How you control volume on this system? In software on PC side?
I´m volume control is in DAC. I have triple stereo Buffalo DAC, controled with modified Hifiduino software.
Anyway, back to your post...
Dephonica looks very interesting.
Please help me with a couple of specs on the website that don't make sense to me..
Low latency: about 50 msec when using IIR filters and crossover output stream into ASIO device
That seems like a very high latency for an IIR mode.
And this from the FAQ section: For example, FIR filter with 8000 taps adds about 250 msec of latency.
That equates to an 8000 tap delay using a 32k sampling rate, which seems unheard of?
What are they saying, you think?
What sampling rate(s) are available?
What are you using? How many taps?
Sampling rate available is anyone. Dephonica make the filters acording to sample rate on the fly. In my sistem, 96k is maximun sample rate without drops. I have a 4 core celeron at 1,7 Ghz, 35% CPU load. Maybe the problem is in the mother board data bus.
I use 48000 taps at 6 channels. Delay is about 1 sec. For video reproduction, is neccesary to use player with audio-video sincronization like VLC.
Regards
A fwiw aside, i would not use a plate amp.
Limits experimentation...
I'm not a fan of plate amps, though I'll need to consider how I work around amping the subs, especially if they end up in the middle of the room somewhere
All those subwoofer sound way overkill. If lucky, only two in total in necessary. Look here:
https://www.harman.com/documents/multsubs_0.pdf
Great shallow subwoofers is here:
Shallow Subwoofers – Sbacoustics
Maybe you find that you dont get better performance, just throwing subwoofers in the room.
Rule of thump in this hobby is that practice with things, makes perfect. You can prepare/read forever and then when you have spent 4 weeks building a cabinet og waiting 2 month on a driver, the result may just be mediocre.
Also, subwoofers are not a substitute for building small mains. Geddes seem to prefer using mains that do not go very low. In you room around 20 squaremeters, I would say that a 10" woofer is minimum. Better is 2*8", 12" or 15" 🙂
Even a 8" is a whole other thing than a 6" in perceived quality.
The harman document covers the Welti approach which I mentioned before, I'm just not sure how low the mains needs to go, could be an option though
DBA requires quite a few subs so I may not have much choice if I decide to do that, if I understand it correctly I should be able to build small mains with DBA.
Think I need to read more about Geddes to see how low the mains need to go, ideally I'd like to keep them above ~80Hz if possible.
I need to read up about options but my main concern with a PC build is making sure it doesn't introduce a lot of EMI/RFI into the system.PC solution is much more powerfull.
you might want to plan from the very beginning to be modular and upgradeable.
that's a key thing for me yes
You have a lot of budget available, which removes several hurdles, but there are many other challenges before you which are not funding related. After reading your posts, I think one of your biggest challenges is that you do not yet know what kind of loudspeaker you prefer. Or put another way, which attributes of speaker excellence are most important to you... Knowing this leads some people to pursue something like a JBL M2 or Gedlee Summa, while others pursue somthing like a Magico M3 or KEF blade. No one can tell you what your preference is, you must discover that.
I'm likely to not start building the mains until after lockdown is over and I can get out and listen to some different speakers.
I keep thinking though that it would be nice to build a few different types anyway. Let's see how badly the bug grabs me 🙂
I agree with hifijim 🙂
You need to try different things, like swapping drivers, filters, cabinets - more than ones - actually many times. Cause this will help you get way more in touch with what you like and prefer.
The fusion amp is a good little all-in-one package. And in the beginning, it will surely be your luck and hard work in understanding how to use it - that will set the limits.
Look at Kii3 - 70$ tweeter. Yep - but they did it right - for a lot of people to be thrilled with the performance.
So a lot of the burden to create world class sound.... lies on your shoulders - in choices - in skill... not necessarily in the this or that driver, amp or DSP.
makes sense, I guess I'm going to have to learn how to design cabinets at some point then!
in reality I'm a good 12-18months+ away from considering that, lots of other things to learn 1st.
I'll confess having looked at it things seem rather daunting,
Last edited:
hmm there seems to be conflicting advice on the range needed for the mains when doing the Geddes approach, Geddes "insists on full range", other people state that 80Hz makes more sense.
It seems strange to me to make the mains go below 80hz given the use of subs in the room...
It seems strange to me to make the mains go below 80hz given the use of subs in the room...
That is because he advice to make an overlap. So by simply having mains that are big enough to run down below 80-100hz... without a high-pass filter. Then you gain the advantage that there is less phase issues and your subs will blend more easily with the mains - without annoying suckout in the crossover frequency between subs and mains.hmm there seems to be conflicting advice on the range needed for the mains when doing the Geddes approach, Geddes "insists on full range", other people state that 80Hz makes more sense.
It seems strange to me to make the mains go below 80hz given the use of subs in the room...
An example at designing cabinets can be seen here:
Heissmann Acoustics | Loudspeaker | Heissmann-Acoustics
You can easily see the different shapes and driver choise - and then in the measurements - there's a clear difference between narrow, wide, waveguided or not.
All the good speakers - no matter brand or price. Usually have an even response and nice smooth power response. You can do this by having the drivers use the cabinet - more or less - as a waveguide. If the cabinet is narrow - you will get less support in the lower frequencies... the longer wavelengths simply weakens and "escape" around the cabinet instead of being forced toward the listener. The narrow cabinet is just more easy to design when it comes minimum effects(difractions) from the baffel edges.
All cabinets will need to take the beaming effect into account. A bigger driver will start to emit it's higher frequencies in an increasing narrowing pattern. So as soon as the tweeter takes over... it will be small in comparison to the frequency it plays... therefore being omni... playing in all directions - before starting to beam when it goes higher in frequency.
This is what oftens happens in poor designs. You get this narrowing/loss of energy in the cross over region... mostly the upper midrange - which sunds laid back and missing detail. Also the tweeter tend to sound detached and not combined with the midrange.
Heissmann Acoustics | Loudspeaker | Heissmann-Acoustics
You can easily see the different shapes and driver choise - and then in the measurements - there's a clear difference between narrow, wide, waveguided or not.
All the good speakers - no matter brand or price. Usually have an even response and nice smooth power response. You can do this by having the drivers use the cabinet - more or less - as a waveguide. If the cabinet is narrow - you will get less support in the lower frequencies... the longer wavelengths simply weakens and "escape" around the cabinet instead of being forced toward the listener. The narrow cabinet is just more easy to design when it comes minimum effects(difractions) from the baffel edges.
All cabinets will need to take the beaming effect into account. A bigger driver will start to emit it's higher frequencies in an increasing narrowing pattern. So as soon as the tweeter takes over... it will be small in comparison to the frequency it plays... therefore being omni... playing in all directions - before starting to beam when it goes higher in frequency.
This is what oftens happens in poor designs. You get this narrowing/loss of energy in the cross over region... mostly the upper midrange - which sunds laid back and missing detail. Also the tweeter tend to sound detached and not combined with the midrange.
That is because he advice to make an overlap. So by simply having mains that are big enough to run down below 80-100hz... without a high-pass filter. Then you gain the advantage that there is less phase issues and your subs will blend more easily with the mains - without annoying suckout in the crossover frequency between subs and mains.
isn't that advice from before DSP was in common use i.e. a way to manage the integration in a simpler way, just let the mains roll off naturally.
I'm not getting how this reduces phase issues, I'd have thought it would increase them as you now have another LF source which you need to time align.
I'm confused!
I can't seem to find any recommendation on how wide the overlap should be either.
An example at designing cabinets can be seen here:
Heissmann Acoustics | Loudspeaker | Heissmann-Acoustics
You can easily see the different shapes and driver choise - and then in the measurements - there's a clear difference between narrow, wide, waveguided or not.
All the good speakers - no matter brand or price. Usually have an even response and nice smooth power response. You can do this by having the drivers use the cabinet - more or less - as a waveguide. If the cabinet is narrow - you will get less support in the lower frequencies... the longer wavelengths simply weakens and "escape" around the cabinet instead of being forced toward the listener. The narrow cabinet is just more easy to design when it comes minimum effects(difractions) from the baffel edges.
All cabinets will need to take the beaming effect into account. A bigger driver will start to emit it's higher frequencies in an increasing narrowing pattern. So as soon as the tweeter takes over... it will be small in comparison to the frequency it plays... therefore being omni... playing in all directions - before starting to beam when it goes higher in frequency.
This is what oftens happens in poor designs. You get this narrowing/loss of energy in the cross over region... mostly the upper midrange - which sunds laid back and missing detail. Also the tweeter tend to sound detached and not combined with the midrange.
thanks for the explanation, I was aware of the beaming effect, but not the waveguide point, I've ordered the loudspeaker cookbook, think it's worth getting a better understanding of the design choices.
It means you send to Buffalo DAC volume control commands by I2C from Hifiduino?I´m volume control is in DAC. I have triple stereo Buffalo DAC, controled with modified Hifiduino software.
isn't that advice from before DSP was in common use i.e. a way to manage the integration in a simpler way, just let the mains roll off naturally.
I'm not getting how this reduces phase issues, I'd have thought it would increase them as you now have another LF source which you need to time align.
I'm confused!
I can't seem to find any recommendation on how wide the overlap should be either.
It's no trickery. The overlap might need tuning. Nothing is fixed, since it also depends on your room and where you put the mains and subwooferes.
Below around 150hz you are having sound waves so long, that they behave like a complete mix with the room. You use phase to allign the response from each subwoofer to create a total response that as a whole sounds like one. The woofer in the mains should be in synergy with the tweeters and midrange. The point in making the woofers run down with no filter - or 2-order. Is to get a best possible integration between the world of subwoofers(omni sound) and the world of more direct sound(mains).
The mains need to be placed in an order to get good stereo. Then the subs are placed to take advantage of the room to optain an even an smooth response. now these to principles need to "find" each other. And by letting the mains run off slowly and the subwoofers have each of their own filter. You can get a smooth cross over - blend.
You see. The subwooferes interacts with each other to create an even soundfield - adding and subtracting from each other - this is good. But when going from subwoofers to mains - we slowly become directional - wavelengts become shorter. So this is why a smooth transition is needed, so we do not get suckout that we cant fill up. Cause as a rule of thumb. We rarely boost, cause we very quickly run into power limitations and stress on drivers. A suckout is a destrcutive interference between waves... so you can actually keep putting energy into this cancellation and never get the result you wish for. This is why individual phase and summed response is a very important matter.
It means you send to Buffalo DAC volume control commands by I2C from Hifiduino?
And does this mean that you're not using a pre-amp at all then?
I've tried directly connecting a DAC/streamer with volume control into a power amp in the past but it sounded far too harsh to my ears, not sure if this was just the make & model. I like the idea as a way to simplify the chain, not sure if it's workable though, would be keen to hear peoples thoughts.
It's no trickery. The overlap might need tuning. Nothing is fixed, since it also depends on your room and where you put the mains and subwooferes.
Below around 150hz you are having sound waves so long, that they behave like a complete mix with the room. You use phase to allign the response from each subwoofer to create a total response that as a whole sounds like one. The woofer in the mains should be in synergy with the tweeters and midrange. The point in making the woofers run down with no filter - or 2-order. Is to get a best possible integration between the world of subwoofers(omni sound) and the world of more direct sound(mains).
The mains need to be placed in an order to get good stereo. Then the subs are placed to take advantage of the room to optain an even an smooth response. now these to principles need to "find" each other. And by letting the mains run off slowly and the subwoofers have each of their own filter. You can get a smooth cross over - blend.
You see. The subwooferes interacts with each other to create an even soundfield - adding and subtracting from each other - this is good. But when going from subwoofers to mains - we slowly become directional - wavelengts become shorter. So this is why a smooth transition is needed, so we do not get suckout that we cant fill up. Cause as a rule of thumb. We rarely boost, cause we very quickly run into power limitations and stress on drivers. A suckout is a destrcutive interference between waves... so you can actually keep putting energy into this cancellation and never get the result you wish for. This is why individual phase and summed response is a very important matter.
kind of you to explain, that makes a lot more sense now. I think I was forgetting the point about frequencies slowly becoming directional.
I'm still not sure how to pick the right overlap though, appreciate it's room dependant - I guess I more concerned with making sure that the mains have the right drivers in them and can go low enough.
maybe as already mentioned what I need to do is try a few different things out including the different sub integration strategies.
I've got a set of bookshelves & 2 different main speakers each with different ranges, and an existing sub that I could experiment with.
think everything is pointing at building a few subs as a 1st step.
overlapping mains is just so there are more LF sources which results in smoother bass in the room. It would be better to take the LF off the mains and have two more subs as the subs will play bass notes with lower distortion. I would just build mains that go down to 80Hz and route all bass to subs. From my experience of running sound systems main speakers sound much cleaner when not having to produce bass notes.
isn't that advice from before DSP was in common use i.e. a way to manage the integration in a simpler way, just let the mains roll off naturally.
I'm not getting how this reduces phase issues, I'd have thought it would increase them as you now have another LF source which you need to time align.
I'm confused!
I can't seem to find any recommendation on how wide the overlap should be either.
thanks for the explanation, I was aware of the beaming effect, but not the waveguide point, I've ordered the loudspeaker cookbook, think it's worth getting a better understanding of the design choices.
overlapping mains is just so there are more LF sources which results in smoother bass in the room. It would be better to take the LF off the mains and have two more subs as the subs will play bass notes with lower distortion. I would just build mains that go down to 80Hz and route all bass to subs. From my experience of running sound systems main speakers sound much cleaner when not having to produce bass notes.
I agree. This was what I tried to explain - sorry if I got to explainy 😱
I cross my mains too.... they can play pretty deep. But I relieve them from the burden by giving them a low order filter from around 80hz. By low-order - I mean I kinda imitate a natural rool-off from a driver in a closed box, but I do it by filter to let the subwoofers do the hardest work. Also - closed boxed are easier to work with.
overlapping mains is just so there are more LF sources which results in smoother bass in the room. It would be better to take the LF off the mains and have two more subs as the subs will play bass notes with lower distortion. I would just build mains that go down to 80Hz and route all bass to subs. From my experience of running sound systems main speakers sound much cleaner when not having to produce bass notes.
I'm a lot happier with 80Hz, think that would give me more freedom to try different mains.
I'm still not sure which of the bass integration approaches I'm going to try yet. I very much like DBA but can't really seem to fine designs for wall mounted subs.
Geddes seems to be the most likely approach, mainly because it will integrate well with my current main speakers.
I'm starting to look at tools as well, I've got access to a fully equipped woodshop however getting into it might be an issue during lockdown. Trying to figure out what the bare minimum tools would be to make a sub so I can avoid having to sink too much money into tools.
Just by having to mains that play what is now possible... then add two subs - one in a corner close to one of the mains and one on the oppisite side wall. Then you play all of them together and find your rooms resonance - just measure in you listening position. Then you start to simply dampen that resonance with EQ.
Now you have a very rough start to what multi-subs can do. Then you slowly build on top of this, by tuning and measuring until you get a better and deeper understanding of what is going on. Dont worry about all the fanzy details and all the beautiful curves you see everywhere. The key is to get started and finding out how it can work in your room.
Now you have a very rough start to what multi-subs can do. Then you slowly build on top of this, by tuning and measuring until you get a better and deeper understanding of what is going on. Dont worry about all the fanzy details and all the beautiful curves you see everywhere. The key is to get started and finding out how it can work in your room.
If you have the correct room shape DBA obtains almost perfect results. For wall mounted subs I don't think there are any existing designs but I would make simple ported or sealed box subs. You can adjust the form factor of an existing sealed sub design you only need to keep the internal volume the same. For a ported box design you also need to keep the port dimensions constant (length and diameter). Also using Hornresp to design sealed and ported subs is easy, so just give it a go.
If you can cope with the boxes having a bit more depth I would use standard drivers as I have never seen as good kipple results from shallow mount drivers and they are expensive. If you can deal with having to use a lot of EQ to get low and amplifiers capable of large output voltage swing then pro audio drivers* are a good choice (cost effective, well behaved at high excursion). There are obviously higher end options but something like 8*18TBX100 in a DBA would be pretty cool.
*They will be more efficient (as in less real power dissipated for given output) but will require more voltage than a Hi-Fi driver in parts of their frequency range. This is because their motor is more powerful and their moving mass minimized which pushes up their self resonant frequency, the result of this is a box with a non flat response that requires EQ but with greater output potential.
If you can cope with the boxes having a bit more depth I would use standard drivers as I have never seen as good kipple results from shallow mount drivers and they are expensive. If you can deal with having to use a lot of EQ to get low and amplifiers capable of large output voltage swing then pro audio drivers* are a good choice (cost effective, well behaved at high excursion). There are obviously higher end options but something like 8*18TBX100 in a DBA would be pretty cool.
*They will be more efficient (as in less real power dissipated for given output) but will require more voltage than a Hi-Fi driver in parts of their frequency range. This is because their motor is more powerful and their moving mass minimized which pushes up their self resonant frequency, the result of this is a box with a non flat response that requires EQ but with greater output potential.
Last edited:
Just by having to mains that play what is now possible... then add two subs - one in a corner close to one of the mains and one on the oppisite side wall. Then you play all of them together and find your rooms resonance - just measure in you listening position. Then you start to simply dampen that resonance with EQ.
Now you have a very rough start to what multi-subs can do. Then you slowly build on top of this, by tuning and measuring until you get a better and deeper understanding of what is going on. Dont worry about all the fanzy details and all the beautiful curves you see everywhere. The key is to get started and finding out how it can work in your room.
yes I'm pretty much on page with giving it a go to be honest, I'll quickly have 2 subs to play around with.
think my focus now is moving to DSP, need to figure out what to do there, I've posted something in the PC subforum as not quite sure if I should try self build given people's comments earlier in this thread.
https://www.diyaudio.com/forums/pc-based/367226-shelf-versus-pc-dac-streamer-dsp.html#post6515470
If you have the correct room shape DBA obtains almost perfect results. For wall mounted subs I don't think there are any existing designs but I would make simple ported or sealed box subs. You can adjust the form factor of an existing sealed sub design you only need to keep the internal volume the same. For a ported box design you also need to keep the port dimensions constant (length and diameter). Also using Hornresp to design sealed and ported subs is easy, so just give it a go
Room is 5.38m x 3.33m so I think should be ok, the challenge is that it has a fireplace with alcoves at either side. Unfortunately DBA wants to position the SUBs on the chimney breast, I can't place them in the alcove or I could have built the wall out level and wall mounted the subs quite neatly. Not sure if there's any latitude with DBA to mount the subs nearer to the sideways than the ideal?
I guess I could still level the walls off and mount them in the correct place.
if I just need to maintain the same volume then that might give me some more options, will have a play around with Hornresp.
The other thought I had was actually to build horns into the alcoves, I don't know enough about horns to know if I could then get the sound to output in the right location though - need to do some research
Another potential issue for DBA though is that on 1 of the side walls I have a recessed single pane french door - I'm contemplating replacing this with a sound studio type door instead to help with acoustic symmetry anyway, but not sure if this would be needs for DBA in it's own right.
If you can cope with the boxes having a bit more depth I would use standard drivers as I have never seen as good kipple results from shallow mount drivers and they are expensive. If you can deal with having to use a lot of EQ to get low and amplifiers capable of large output voltage swing then pro audio drivers* are a good choice (cost effective, well behaved at high excursion). There are obviously higher end options but something like 8*18TBX100 in a DBA would be pretty cool.
*They will be more efficient (as in less real power dissipated for given output) but will require more voltage than a Hi-Fi driver in parts of their frequency range. This is because their motor is more powerful and their moving mass minimized which pushes up their self resonant frequency, the result of this is a box with a non flat response that requires EQ but with greater output potential.
I'm not too worried about needing lots of power or using standard drivers rather than shallow if it gives the right result, quality's the key thing.
B&C 18TBX100's look fairly reasonable, certainly might be able to stretch to something higher end though.
I'm working my way through various DBA threads, think I need to figure out what's viable and just give it a go! I might start with Geddes 1st though as I'll only need to build a couple of subs, that might give me a good way to learn.
decisions decisions.....
Can only reiterate that Big subwoofers are not a substitut to tiny mails ie. one 6” per side.
Before building anything new, I would test with what you got. For trial you can easily test with subs, using a smaller woofer in a cabinet - or one of your 2-ways, with suitable filtering. One of Geddes arguments is that you don’t need to use the same kinds of subs.
Before building anything new, I would test with what you got. For trial you can easily test with subs, using a smaller woofer in a cabinet - or one of your 2-ways, with suitable filtering. One of Geddes arguments is that you don’t need to use the same kinds of subs.
Agree, but it is only for testing. Would be sad if the best position in practice is under the couch, and the enclosure can’t fit there.
Many, including myself, have found the BMS 18n862 to be a great driver.
That looks an interesting driver.
I think I might need to do more research. I'm a bit unsure how much of my budget I should be putting into sub drivers as a %.
and it doesn't help not knowing if I'm going Geddes or DBA given there's a big difference in the number of drivers.
I'm wondering if starting with Geddes and putting more money into higher quality drivers might be a better place to start, and then if I ever do want to try DBA I can do it over a long time period using the more expensive drivers.
Can only reiterate that Big subwoofers are not a substitut to tiny mails ie. one 6” per side.
Before building anything new, I would test with what you got. For trial you can easily test with subs, using a smaller woofer in a cabinet - or one of your 2-ways, with suitable filtering. One of Geddes arguments is that you don’t need to use the same kinds of subs.
How do you mean you don't need to use the same kind of subs? just for the testing you mean?
Agree, but it is only for testing. Would be sad if the best position in practice is under the couch, and the enclosure can’t fit there.
I'll be in a dedicated listening room so I have more freedom than most to place the subs where I want. hopefully not an issue.
- Home
- Loudspeakers
- Multi-Way
- High-end full active DSP 2.1