• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Help me decide what speaker to build :)

Meander

New Member
Joined
Jun 5, 2024
Messages
4
Likes
1
Hi i want to build studio monitors for mixing, mastering and also just listening to music. A few things are already set. They will be 3 way with seperate closed boxes for bass and mid and i will use a hypex fa253 for dsp and amplification. The hypex plates can do fir filters since the latest update, so with some tweaking a fully linear phase system could be possible. Source will be my cheapo 2nd gen 2i4 Focusrite interface but I will upgrade to an RME with AES outputs in the not too distant future to have one less adda conversion. The speakers will stand on my desk and i will be listening in the nearfield 1m before i move to a bigger space someday. I want to make the cabinets from mdf or multiplex and they will be rectangular boxes, no fancy shape. Max spl is not super important around 90-100dB without detectable distortion is enough. I do want them to go very deep, down to 25-30Hz which should be possible with eq at the cost of max spl.

I want to use the 10" seas l26roy woofer, probably the 4layer vc version, which has lower qtc so that i can make the cabinet small and boost the bass with a linkwitz transform. The question is what midrange and tweeter i will use. There is the seas kingroy mk3 kit which uses a coaxial mid tweeter combo and has nice documenation for the cabinet and ready made dsp files for the hypex plateamp. Measurements look good aswell and it would be cheaper than the alternative. Coaxial is nice for nearfield listening which will be my usecase for at least the next year or so. The drivers cost about 1600€

The other version would be to use really nice individual drivers. Purifi 5.25" midrange and sb satori beryllium tweeter with waveguide, both with 8ohm versions so that the dacs in the hypex plateamp have to ouput more signal and thus have slightly better resolution. The smallest possible center to center distance between those two drivers is 16cm which is about 2150Hz in wavelength. The tweeter should be able to be crossed below that, how low depends on how much distortion i can tolerate so i will have to experiment. The purifi starts to beam around 1400Hz so I will try to cross there to have nice wide dispersion at all frequencies. 1/4 Wavelength of the driver spacing between mid and woofer suggests 440Hz as the max crossover point. Another question is the mid tweeter position on the baffle. Should i offset them from the center to minimize diffraction and/or use foam on the baffle? The drivers in this configuration cost 2320€.

What do you think, are the extra 720€ worth it for the nicer drivers? I guess in the midfield the more expensive version will sound better? Do my plans make sense at all? What are your thoughts on an atc style dome mid instead of the purifi?
 
Last edited:
The Seas coaxial has somewhat uneven tweeter response. I would go with a KEF coaxial, a couple of their patented features like the tangerine waveguide significantly improve performance. A pair of refurb Q150 is $350. One could also make the argument of purchasing a pair of refurb KEF LS50 Meta for $1200 and integrate them as a 3 way with no physical modification needed. Just place them on top of your woofer modules and slap on a 300Hz digital high-pass.

I personally wouldn't buy the Purifi mid + SB Beryllium tweeter for this application, you just won't be generating the excursion/SPL to merit the expense. Even cheap drivers like a FaitalPro 4FE42 + SB Acoustics SB26ADC will sound amazing at 95-100dB.

ATC style 2-3" mid-dome results in extremely wide dispersion. Great for doing client showcases in a large studio, not people typically use for near-field mixing.

In my experience you won't need to use a driver offset or baffle foam if the tweeter has a waveguide, it eliminates enough baffle diffraction already.
 
Purifi 5.25" midrange and sb satori beryllium tweeter with waveguide,
This would be my suggestion for your use case and budget, esp. because you want low distortion. As you probably know you've picked some of the lowest distortion drivers out there, just go for it.

Another question is the mid tweeter position on the baffle. Should i offset them from the center to minimize diffraction and/or use foam on the baffle?
If you are spending this much on drivers, spend a little more on a router / bit so you can round over the corners. Or pay to get the baffle CNC'd in a 3D shape with good roundovers (you can integrate the waveguide at the same time.) Putting foam on the baffle doesn't really work from what I've seen.
 
Hi i want to build studio monitors for mixing, mastering and also just listening to music.
For "real monitors" (audio production), I'd recommend buying monitors because they are a "known-quantity". If you build them yourself you may get some strange/unknown characteristics that make it hard to get productions that "translate" well to everybody else's system. And if you end-up with a problem it could be difficult and/or expensive to fix.

...I've never heard them myself but the JBL LSR308 is highly rated and about $700 a pair. But it's only got an 8-inch woofer so you may want a sub. (There are several other very-popular 8-inch monitors in the same price range.)

There is the seas kingroy mk3 kit which uses a coaxial mid tweeter combo and has nice documenation
A kit will help if there are reliable measurements so you know what you'll end-up with.

If you go DIY, at a minimum I recommend getting some speaker design software (WinISD is free). It will allow you to optimize the woofer & cabinet and it will model & predict (low frequency) performance before you buy or build anything. It's probably worthwhile running the driver & box in the kit through the software too. You may want to change something.

There is other software for other aspects of the design but I've never used it. In any case, it's very important to get the woofer & box "working together".

Should i offset them from the center to minimize diffraction and/or use foam on the baffle? The drivers in this configuration cost 2320€.
"Traditional" vertical alignment minimizes horizontal interference as you move left-to-right in the room. Head/ear height doesn't vary as much with different people or between sitting and standing so it's less of a problem. If the speakers are pointed toward you and you're not moving around, it probably doesn't make much difference how the drivers are arranged.
 
Yes, there is a good case to be made that for $2500+ you can get Genelec or Neumann plus subs and not have to wonder.

However, with drivers like this, I've seen some builds on DIYAudio.com that rival or beat their performance. It can be done.

So OP, if you think building speakers is a fun project I wouldn't discourage you. But if you just want to get to the part where you are mixing good music, it's probably smarter to get something off the shelf.
 
Thanks for the advice :)
For "real monitors" (audio production), I'd recommend buying monitors because they are a "known-quantity". If you build them yourself you may get some strange/unknown characteristics that make it hard to get productions that "translate" well to everybody else's system. And if you end-up with a problem it could be difficult and/or expensive to fix.

...I've never heard them myself but the JBL LSR308 is highly rated and about $700 a pair. But it's only got an 8-inch woofer so you may want a sub. (There are several other very-popular 8-inch monitors in the same price range.)


A kit will help if there are reliable measurements so you know what you'll end-up with.

If you go DIY, at a minimum I recommend getting some speaker design software (WinISD is free). It will allow you to optimize the woofer & cabinet and it will model & predict (low frequency) performance before you buy or build anything. It's probably worthwhile running the driver & box in the kit through the software too. You may want to change something.

There is other software for other aspects of the design but I've never used it. In any case, it's very important to get the woofer & box "working together".


"Traditional" vertical alignment minimizes horizontal interference as you move left-to-right in the room. Head/ear height doesn't vary as much with different people or between sitting and standing so it's less of a problem. If the speakers are pointed toward you and you're not moving around, it probably doesn't make much difference how the drivers are arranged.
Using speakers that other producers also use definetely makes sense, but i also like building stuff man its hard to decide. I already have Yamaha HS-8 speakers so the jbls would be quite similar in SQ i think. I want something really hiend, similar to the kh310 or amphion one25a. I have Win isd and a measurement microphone and i have build speakers before. Nothing this serious, my previous builds were both fullrange bassreflex speakers with visaton bg20 and tangband 2145.
The Seas coaxial has somewhat uneven tweeter response. I would go with a KEF coaxial, a couple of their patented features like the tangerine waveguide significantly improve performance. A pair of refurb Q150 is $350. One could also make the argument of purchasing a pair of refurb KEF LS50 Meta for $1200 and integrate them as a 3 way with no physical modification needed. Just place them on top of your woofer modules and slap on a 300Hz digital high-pass.

I personally wouldn't buy the Purifi mid + SB Beryllium tweeter for this application, you just won't be generating the excursion/SPL to merit the expense. Even cheap drivers like a FaitalPro 4FE42 + SB Acoustics SB26ADC will sound amazing at 95-100dB.

ATC style 2-3" mid-dome results in extremely wide dispersion. Great for doing client showcases in a large studio, not people typically use for near-field mixing.

In my experience you won't need to use a driver offset or baffle foam if the tweeter has a waveguide, it eliminates enough baffle diffraction already.
I know in the documentation they say the response is the smoothest at a listening angle of 15°. Do you suggest using the KEF as the monitor or just using the coaxial driver? I cant find any data on the drivers alone which makes designing a cabinet harder.

This would be my suggestion for your use case and budget, esp. because you want low distortion. As you probably know you've picked some of the lowest distortion drivers out there, just go for it.


If you are spending this much on drivers, spend a little more on a router / bit so you can round over the corners. Or pay to get the baffle CNC'd in a 3D shape with good roundovers (you can integrate the waveguide at the same time.) Putting foam on the baffle doesn't really work from what I've seen.
Nice too hear that the driver selection makes sense for what i want to do. How big should the radius of the roundovers be? I have a router and always make the edges round when i build something out of wood but the biggest radius i have is like 5 or 6mm. Does that help or is it just for looks? Im also a little concerned about the woofer surround. Its really big will that cause reflections and interference?

Yes, there is a good case to be made that for $2500+ you can get Genelec or Neumann plus subs and not have to wonder.

However, with drivers like this, I've seen some builds on DIYAudio.com that rival or beat their performance. It can be done.

So OP, if you think building speakers is a fun project I wouldn't discourage you. But if you just want to get to the part where you are mixing good music, it's probably smarter to get something off the shelf.
Good Point. Neumann KH310 cost 2800€ used when you get a good deal, the sealed 750 dsp sub costs 1500€ + another 250€ for their custom measurement mic which i would get if i go down this route. So 4550€ for 2xKH310 + 1x sealed sub + Mic. I dont really want genelecs because they are not sealed. I know they sound good, there is a pair of the ones in a studio where I study. But I prefer the sound of closed speakers. My "hihend" speaker costs 2320€ for drivers + 1160 hypex + 200€ for cables, damping material, wood and paint if i make them myself. So about 3680€ but also a few weekends of time and probably hours of measuring and fiddling with crossover settings....
 
Hi,
it's gonna be long journey!:) Build what you think is good, then measure the system, all drivers individually (see VituixCAD measurement manual), use VituixCAD to figure out the crossover stuff, implement, measure the whole thing to confirm. At this point you've got million more questions, and a chance to develop it further which likely means build another prototype. Rinse and repeat until quality you seek is met.

How big should the radius of the roundovers be?
In general, as big as possible, but as big as fits is about as good. So, if you have 5mm radius roundover router bit only, then do not make the enclosure extend more than 5mm from the driver chassis ;) In other words, make your box only ~5.25" + 1cm wide for the woofer, and even smaller for the tweeter assuming it's the typical ~4", if you make it bigger, then figure out if you can slant / roundover the edges, starting right beside the driver/waveguide. Think the roundovers or slants are there for minimizing flat area around the transducer(s) as rule of thumb and now you can come up with a plan that you can implement with your tools and aligns with your project goal. You can find more about this stuff if you want, I and others have written a lot about this stuff with examples on forums. You can play with VituixCAD diffraction tool, or any other tools to simulate edge diffraction to build intuition how it plays out and come up with the same generalization.

However, this generalization doesn't take account what kind of a system you are building, so there is some leeway of course. For example, in my opinion tweeter (waveguide) should not be on a baffle but a separate freestanding waveguide for optimal edge diffraction performance, but this increases c-c distance between tweeter and mid, which affects lobing at crossorver, which might be annoying for near field listening context, depending how much you move vertically as you work, how listening distance and room acoustics are and so on. Overall, plan the speakers for your room and for how you work, optimize the speakers for yourself. That's the only reason to build speakers in the first place, tailor fit. It's a lot of work, but a lot of fun time as well so better just get started and evolve as you go. The drivers seem nice, and if you have a DSP, there is a lot of opportunities to arrange the system to a physical structure, if your first one is a let down for some reason, you could just build better one. Have fun!
 
Last edited:
Hi

Welcome to ASR.
I understand the call to build the speaker yourself. The seduction is in the thought that for the same cash outlay, you could end up with a product that is superior to commercial offering that cost multiple time... but ...Real costs are another matter. Not only time, but material. You may need to measure your final product carefully and that may take time, research and knowledge...
You seem to like the Neumann KH310. I'd go for these and then supplement these with a pair of DIY subwoofers, using the best drivers you can afford. Subwoofers are relatively straightforward to build. There are software that would predict their output relatively well, and measuring subwoofers do not require extensive knowledge, time and material (as in instruments) as a "full range" speaker. Subwoofers are relatively easy to DIY, especially if you have the woodworking skills and tools.
Then you use a DSP to integrate the subwooferS, as in multiple, at least 2, to the excellent K-310, relieving these from bass duty, thereby lowering their already low THD, perhaps increasing their SPL output.
It will probably take some time to integrate the subs with the KH-310, correctly. This will require DSP, their are several solutions, I am a fan of miniDSP. If you go that route, at the end you could have a superb system that measures supremely well, say 20 to 20,000 Hz within ± 2 dB or less, is capable of elevated output within the audible spectrum and allow you to enjoy music and work at producing it, to the most extracting standards of quality. The cash outlay would also be lower as well as the fuss and the wait. IMHO.

Peace.
 
People always fall for the same illusion. "Any idiot can DIYS build a speaker, so I'm highly qualified." "I just take the best chassis speak of mouth say's there are and end up with something "hiend".
Now, step back a little and listen to a guy in the Homedepot saying just that. Is your first idea, "that is a real genius" or do you thend to think "what an idiot"?
Is picking two or three drivers and a plate amp DSP all you need to build world class monitors? If it was that easy, why do most people pay so much for commercial "hiend" speaker? Is it possible that in reality it is not that easy?
Maybe think about it.
A professional speaker developer once told me "If things go well I need less than a week for a two way. If I'm really, really lucky, a three way takes me 4 month."
Which was a guy who studied this stuff on a university and had decades experience, wrote a few books too.

What you may be able to build, is a closed two way, with an f3 of 80 Hz and combine it with a subwoofer. This makes your monitor smaller (=closed) and gives the tweeter and midwoofer a well sized baffle. No one says that you have to use only one sub, you may do two, use thema as stands, place it near the monitor top, maybe even in one cabinet.

Sorry if I hurt your feelings.
 
Last edited:
I won’t say it can’t be done but usual a GOOD monitor design takes a long way to go.

Even reputable monitor manufacturers go from a simulation to a first prototype. Then measuring its behavior in a dead chamber, take the inputs for the creation of a next prototype ….. until in the end it meets all design criteria. Only then it gets released for mass production and sale.

You expect to achieve the same from a first self created setup? If I were standing in your shoes, since you appreciate the KH-310, I would go for that. You don’t only buy a physical KH-310 but also lots of built-in design knowledge that goes along with it.

… and if you really must make something of your own, either now or lateron in time, then go for something that's less critical: a subwoofer.


Cheers,
JaapD.
 
Last edited:
I don't know what means "hiend" for you. If you are not focus on pricy stuff why don't you go with this:


Cut it to one or two subs and you will have something realy good.

Add a PR (like done in a post) if you want to decrease disto at higher SPL (probably waste of Time and money for a nearfield listening.) but in this configuration measurments show that you will probably be near of a KH-310.

Plus you win some place on your desk and have the flexibility to put subs where it sounds best in you room ...
 
A professional speaker developer once told me "If things go well I need less than a week for a two way. If I'm really, really lucky, a three way takes me 4 month."
I had help from a pro with more credentials and exerience than I can ever dream of, and my three way took me/us over 18 months.
 
I have to say, this is the first DIY forum I've come across where people try and talk someone OUT of building their own speakers... :confused:

There are definitely some cons to DIY, but in my experience at least, SQ isn't one of them. An active system using the drivers you are thinking about is going to sound pretty sweet. 3-ways have always been considered a lot harder to build, but that's more because of how much more complex a passive crossover gets when adding another driver. DSP solves all that. I've got a 4-way running as my main system, and it wasn't really any harder to build than a 2-way. You just spend more on amps. There's no reason to be scared of more drivers as long as you know what you're doing, and if you're already converting c-to-c distances into frequencies to figure out crossovers, it feels to me like you generally know what's up. I'd jump in and go for it.
 
but that's more because of how much more complex a passive crossover gets when adding another driver.

3 ways are hard because it's hard to get accurate data down low where the mid to woofer crossover will be.
 
Here's a model of your L26RO4Y in a 11.6L sealed enclosure with a LT to boost the bottom. You'll run out of xmax with around 165W input and hit about 108dB

The dashed blue line is in an equivalent vented enclosure. I can provide tuning if desired.
1724199191796.png
 
I have to say, this is the first DIY forum I've come across where people try and talk someone OUT of building their own speakers... :confused:
It is not a DIY forum.


There are definitely some cons to DIY, but in my experience at least, SQ isn't one of them. An active system using the drivers you are thinking about is going to sound pretty sweet. 3-ways have always been considered a lot harder to build, but that's more because of how much more complex a passive crossover gets when adding another driver. DSP solves all that. I've got a 4-way running as my main system, and it wasn't really any harder to build than a 2-way. You just spend more on amps. There's no reason to be scared of more drivers as long as you know what you're doing, and if you're already converting c-to-c distances into frequencies to figure out crossovers, it feels to me like you generally know what's up. I'd jump in and go for it.
It certainly seems easier/faster when having a digital XO.
 
I have to say, this is the first DIY forum I've come across where people try and talk someone OUT of building their own speakers... :confused:

There are definitely some cons to DIY, but in my experience at least, SQ isn't one of them. An active system using the drivers you are thinking about is going to sound pretty sweet. 3-ways have always been considered a lot harder to build, but that's more because of how much more complex a passive crossover gets when adding another driver. DSP solves all that. I've got a 4-way running as my main system, and it wasn't really any harder to build than a 2-way. You just spend more on amps. There's no reason to be scared of more drivers as long as you know what you're doing, and if you're already converting c-to-c distances into frequencies to figure out crossovers, it feels to me like you generally know what's up. I'd jump in and go for it.
Haha, im surprised aswell. And yes my thoughts exactly, without digital dsp i probably wouldnt bother to try this. Imagine soldering in and out resistors/capacitors for every eq change.... I will give updates when i get around to building something. Probably in november.
 
Back
Top Bottom