials without paying for the design expertise. This is only true though if you can design ( or clone ) a speaker properly.

REMEMBER - the chain is only as strong as the weakest link. it can be the components, it can be the build quality or it can be the design itself. for the average speaker - every part of it is the weakest link - shit design, shit build quality and shit components. for something like JBL VTX F35 there are no weak links - everything is world class. and for the average DIY speaker the weakest link will be the design, while components and materials will all be high quality.

THE KEY THING TO UNDERSAND is you can’t compensate for one weak link in the chain by over-specifying other links. more specifically you can’t compensate for crap design by using expensive drivers. this is why real pros just buy JBL speakers and don’t fuck around.

knowing that as DIYers our weakest link will be the design we should be focusing on optimizing the design rather than finding the “best” components …

towards that end …

towards that end with a 3-Way + Sub system we’re really trying to keep distortion, compression, modulation and all that nasty stuff down low where it belongs.

both low in frequency and low to the ground physically.

from subwoofer to midrange every frequency band is designed to off-load mechanical and thermal stress from the frequency band above it.

the subwoofer will be very large and heavy and have very high moving mass and inductance and very large ports with prominent resonance is the cabinet, port, cone and suspension but all that crap will be kept down low in frequency where it is largely inaudible or irrelevant to music enjoyment …

and it will also be kept low to the ground physically where that 200 lbs subwoofer can’t tip over and fall on somebody …

next up the woofer will have a cone that weighs 1/2 as much and VC that weighs 1/4 as much with cone and suspension optimized less for durability and more for sonics …

woofer cabinet will be much smaller, with much smaller ports and more damping material - this will reduce resonances and colorations as well versus the subwoofer enclosure …

the lower weight of of woofer and its enclosure will allow it to be safely placed on top of the subwoofer without danger of maiming anybody if it tips over …

next up the midrange can be fully self contained “bucket” driver, like this:


at this point the moving mass is very low and voice coil and suspension mass is almost nonexistent even with a huge inside-outside 3" VC … because the wire is now very thin and the winding and former are very short and suspension is very light because it has almost nothing to do with the bucket rear enclosure controlling excursion to only a few millimeters …

with low VC mass, low inductance and virtually no cone displacement we are side-stepping most of the mechanisms that color the sound of large subwoofers such as induction modulation, suspension stiffness and so on …

that B&C SW woofer may have linear BL curve in the center … but it still has nonlinear suspension, still has asymmetrical inductance, still has a cone designed to withstand extreme forces rather than be free from coloration through breakup …

by contrast the midrange cone is made of felt paper that is soft almost like a silk dome tweeter - combined with a light weight VC it can be super smooth in the midrange because it doesn’t need to worry about creating huge forces over huge excursions …

by being flat to 3 khz and with no real breakup this midrange can comfortable mate to ANY tweeter ( short of a true ribbon ) thus allowing you to use any compression driver and horn combo you want - both saving you money and allowing you flexibility in cabinet design.

by contrast if you try to do a 15" 2-way you will most likely have to step up to a compression driver with 1.4" throat ( versus the standard 1" throat ) and that will impact off-axis response at high frequencies by introducing significant beaming.

with a 15" 2-way you will have to use a large and expensive compression driver, a large and expensive horn, a large cabinet to accommodate the large horn …

and your “reward” for all this will be high frequency breakup and beaming associated with the large Compression Driver and large diameter horn throat …

you may then ask - if this is really such a bad idea - why do companies like RCF make such speakers ?

the RCF TT 5A linked above uses a 4" ditanium dome crossed to 15" woofer at 650 hz … it is a very expensive speaker, even though it doesn’t really have subwoofer bass performance. so why would RCF make a speaker like this ?


to cut through the noise in a club you need extreme high frequency capability that only a large Compression Driver can offer. and once you have condemned yourself to using a $500+ compression driver you may as well try to recoup some of that investment by eliminating the midrange driver and saving some money that way.

in other words RCF HAS NO CHOICE. they can’t use a smaller compression driver in this speaker because it wouldn’t have enough clean output for the application.

but you’re building a home speaker - you don’t need an enormous 4" titanium dome CD. instead of a $500 compression driver with two part moving assembly ( titanium dome and Mylar suspension ), 5 slit phase plug, huge 4" motor and impossibly tight manufacturing tolerances … you can use a $200 compression driver with a simple 1.75" polymer ring diaphragm and plastic phase plug …

and put that $500 you saved towards a midrange …

well to be fair a 2-way plus sub using say a 2" CD and 12" woofer would be perfectly fine for home use.

i used that RCF TT 5-A as an example of how costs spiral out of control when you try to stretch a 2-way design past its comfort zone.

if you’re willing to accept slightly lower performance a 2-way design ( with subwoofer ) will be a very cost effective and elegant solution.

the only real problem with such a design is that there are already many such speakers you can buy at competitive prices and there is no real reason to build one as a DIYer.

you can buy a great sounding 12" 2-way plastic DSP box from JBL or RCF and then simply DIY a sub for it and call it a day, which is frankly what i would do.

you will get great performance and bang for the buck that way but you won’t get a lot of respect in the DIY community for taking such an easy way out.

the other problem with plastic speakers is you will have to turn it up to near deafening levels to make people understand what it’s about whereas with mahogany cabinets like Sonus Faber you won’t have to play anything at all in order to convince people that your speakers are expensive.

with a DIY 3-way + Subwoofer you would be able to create something interesting enough that people on DIY Audio and elsewhere will want to ask you questions about it and it will be a conversation starter in your room as well …

and as a bonus it will also have the ultimate performance …

and be uniquely yours …

whereas with a 2-way Full Range system your options will be so restrained that you will either have to build basically a clone of somebody else’s design that is known to work well or take a risk by stretching drivers to the edge of their limits and possibly end up with something that you won’t be 100% happy with …

if you do decide to build that 15" full-range 2-way anyway i will accept that choice of course - i am not trying to stop you. I just want to make sure you’re making an informed decision.

I was going to mention the amplifier as well. If you DIY a 3-way active speaker with DSP and separate amplification for each driver, the amplification and DSP ends up being the most expensive component of the system. If that is the case then spending a bit more on the best drivers doesn’t seem like a huge deal to me. Amplification/DSP for the one speaker already costs $2000, so why not spend $600 on the woofer instead of $300? It doesn’t make a huge difference in the bottom line. Real manufacturers have access to cheap amplification from Texas Instruments and such, but the DIY market is limited to very expensive finished amps from Powersoft, Crown, QSC or maybe boards from Hypex, Pascal, ICEPower, etc. There is no cheap solution to the amp problem.

My design was always a three way with a 15" woofer, a midrange, and a tweeter/compression driver. The woofer/subwoofer only has to play to like 400Hz.

I discovered Eighteen Sound also has an Air Motion Transform driver:


That’s interesting. I have never seen anything other than compression drivers in pro audio.

Measurements here:


It has lower distortion than any of the compression drivers I have seen. Very nice.

actually there is.

used Multi-Channel QSC Commercial amplifiers on Ebay.

these follow the same dynamic as used Dell servers. after 5 years you can buy a Dell Server on Ebay for $500 which was $20,000 new.

why ? because those servers are bought by companies like Google who don’t care about the price but they will ONLY get brand new equipment direct from manufacturer and will only use it with full warranty and support. once that’s out they dump it on Ebay and they don’t care what price they get for it.

the reason they don’t get higher price is because the only potential customers for such servers are other companies like Google that also have same policies of not using outdated out of warranty gear, thus they are forced to basically throw these still perfectly good servers out, selling them for peanuts.

certain types of installation amplifiers follow the same dynamic. the buyers on Ebay are looking for lightweight portable DJ amps to drive subs - they don’t want heavy install multichannel amps that are getting offloaded. as a result you can often buy a QSC amp that was $3,000 new for about $400.

typically what goes bad in QSC amps is the potentiometers. they did in mine. if you can’t replace them you can just short them so it’s always at 100% volume and control the volume elsewhere. in my case i just fiddle with it until it starts to work, but my speaker is passive - if directly connected to tweeter the noise made by these rusty potentiometers when you twist fiddle with them may blow the tweeter up.

essentially the only amps you will have to buy new are the subwoofer amps, because those are in such a high demand that you won’t find any deals on them used anyway.

so basically your highest power amp you have to shell out for because people are always looking for amps with more power for latest subs and the demand is greater than the supply, but all the lower powered multi-channel amps you can get for cheap because supply outstrips demand for those …

so your amplification cost will be mainly a function of how loud you need your subwoofers. all the other frequency bands will be powered at a reasonable cost so long as you aren’t locked by a speaker like JBL VTX F35 to a specific amplifier model because it is the only one that has the DSP for them … in which case you will spend an arm and a leg.

i’m going to now contradict myself and argue the opposite point of what i was arguing before.

that’s because i want you to see the cons and pros of both approaches.

by doing a 3-way you will have a discontinuous directivity / beamwidth versus frequency.

one gigantic horn like in that B&C prototype will have a consistent off-axis response with frequency and can be optimally matched to the woofer.

on other hand a midrange will throw a monkey wrench into that equation unless you load it in a waveguide, as the JBL VTX F35 does. then obviously you can adjust its beam width to match the woofer and tweeter, but at that point you’re looking at a fairly large and complex speaker.

for this reason if you don’t need maximum performance the 2-way may be the way to go. many of the budget 3-ways are actually worse than high-end 2-ways for this reason. some old-timers just want a 3-way and PA speaker makers are happy to oblige them by building a crappy 3-way. but it’s only the high end 3-ways like JBL VTX F35 that have a proper midrange waveguide that are actually a solid upgrade from a good 2-way.

what i’m trying to say is i would sooner go for a 2-way + Sub than a 3-way without a sub. 2-way + sub is the most common configuration for mobile DJ while 3-way without a sub is probably the least common. and i agree with this.

2-way plus sub is relatively simple and portable and works well. 3-way plus sub would also work well if done right but most aren’t done right because they want to skimp on the midrange waveguide to keep speaker small and cheap. a 3-way without a sub is in some sense the worst of all worlds because it would be hard to transport, hard to raise up to ear level and still not provide the kind of bass club goers expect.

as always, this is just food for thought - i am not trying to steer you one way or the other.

and you may not have the same level of bass craving as me or drunk crackheads at the club.

AMT in Pro Audio were popularized by Beyma, who has several models. Beyma is that company i mentioned that wrote a scientific paper on subharmonic distortion in tweeters. they probably just wrote it to promote their AMT tweeters.

the one from Eighteen Sound is very similar to the higher end models from Beyma.

aside from cost ( which is very high ) the other issue would be beaming. these are really meant as line array drivers but a line array of those would cost as much as a small car. in pro audio such prices may be acceptable if you’re covering a crowd of 10,000 people who all bought tickets … but if you’re listening alone at home that is way too expensive IMO.

and without a line array you would be left with basically a laser beam in a vertical domain at high frequencies. your ear would have to be at exactly the same level as tweeter. you would either be able to listen sitting down or standing up - but not both.

theoretically you could 3D print an acoustical lens that would widen the beam of the AMT by slowing down the sound at the upper and lower sides of the driver thus curving the wavefront to achieve wider vertical coverage … it would actually be a pretty cool DIY project.

same but opposite approach is used to narrow the beam of compression drivers when used in arrays.

there are actually multiple approaches to doing this. for example here is one:

the other one is V-Dosc:


they use different methods but accomplish the same thing - introduce delay at the center to turn a curved wavefront into a straight vertical one thus focusing a natural wide dispersion pattern into a horizontal laser beam. such laser beams are then arrayed for optimal coverage.

you can do the opposite and turn a line source ( AMT ) into a virtual point source by using similar approaches but in reverse …

this has been tried by some Ribbon drivers for example:

the foam on top and bottom is designed to create acoustical delay to widen vertical dispersion …

yes distortion of AMT is lower than that of compression. the price is higher.

but the real issue, as i said, is making their dispersion pattern work for your needs.

if ribbons ( whether true ribbons or AMTs ) didn’t have this issue with being neither line sources ( too short ) nor point sources ( too long ) they would be pretty much a no brainer for home high end.

for pro audio, which is all about line arrays you would think they would be jumping all over AMT that is so easy to turn into a line array … but compression drivers with a V-Dosc or other wave adapter can also be turned into a line source and Compression Drivers are more efficient and when competently designed like BMS will be lighter and cheaper due to not having to immerse the entire diaphragm in magnetic field but only the voice coil thus saving on neodymium ( cost ) and iron ( weight ).

to summarize the reason AMT is niche in prosound is cost and weight and slightly lower output than compression drivers, though actually the output of prosound AMTs is quite high - they are about 10db less efficient than CD but they handle maybe 2X to 4X more power thus partially compensating for lower efficiency …

and when you consider that CD are often limited by distortion rather than power handling that makes AMTs somewhat competitive with CD from perspective of output … but it will still be more cost and weigh than a BMS CD which are just super tiny and lightweight …


or maybe drunk people at rock concerts simply don’t have such stringent THD requirements as you …

if doing a line array you can use cheaper ribbons than Beyma or Eighteen Sound and still have high output at low distortion.

of course line arrays are inherently expensive and you would have to be absolutely 100% sure a point source speaker can’t possibly do the job before you step up to that kind of expense.

and if you’re doing a point source your life will be easier with a CD than AMT.

but yes CD does have higher distortion than AMT.

AMT when well made ( which Beyma is ) quite possibly has the lowest distortion of all types of HF drivers

but as i said distortion isn’t the only consideration …

I would probably just use the best tweeter I can find and set limiters.

I just want to build really good speakers. They aren’t for anything in particular. I just want them to be as good as I can make them.

1 Like

actually good limiters is one of the things that separate cheaper DSP from more expensive DSP so gain the $$$ savings may not really be there …

nothing wrong with that either. it’s a hobby so you get to define what it’s about.