Exploring High-End Component Upgrades for DIY Active Speakers with DSP

These are the design notes I used to build up my woofer filter. I am all ears as to the best way to accomplish this with a product/software which allows the filters to be built up graphically and independent of any measurements. I would prefer to do full room correction as a separately implemented pass once I can measure the output against the designer's reference calibration file.

Woofer:
LPF2=300Hz, Q=0.85 (use 2 of these to get 4th order cutoff)
HPF2=20Hz, Q=1.5 (very much your choice as to how to trim this, with vinyl playback need to be a bit careful)
LPF1=30Hz (puts a tilt into the response mainly for baffle step)
Boost ?? Probably ballpark 10dB required
*I don't think a separate baffle step function was used but I'm not 100% certain so try to solve it without it; there isn't one in my tube amp version

Additional footnotes provided to me later...

Woofer lowpass should be built with two B2 filters. Add a PEQ at the same point with Q=1 and +1 to +2 dB.
Woofer tilt is first order lowpass.
Woofer highpass, select B2 at about 20 to 25Hz. Add a PEQ at same frequency, maybe Q=1, +3-6dB
Add one more PEQ if needed for balancing the 20-100Hz range but you might not need it - be persistent with everything else first.
 
Danville looks promising but the 4 channel dac version not available yet.
So unlike you said, it is available.

Danville has 3 different Nexus products listed the 8, 4 and 2 channel versions. They first appeared on the scene in 2020/2021 (though been making Air Traffic Control product for decades). Only now in late 2023 and 2024 have they been promoting early adopter specials.

Does the purchase of a Danville unit come with a limited user license for diy purposes? Is there no recurring annual fees to continue to use the program.
Yes it does come with a license of audio weaver. Al Clark will even help you install it and setup the parameters for your needs. Actually lets include him in this discussion @danville
 
Last edited:
There are various options:

From fully DIY’ed ADC/DSP/DAC solutions. Here’s one from @5th element

1722636023082.png

Reference:
http://www.5een.co.uk/DSP 2.htm

Then there are ready built solutions from American Danville Signal, Australian DEQX, and Hong Kong SAR miniDSP.
The pro ones have been around for a long time, including from dbx, Lake Processing, Soundcraft etc.

Then there's re-using/recycling a commodity personal computer and multichannel audio interface and some software. Or one can use a Raspberry Pi and the HATs

For anyone interested, the Nexus was reviewed in the most recent Voice Coil (August 2024) magazine! Nice!

Rather than get a paralysis from (over)analysis, I would suggest that the OP find an option (s)he is comfortable with, and dive in and start using and having fun!

Remember, making mistakes is how you learn! So ask questions and the community can help!
 
Last edited:
  • Like
Reactions: MtlJazz
Hi OP,

In particular, I disagree with your entire premise that all A/AB amps sound better than all Class D amplifiers. They certainly tend to be more expensive though.

My own experience is that I like some class D a lot more than I like some Class A amps which cost 10x-100x more. I've gone from A/B to Class D and back to A/B as my mood and opportunity suited me. Right now I'm running a Luxman integrateds side by side with a Hypex powered center channel. I'm pretty sure my next speaker build will be going away from linear amps, and more towards all-in-one solutions. While I really really like my Luxman's sound quality I find myself more listening to movies than music and some of the glamour is wasted on me. If I did want to add more tube/liquid sound back I'd probably try a nice tube preamp if I could find one with HT bypass.

Best,

Erik
 
My setup is active with multiple stereo DACs and class D amps.
I use VST DSP to handle the crossover/EQ settings on Windows/Mac.
  • RME Digiface USB outputs 2x TOSLINK to 2 stereo DACs.
  • 2 stereo DACs: one to ICEpower for the woofer and the other to Purifi for mid/high. The mid/high part uses a passive crossover. Perhaps I'll change to full active in the future if I have more money.
So, it's a hybrid active/passive system.
With the RME Digiface, you can choose whatever high-end stereo DACs you like, as long as TOSLINK isn't a problem for you.
If you're not bound by TOSLINK, you can even choose some multichannel I2S output board from miniDSP or DIYINHK. Or if you want to use AES, you can go for the MOTU 112D.

I don't think class A/AB sounds better than modern properly implemented class D. Give it a chance, whether by listening to a friend's setup or trying from a vendor that offers a money-back guarantee.
 

Exploring High-End Component Upgrades for DIY Active Speakers with DSP​


People that consider DSP to be worthwhile in closely controlling transfer functions are likely to lean towards technical performance and away from high-end performance. This makes it difficult, though not impossible, to get them interested in expensive high-end hardware that doesn't increase technical performance in a significant way. I think you will find that a few decades back class AB was often used for tweeter amplfier modules and sometimes midrange ones when typical class D amplifier modules were not audibly neutral at higher frequencies. As class D design evolved this became less the case. Here and here are a couple of papers by Genelec discussing amplifier modules for speakers. Note how they are discussed largely as commodities rather than something that can bring value and help distinguish Genelec speakers from the competition.

At present people that are receptive to high-end components are far more likely to look for it with exotic components for passive crossovers rather than active crossovers. The few high-end enthusiasts that have made the move to active crossovers are in turn more likely to be interested in analogue rather than digital active crossovers. Given the technical advantages are with DSP active crossovers (and the costs to manufactuers though currently less so for DIY hobbyists) it is only a question time. It is coming but what form will DSP active crossovers take that successfully attract high-end enthusiasts? Class A?, valves?, excessively high sampling rates?, separate exotic cases to house the electronics with exotic cabling, shelving, isolation devices, spikes?,... Don't know but when enough high-end enthusiasts want DSP active crossovers we will find out.
 
I want to thank you all for contributing to this thread. I learnt a lot.

I have the MiniDSP 2x4HD for my subs and I love the ease of use.
The rest of my system are high-end with snake oil power cables, Holo Audio May KTE day, PS Audio BHK 250 power amp and the list goes on.
I can absolutely hear the difference in cables and amps. That's why I am reluctant to by a minidsp for my active speaker project.

Here is what I would like to have.
I have built a music pc running roon and Dirac. What I would like is a program like CamillaDSP to site after Dirac.
CamillaDSP looks like an okay product but it's a one man show and not easy to use. So, if one of you has a suggestion for a program that can replace CamillaDSP, then I am all ears.

I am thinking of using the okto 8 pro dac. But many in this thread has suggested to look at the professional world. So, what's out there?
My music pc is running windows 11 because of Dirac.
 
I have built a music pc running roon and Dirac. What I would like is a program like CamillaDSP to site after Dirac.
CamillaDSP looks like an okay product but it's a one man show and not easy to use. So, if one of you has a suggestion for a program that can replace CamillaDSP, then I am all ears.

Given this is a DIY forum and camilladsp is a small one man open source program that is close to what you want there is an obvious suggestion...
 
@ jjasniew

There was a huge problem with these speakers, because they were only marketed to be smaller, putting out the same bass like a larger, passive speaker. All the argument was "small makes noise like large". Something that customers interested in HIFI, at that time, did not care about. All what Philips raved about was "give up your large speakers, your wife will love it". By the way, the same mistake is repeated by the people still building them today. Instead of building something better, they still focus on size. Maybe a thing more important in the Netherlands?

The Philips management, as allways, was too arrogant to look at the high end market, they wanted large numbers and to replace the competitions speakers and amplifiers at the same time. Instead of looking at demand, they decided what people had to buy. I was told they also didn't give the dealership a fair share of the retail price. If you look at what is left of Philips today, you will see that constant arrogant, stupid decisson making have shrunken it from a once giant company to a dwarf, selling of any branch they were unable to run profitable.

I was very interested in these speakers, because we had an uncle working at Philips Hamburg who could have given us all parts to build them for free. We could not find a single pair of these speakers on working dispaly in the whole 2 million city of Hamburg. The dealership usualy did not want to display them, even for the reason they did not fit the usual switch board used to compare passive speaker or just not having wall mains in place.
Also, at that time, speaker size was only a problem if they were too small! People upgraded, downsizing wasn't invented.
So the resume about them was always "impressive bass output for the size". Just like the rest of the frequency spectrum did not exist or count.
The price advantage to have build in amplifiers was none, as at that time, a preamp was more expensive than a good integrated amp. So you had to buy some high end preamp to drive the medium quality active speaker, build from low price Philips OEM chassis. I got to mention, even as they were quite cheap, no one who did serious DIYS speaker building used any Phillips OEM chassis. The sound was not really impressing and reputation low, even as modified versions could be found in some more "HIFI" speakers.
I did not even touch the point of the ugly, cheap design of the Phillips speaker range! I don't think they spent even a single Gulden for a designer. Was done for free, during lunch brake, from a trainee who happened to have a pencil in his jacket. The grill was made from some left over curtain cloth. The board of directors had rejected it as too cheap for their private dining room.
.
So if we talk high end, the Philips active doesn't match the bill at all. Even the most high end amp would not have improved much.

On the other hand, Backes&Müller, who build only full size, fully active speaker, some as high as the room, where very succesfull! So it was not the principle that was not accepted, but the Philips product focusing on small size and ignoring sound quality.
 
I can say with full confidence that concentrating on some modest room treatments along with the MiniDSP on board dacs fed a digital signal with Class D or ICE Amps is all you really need along with a measurement mic and the knowledge to use it with software……..expensive A AB or tube amps are simply a waste of money and current. How you design your speaker system and enclosures along with how you filter them ( this is where the mic and software comes in) is where the greatest gains are. Some diffusion here, some absorption there…….proper placement and you’re done.

Get a DSP unit with 8 outs and 8 mono amps…..eliminating cross talk is a huge step towards clarity…….plenty of current on tap to deliver a solid dynamic range is another.

Get the bass right from the get go and the rest falls into place when the room isn’t resonating like a huge enclosure. You don’t need expensive and large bass traps……..four subs tuned and your good to go…..if you can clean up 30-100hz you’ve already won more than half of the battle.
 
Motion control is very tricky except for the subwoofers (simply tricky). I failed at repeatably attaching a piezo sensor with sufficiently high Fs to a voice coil former with an archery glue (high young module, capable of sustaining high G) at DIY conditions, not talking of anything else. Ideally, it would help if you used it inside an adaptive DSP-controlled loop to adjust for varying Fs, dumping, etc, which is another trouble. Unfortunately, it is of limited benefit for lowering distortions at higher frequencies. The current drive is much simpler and more effective.

I have a question. Suppose, you planned a base 3-way DSP project, with miniDSP, Dayton reference series speakers, 6 channels of TPA3255 amps, just enough power supply, and basic woodwork. Now, you suddenly have an extra $1000 to spend on it. What would you consider most beneficial:
1. Faster DSP (more MACS/FLOPS)
2. Better DAC/ADC (SINAD)
3. Better amps (THD+N)
4. Bigger power supply
5. "Better" Cables
6. Better drivers
7. More drivers (going to a line source vs point source)
8. Looks
9. Room treatment
10. Something else
?
 
Have a look at the Linea Research ASC-48 for a 4-in 8-out DSP processor with good ADC and DACs as well as AES I/O. I have the Danley branded equivalent with FIR option and am very happy with its performance. I use a MiniDSP SHD Studio as a digital pre-amplifier and Dirac processor. System is stereo active 3-way with powered subs in stereo.

The MiniDSP is OK for its role.

You can find the ASC-48 for under $2000 used. (i.e not inexpensive)
 
  • Like
Reactions: MartinAndersen
If there is money left after DSP and amps, the drivers should get a close look. What you spoil at this point will be impossible to regain with something else on the list. Even if I see the DSP thingy with mixed emotions. I woudn't touch a minDSP, but that's me.
Improvements near to being not audible may cost large. Same to amps. If no one screwed up, hard to beat a sub 100$ TPA3255 with matching 50$ power supply or some PA amps.

Depending on your room there may be some room treatment neccesary. Something that integrates into the room, not huge barrels to store bass and drink it later.
Not any room makes room treatment a must, there are many rooms that are quite OK the way they are. Anything inside your listening room is "treatment", I think of carpets, curtains and sofa's etc. If you live inside a fancy loft, with just a few seats bare walls, windows and floor, things look different.

Not as a last I would put looks. If you are a grown up, you may not like optics that show your inability to build decend furniture like speaker cabinets or the lack of funds to buy them. No need to go all "Bubinga Root Elephant Rose Nut", just some sanding, a coat of primer and cheap self adhesive foil turn your ugly speaker shag into something more civilized. At least from normal distance. The "I do it later excuse" doesn't count any more, it ran out in 1999. For a majority a well finished speaker cabinet is a proof of sonic quality.

The small rest of the list may be nice to think about but not sound improving.
 
Motion control is very tricky except for the subwoofers (simply tricky). I failed at repeatably attaching a piezo sensor with sufficiently high Fs to a voice coil former with an archery glue (high young module, capable of sustaining high G) at DIY conditions, not talking of anything else. Ideally, it would help if you used it inside an adaptive DSP-controlled loop to adjust for varying Fs, dumping, etc, which is another trouble. Unfortunately, it is of limited benefit for lowering distortions at higher frequencies. The current drive is much simpler and more effective.

I have a question. Suppose, you planned a base 3-way DSP project, with miniDSP, Dayton reference series speakers, 6 channels of TPA3255 amps, just enough power supply, and basic woodwork. Now, you suddenly have an extra $1000 to spend on it. What would you consider most beneficial:
1. Faster DSP (more MACS/FLOPS)
2. Better DAC/ADC (SINAD)
3. Better amps (THD+N)
4. Bigger power supply
5. "Better" Cables
6. Better drivers
7. More drivers (going to a line source vs point source)
8. Looks
9. Room treatment
10. Something else
?
10. Something else. Although it may kind of fit in #1, too.
An open-architecture DSP like Q-Sys, Symmetrix, Biamp, Xilica, etc. The open-architecture part is key. (and for me the processors' FIR capability)
Used on ebay for under $1000 easy.
Expands experimentation and tuning capability ten-fold over conventional DSPs, like the minidsp's and ASC-48, that i also own and use.

Imso, better multi-way tuning execution far exceeds better electric gear.

This is assuming there is a good acoustic speaker design to work with, and all the drivers are fit for the design goals. Otherwise, I'd spend here first.
 
Imso, better multi-way tuning execution far exceeds better electric gear.
This is only possible if you were able to find the optimal way to obtain the frequency response and its correction. There are many subtleties and nuances in setting up a multi-band active speaker system in a room.

But yes, if you understand the basic principles of setting up a multi-band speaker system, then as a result, no serial speaker system will be able to surpass your well-tuned active system, and the secret here is quite simple, a serial speaker system does not take into account your listening conditions, and the room affects the perception of sound quite significantly, and with good correction of the room's influence on sound, the sound quality level becomes noticeably better.
 
  • Like
Reactions: MtlJazz
Pro-audio Open Architecture DSP boxes are expensive. Yes, the latest use faster DSPs, the latest and greatest codecs, and more of them. Old ones are weaker than miniDSP. They are not sold to end customers, UI is ... cryptic at best. There will be no customer support. Buying a SigmaDSP/Studio-enabled eval board from ADI may be a far better option (IMHO), In any case, you have to know what you are buying and have training & experience.