I have played around with Equalizer APO for trying out a digital XO for my speakers. But now when my new speakers are getting closer to being finished I need a good audio source for them. With my old speakers with passive filters I use optical out to my amplifier which has a DAC that does the job. But that option wont do with active filters since I need two amplifiers. So I am looking into which solution is the best one, at a budget.
Shall I go for a MiniDSP 2x4 HD? Seems simple enough to set up, get some room correction and FIR-filters. But I've heard that sound quality can't match a good soundcard or good DAC. Over here that will set me back 300€, and close to 100€ more if I go with an UMIK-1.
The other and basically free alternative is to make filters in Equalizer APO, as I understand it you can do phase correction that way too. But for that I need a new sound card, which puts me almost in the same price range as the MiniDSP-solution anyway. And maybe a mic aswell.
Which will perform the best? Will it be alot easier to get things working with the MiniDSP? Is software + soundcard inferior? I won't use any other audio source than my computer so I wont be held back by a software solution in that aspect.
I am reluctant to settle for old fashioned passive filters since I want to play around and modify my speakers. And as I understand it digital filters will for the most part get me better results, at least with my level of knowledge in filter design.
The whole reason I'm into DIY to begin with is to keep costs low. So I wont just throw money at the problem until it gets right. I want to get as close to my goal as I possibly can at my first try. :/
Shall I go for a MiniDSP 2x4 HD? Seems simple enough to set up, get some room correction and FIR-filters. But I've heard that sound quality can't match a good soundcard or good DAC. Over here that will set me back 300€, and close to 100€ more if I go with an UMIK-1.
The other and basically free alternative is to make filters in Equalizer APO, as I understand it you can do phase correction that way too. But for that I need a new sound card, which puts me almost in the same price range as the MiniDSP-solution anyway. And maybe a mic aswell.
Which will perform the best? Will it be alot easier to get things working with the MiniDSP? Is software + soundcard inferior? I won't use any other audio source than my computer so I wont be held back by a software solution in that aspect.
I am reluctant to settle for old fashioned passive filters since I want to play around and modify my speakers. And as I understand it digital filters will for the most part get me better results, at least with my level of knowledge in filter design.
The whole reason I'm into DIY to begin with is to keep costs low. So I wont just throw money at the problem until it gets right. I want to get as close to my goal as I possibly can at my first try. :/
I've done both myself. Started with minidsp but had issues with device not recognized by PC randomly to the point I gave up. Then I used eapo for about a year but eventually went back to another 2x4hd that I've had zero issues with and prefer sound quality of vs the behringer ucd204hd I was using with eapo.
If you want to use fir filters eapo will support a lot more taps than the minidsp. Other than that I think it comes down to dac preference with any reasonable config.
Eapo would sometimes glitch out too which is why I ended up going back to the minidsp. while my first minidsp must have been bad the new one has been just fine.
I've written software to automatically generated 2 way crossovers and peqs in fir for eapo. I hated the user interfaces for that so I essentially wrote my own. That way I could change things just be moving sliders and it would automatically make a new fir impulse and wire it into eapo. Worked really well and I miss it with the minidsp.
If you want to use fir filters eapo will support a lot more taps than the minidsp. Other than that I think it comes down to dac preference with any reasonable config.
Eapo would sometimes glitch out too which is why I ended up going back to the minidsp. while my first minidsp must have been bad the new one has been just fine.
I've written software to automatically generated 2 way crossovers and peqs in fir for eapo. I hated the user interfaces for that so I essentially wrote my own. That way I could change things just be moving sliders and it would automatically make a new fir impulse and wire it into eapo. Worked really well and I miss it with the minidsp.
Bi-amping or tri-amping with good sound quality is an expensive proposition, it seems to me. To keep it low cost, there are going to have to compromises. The digital filtering part is pretty cheap and easy, and for the most part so is room correction (better not to go overboard with room correction, usually sounds better if kept pretty simple). The expense comes with dacs and power amps. Maybe the best way to do a dac would be to use something like ES9038PRO to make a good 8-channel dac (so all the channels can work in perfect sync with each other). TP seems to be working on one, but their dacs may not be all that low cost by time time all is said and done. Then there are power amps. If the power amps aren't up to the same standards as the dacs, then sound quality will most likely suffer. Chip amps might work okay and can be fairly low cost, except maybe for bass. A really good implementation would seem like it ought to be able to do bass pretty well, but some people still seem to express some disappointment. Maybe a Hypex switcher would be a better choice LF, or maybe for everything.
Myself, I gave some thought to going with active crossovers, but decided it would be too expensive given the sound quality I want. I ended up rebuilding the passive crossovers in my speakers with better components, and added stereo sealed subs. All the speakers are mounted on Primacoustic Recoil Stabilizers (which help a lot in a way that DSP can't do). I tuned the LP filters in the subs using a mic and software RTA to smooth out overall FR as much as possible and left it at that. How good does it sound as is? Easily good enough to mix hit records, I think would be a reasonable claim. The one stereo dac is very good and so is the one stereo power amp. It ended up addressing a number of things that helped SQ, and that would still need doing even if electronic crossovers were used. I feel like I spent my budget where it would do the most good and left it at that.
Myself, I gave some thought to going with active crossovers, but decided it would be too expensive given the sound quality I want. I ended up rebuilding the passive crossovers in my speakers with better components, and added stereo sealed subs. All the speakers are mounted on Primacoustic Recoil Stabilizers (which help a lot in a way that DSP can't do). I tuned the LP filters in the subs using a mic and software RTA to smooth out overall FR as much as possible and left it at that. How good does it sound as is? Easily good enough to mix hit records, I think would be a reasonable claim. The one stereo dac is very good and so is the one stereo power amp. It ended up addressing a number of things that helped SQ, and that would still need doing even if electronic crossovers were used. I feel like I spent my budget where it would do the most good and left it at that.
Last edited:
In what way did Eapo glitch out so seriously that you ended up with MiniDSP again? Do you think Eapo with anything better that the Behringer would rival or surpass the MiniDSP.
Nice about the software, I guess it wasn't really fit for distribution? 😛
Yeah it might get expensive, but as it happens I have some decent amplifiers lying around already. Vincent SV-234 should be enough for my needs. So there won't be any additional costs there. The ES9038PRO looks really nice, but it is far from affordable. Looks like 1000-2000$+ in the worst cases. Found one from HiFime thats quite afforable, but that will only give me 2 channels. You know of any other solutions with 4 channels or more? Are there any others to recommend in your opinion?
When it comes to room correction I don't intend to overdo it.
Last edited:
Four channels are probably doable with one of these: ES9028PRO ES9018 32bit Audio DAC PCB - DIYINHK
The connect the 8 channels into groups of 2 for processing by the output stage. Unfortunately, they only brought out one stereo pair of those. But, they did parallel the other channels into two sets. It is probably possible to attach small wires to the traces for the extra 2 channels with a very small solder iron tip, this gauge solder, and some thin wire, maybe #30 or smaller. I could post a pic of where the extra channel wires would have to be attached, if needed.
Perhaps one nice thing about that board is that I have gone through the exercise of sketching out the schematic, which I could share (although it could use some cleaning up 🙂 ). If interested I would recommend to get the board with the ES9028PRO installed. The rest of the components you have to diy, but the board is labeled with parts values for everything including a stereo output stage. An output stage for the extra two channels would probably have to be outboard. An Arduino or MCU of your choice can setup the dac registers, although it will run in a basic default mode with no MCU. Diyinhk also sells a volume control board that may be able to control it, don't know. They also have a more complete and new design ES9038 PRO board, but it costs more. May include more parts though.
You would also need a multi-channel (4 or more channel) USB to I2S board, or perhaps two stereo of SPDIF feeds (I would have to check on the SPDIF). The it would work with I2S is there would be one BCLK, and one LRCK, and a DATA1 and a DATA2 for the two stereo pairs.
The connect the 8 channels into groups of 2 for processing by the output stage. Unfortunately, they only brought out one stereo pair of those. But, they did parallel the other channels into two sets. It is probably possible to attach small wires to the traces for the extra 2 channels with a very small solder iron tip, this gauge solder, and some thin wire, maybe #30 or smaller. I could post a pic of where the extra channel wires would have to be attached, if needed.
Perhaps one nice thing about that board is that I have gone through the exercise of sketching out the schematic, which I could share (although it could use some cleaning up 🙂 ). If interested I would recommend to get the board with the ES9028PRO installed. The rest of the components you have to diy, but the board is labeled with parts values for everything including a stereo output stage. An output stage for the extra two channels would probably have to be outboard. An Arduino or MCU of your choice can setup the dac registers, although it will run in a basic default mode with no MCU. Diyinhk also sells a volume control board that may be able to control it, don't know. They also have a more complete and new design ES9038 PRO board, but it costs more. May include more parts though.
You would also need a multi-channel (4 or more channel) USB to I2S board, or perhaps two stereo of SPDIF feeds (I would have to check on the SPDIF). The it would work with I2S is there would be one BCLK, and one LRCK, and a DATA1 and a DATA2 for the two stereo pairs.
Last edited:
The miniDSP HD should exceed the specs of an ideal 16 bit DA easily, it should also have lower harmonic distortion and better SINAD than the amplifiers your using. The main thing to be careful of using such DSPs is preserving this high performance by avoiding degrading the signal to noise ratio.
For example if a shelving filter is used to extend the bass of a sealed box speaker by an octave (+12dB boost) this will be implemented by the DSP as a reduction in level of all the higher frequencies. As the level is reduced by 12dB this means that the signal amplitude is now 12dB closer to the noise.
Another problem is if the gain of the power amplifiers is too high and digital volume control is used, the digital volume again works by reducing signal amplitude with the presence of fixed noise amplitude. This also applies if you are using the DSP to sensitivity match drivers rather than adjusting their amplifier gains.
In my case I need a lot of volume adjustment range and quick setup of my system so am in the process of constructing a 8 channel balanced volume controller. I am also implementing some of my equalisation using analogue circuits inside the same unit to avoid wide band cuts.
One thing that might be worth considering for an easy all in one solution are Hypex plate amps that have DSP and amps built in:
Hypex Electronics B.V.
their amplifiers measure very well:
Review and Measurements of Hypex NC400 DIY Amp | Audio Science Review (ASR) Forum
For example if a shelving filter is used to extend the bass of a sealed box speaker by an octave (+12dB boost) this will be implemented by the DSP as a reduction in level of all the higher frequencies. As the level is reduced by 12dB this means that the signal amplitude is now 12dB closer to the noise.
Another problem is if the gain of the power amplifiers is too high and digital volume control is used, the digital volume again works by reducing signal amplitude with the presence of fixed noise amplitude. This also applies if you are using the DSP to sensitivity match drivers rather than adjusting their amplifier gains.
In my case I need a lot of volume adjustment range and quick setup of my system so am in the process of constructing a 8 channel balanced volume controller. I am also implementing some of my equalisation using analogue circuits inside the same unit to avoid wide band cuts.
One thing that might be worth considering for an easy all in one solution are Hypex plate amps that have DSP and amps built in:
Hypex Electronics B.V.
their amplifiers measure very well:
Review and Measurements of Hypex NC400 DIY Amp | Audio Science Review (ASR) Forum
By glitch I mean it would stop working. I'd only get signal to the main outputs (woofers) and nothing on the tweeters. This usually happened after win10 updates so perhaps eapo isn't solely to blame here but it seems the way it registers itself as an apo is goofy. I'd have to sometimes adjust registry permissions and such to get it working again. These issues are documented in various forums. I bought a minidsp for a second system and after messing with eapo one time too many decided to just try it as I was trying to decide where to go next and found I had none of the issues I had with the first minidsp.
The behringer is quiet and very good sounding. I lived with it for about two years and when I say I prefer the minidsp it's very slight and maybe just my imagination. But i think j found the behringer casts a bigger wider picture but maybe lacks a little detail. The minidsp seems a bit flatter and less 3d but more detail.
There's other options too. Eapo isn't the only game in town. Frankly I think the benefits of eq far outweigh the details and compromises of how you do it and the dacs you use. I cant speak for how a better dac would sound with eapo but the issues I had there were not the dac or the drivers. It was eapo. Lots of people seem to have no issues with eapo but some do if you do some research. I also spent a year with an rpi solution but that was too unstable for my tastes too although that was definitely more about the hats drivers and firmware than the rpi or the moode software. The minidsp has been solid for me, knock on wood and I don't think you'd regret going that way assuming you get a solid unit.
Also I have a really sensitive horn system 106db lf and even more sensitive on the hf. I have no noise or hum from the minidsp unless my ear as at the mouth of the hf horn. The behringer was definitely a bit hissier and I could hear it a little further away.
Last edited:
Hi Borotech,
Any reason why you did not use the pace gui for eapo? Is your code on any repository?
My av pc is dedicated and I run the media players/ audio player in game mode ( kernel prioritizes audio and video over background apps). I have not had many glitches with eapo, other than the occasional lock up, that happened mostly when I was updating/ testing vs. the "production" phases. But I never worked the filters as hard as you.
My new project is a econowave ( two way horn and sub) with active filters. I was planning to use eapo for filters and rew room correction, by the means of HDMI output and multichanel receiver. Do you believe eapo is ok or I should go with minidsp ( I have one classic on a case with two sure stereo amps).
Tks in advance, F.
Any reason why you did not use the pace gui for eapo? Is your code on any repository?
My av pc is dedicated and I run the media players/ audio player in game mode ( kernel prioritizes audio and video over background apps). I have not had many glitches with eapo, other than the occasional lock up, that happened mostly when I was updating/ testing vs. the "production" phases. But I never worked the filters as hard as you.
My new project is a econowave ( two way horn and sub) with active filters. I was planning to use eapo for filters and rew room correction, by the means of HDMI output and multichanel receiver. Do you believe eapo is ok or I should go with minidsp ( I have one classic on a case with two sure stereo amps).
Tks in advance, F.
I”m looking into something similar for a 3way (or possibly 4way) active. I’ve looked into three alternatives.
1) minidsp 4x10HD, all in one box but you are stuck with the inbuilt dac chip
2) equilizer app with peace front end plus 7.1 sound card (cheap option ASU’s U7)
3) minidsp nanodigi + 4 DACs + passive control, dac options are second hand Lucid 88192, four topping d30/d50s, four khadas tone boards. To this I would add ladder attenuators (eizz?) so I can keep the music digital at full bit (no digital volume issues) and then keep the volume control analogue. The nanodigi is digital in and out so let’s you chose your dac.
For amplification I have got two Rotel RMB1066 which can run 6x60w or brig pairs for 150w, these are fairly cheap in the U.K. around £200 second hand.
Just a few thoughts as the minidsp nanodigi is an option not discussed yet.
1) minidsp 4x10HD, all in one box but you are stuck with the inbuilt dac chip
2) equilizer app with peace front end plus 7.1 sound card (cheap option ASU’s U7)
3) minidsp nanodigi + 4 DACs + passive control, dac options are second hand Lucid 88192, four topping d30/d50s, four khadas tone boards. To this I would add ladder attenuators (eizz?) so I can keep the music digital at full bit (no digital volume issues) and then keep the volume control analogue. The nanodigi is digital in and out so let’s you chose your dac.
For amplification I have got two Rotel RMB1066 which can run 6x60w or brig pairs for 150w, these are fairly cheap in the U.K. around £200 second hand.
Just a few thoughts as the minidsp nanodigi is an option not discussed yet.
The minidigi is discontinued but if you have one is a great option. There is an implementation of the lx 5.2.1 with the minidigi blessed by linkwitz. Also, i would not use the sound card as an old recreceiver ( or a nice new one) can give you 7 to 9 channels with more flexibility and quality than a sound card. Its a perfect way to test ideas and configurations before you step un to a more advanced settings. That is the path in going to follow before i decide what dsp dac i will use nlon the final product.
Linkwitz states that all amps that power all chanels must have the same gain. Easy way is to get 6 or 8 of the same type or a multichannel amp. But Nelson Pass stated about the lx mini analog crossover / amps that they decided to use 100 Watts for the woofer and 25W for the hf second driver.
Please keep the thread alive with your build. Really curious to see the progress of your implementation.
Linkwitz states that all amps that power all chanels must have the same gain. Easy way is to get 6 or 8 of the same type or a multichannel amp. But Nelson Pass stated about the lx mini analog crossover / amps that they decided to use 100 Watts for the woofer and 25W for the hf second driver.
Please keep the thread alive with your build. Really curious to see the progress of your implementation.
Last edited:
Check the gigaport ex
8 channel unbalanced out , usb driven with sinad overb115 dB for $200, seems too good to be true.
8 channel unbalanced out , usb driven with sinad overb115 dB for $200, seems too good to be true.
In a recent rant Linus Torvalds stated that many runtime errors could be traced to the lack of availability of ecc memory on the intel platform. Glitching on eapo may traced to that. New amd costumer grade chips
support ecc as all server chips. Some good deals can be found on old stock chips
support ecc as all server chips. Some good deals can be found on old stock chips
In my case, I choses the following:
- External sound blaster xonar U7. Pretty good sound quality.
- 4 estéreo amplifiers. AIYIMA A04 TPA3251 175W * 2.
I would said, good quality for the price. 40€ each at AliExpress.
- Umik-1 microphone + REW + rephrase + deep testing for deep knowledge of what you are doing.
- Three way speaker without passive crossover.
- Active DSP with EAPO. Many tweaks for 3.1 system. Two floors handing, active 3 way + central + sub.
Never got any problem with EAPO crashes. I got an improvement over passive crossover and I am very happy with the final result.
Now I am designing Nautilus shape 3 way speaker for 3D printing, but that is story for another day.
- External sound blaster xonar U7. Pretty good sound quality.
- 4 estéreo amplifiers. AIYIMA A04 TPA3251 175W * 2.
I would said, good quality for the price. 40€ each at AliExpress.
- Umik-1 microphone + REW + rephrase + deep testing for deep knowledge of what you are doing.
- Three way speaker without passive crossover.
- Active DSP with EAPO. Many tweaks for 3.1 system. Two floors handing, active 3 way + central + sub.
Never got any problem with EAPO crashes. I got an improvement over passive crossover and I am very happy with the final result.
Now I am designing Nautilus shape 3 way speaker for 3D printing, but that is story for another day.