Maybe, but if I don't try, I'll never know. A lot of people discouraged me to build line arrays, it would become a huge comb filter mess. That didn't work out so bad...
I only need to figure out if mms is as important as it's often said, on here and elsewhere. I "get" the idea behind that theory and even agree that when you move it back and forth it's probably a good idea to have as low a weight as possible.
If I had no restrictions I would go for SD size over excursion, probably meaning line arrays. But I'm very restricted and have to make the compromise work.
I can really recommend these Dayton drivers:
https://www.parts-express.com/dayton-audio-rss390ho-4-15-reference-ho-subwoofer-4-ohm--295-469
I use them in a relative small enclosure (for a 15" driiver) they deliver high quality deep bas.
I never really liked the XLS series but liked the XXLS series.
I am a fan of closed enclosures for subwoofers - Even though they do not seem to go as deep as bass reflex on paper, DSP is not needed for the low end at all.
And they integrate much better acoustically in most rooms where room gain can be a pain.
https://www.parts-express.com/dayton-audio-rss390ho-4-15-reference-ho-subwoofer-4-ohm--295-469
I use them in a relative small enclosure (for a 15" driiver) they deliver high quality deep bas.
I never really liked the XLS series but liked the XXLS series.
I am a fan of closed enclosures for subwoofers - Even though they do not seem to go as deep as bass reflex on paper, DSP is not needed for the low end at all.
And they integrate much better acoustically in most rooms where room gain can be a pain.
Indeed 15" is going to be too big. I did take a look at the 12" variant from Dayton. However, the Peerless XXLS and Scan Speak 30W do attract me more.
If you compare the XLS with the XXLS there are notable differences.
The one thing that I regret seeing is the venting in both the XXLS and 30W happens trough the dust cap on the cone.
The more expensive 32W has a vented pole piece. I bet the Beolab 90's 10" woofers aren't standard 26W drivers, who knows.
I will use DSP, no doubt you have seen the graphs I posted. That wouldn't have been possible without DSP. To better my current results I just have to use it. The one thing I fear is the sub position will bring more trouble than I can solve. The Arrays sort of smooth out any ill effect by their tall vertical dimension. The subs will be point source and will even face a table partially.
I see a lot of people mentioning the low end isn't that important. What if you were to get the (lower) frequencies to behave, run it way down to at least 30 Hz (or lower) and gradually roll them off from there, preferably without much group delay.
Listen to your favourite songs. I doubt you'd want to be without that feeling ever again. It's just too much fun to really have that low end after you rid yourself of the effects of room modes.
Measure your room behaviour with the free demo of APL_TDA. I can't say that enough. Read that room. You'll be surprised what you find.
Ron,
A thought would be if you are indeed going to turn those black cubes into subwoofer enclosures would be to use smaller speakers that would be on four sides of the box. This would work somewhat like a distributed sub system as the cones would be facing in all directions and the box would act as an omni-directional source. This would give multiple path lengths to any walls or reflective surfaces and help with any single room modes. Just a wild thought off the top of my head. What is the internal volume of the box you will be using if your using those black boxes?
A thought would be if you are indeed going to turn those black cubes into subwoofer enclosures would be to use smaller speakers that would be on four sides of the box. This would work somewhat like a distributed sub system as the cones would be facing in all directions and the box would act as an omni-directional source. This would give multiple path lengths to any walls or reflective surfaces and help with any single room modes. Just a wild thought off the top of my head. What is the internal volume of the box you will be using if your using those black boxes?
I did think about that 
but most sides face unusable areas. Can't use the top, nor the side facing the rear wall or the array. Visually it will have to stay the same (I got away with that by placing those cubes there to let that sink in. There used to be only one, the other side had my stereo mounted on the wall. A sneaky way from me to be able to replace those benches with subs down the road.
I expect I won't get much more than 40 Litre. I have run some sims long ago, copying the Thor subs. I'll redo them to refresh my memory with the most likely candidates.
but most sides face unusable areas. Can't use the top, nor the side facing the rear wall or the array. Visually it will have to stay the same (I got away with that by placing those cubes there to let that sink in. There used to be only one, the other side had my stereo mounted on the wall. A sneaky way from me to be able to replace those benches with subs down the road.I expect I won't get much more than 40 Litre. I have run some sims long ago, copying the Thor subs. I'll redo them to refresh my memory with the most likely candidates.
To come back to the DSP point, just look what happens when measuring a stereo pair:
Uncorrected stereo speakers at the listening spot:
Compared to corrected:
A dramatic change, but the real work that's done is a smaller change than you'd imagine.
Phase correction only happens from 20 Hz to 1000 Hz. Not 6 cycles as some would assume, less than half that, judging all the REW graphs for the moment the room takes over the response. Fix it before that, not after.
Sort of like how you read an IR and gate before the first reflection, only much more work to find. Walking dynamically though literally all the plots REW presents using all the filter tools. Figuring out the cause of any change and resisting the urge to fix them all.
Believe me I tried, lock your head in place of the microphone and it will work, but it won't work when you move more than 10 cm.
Uncorrected stereo speakers at the listening spot:
Compared to corrected:
A dramatic change, but the real work that's done is a smaller change than you'd imagine.
Phase correction only happens from 20 Hz to 1000 Hz. Not 6 cycles as some would assume, less than half that, judging all the REW graphs for the moment the room takes over the response. Fix it before that, not after.
Sort of like how you read an IR and gate before the first reflection, only much more work to find. Walking dynamically though literally all the plots REW presents using all the filter tools. Figuring out the cause of any change and resisting the urge to fix them all.
Believe me I tried, lock your head in place of the microphone and it will work, but it won't work when you move more than 10 cm.
I've looked at the XXLS400 subs for inspiration, both models I found,
the normal configuration:
And the down firing one:
Attachments
I'll probably cover them up with the current fabric, mount the woofer like I did on the arrays.
Proof though, the Beolab 90 does indeed feature a trio of Discovery subs: Starting at 1:08 to show the back of the 26W driver
Proof though, the Beolab 90 does indeed feature a trio of Discovery subs: Starting at 1:08 to show the back of the 26W driver
Is it because of the driver size it is too big?
Because in my experience the XLS or XXLS needs at least 50 liters in order to come close to their optimal cabinets - And this 15" RSS390HO driver will perform great in 50 liters (I know some would like a lower Q, but these will do fine with a Q just below 0,7).
I have had some 23W drivers - I think the 32W is a larger and newer version of the 23W - and that is a really great driver. But even though I had four 23W, the SPL just was not enough to match my system - and I really could not integrate them the way I wanted in my system - they sounded to slow and heavy (there is no such thing as slow bas - but I lack a better description of the sound). Maybe the 32W is better (?)
Deep and well integrated bass is essential in a great HiFi system, in my opinion, and even though I hear many talk about the lack of low end output on most music, I definitely would not give it up now I am used to having it.
Maybe you could be inspired by Earl Geddes take on great bass?
He does have some interesting approaches on great bass and it might be able to fit the WAF factor as well with three or four small subs hidden away in the room.
What about side-mounting an array of smaller woofers / mid-woofers for the low end? I have something like the Hivi B4n in mind, as they have a low fs especially for the size. Not the high end extension, but that's not what you'd want them for, and bringing them in probably from 3-400hz down. I suspect that's too high though and directionality would suffer, but it would be low footprint and "relatively" easy to do if you were building from scratch.
Another option might be an array of 6 or 8 inches on the side, spaced at fairly large c-c for back volume, and crossed at 80-100hz, but then you're still loading those tc9 down a little lower...
On the other hand, the sealed cube subs are appealing, just ensure you join them properly, otherwise they can crack on the joints (speaking from experience here!)
Sent from my iPhone using Tapatalk
Another option might be an array of 6 or 8 inches on the side, spaced at fairly large c-c for back volume, and crossed at 80-100hz, but then you're still loading those tc9 down a little lower...
On the other hand, the sealed cube subs are appealing, just ensure you join them properly, otherwise they can crack on the joints (speaking from experience here!)
Sent from my iPhone using Tapatalk
I'd recommend you download the demo of APL_TDA and run that in your room.
I know the Geddes theory and I've tried to get Geddes try mine. Look at the group delay out in my room:
Spreading subs throughout the room will not get you this. Look at the millisecond scale.
Geddes is convinced group delay below 70 Hz does not matter. I don't see it that way. Combine that plot above, a nice visual of the time alignment with this FR plot:
Frequency and Phase as measured from a stereo pair at the listening spot. (6 cycle FDW view)
And know this is the IR/STEP:
Now that's what I want to improve on. As separately the left and right channel "do" show some irregularities below 70 Hz. Where the left is helped out by the right channel and vice versa.
My girlfriend has given me strict orders not to add anything visual. So all I have available are two tiny feet bench that could be turned into subwoofers.
Once you talk about Linkwitz transform, PEQ and FIR, the box Q isn't that relevant anymore, it wasn't on my arrays either. Each of the TC9 drivers has just over 2 litre, the impedance bump shows at ~145 Hz. Look at that frequency curve again, what Q were we talking about? Afraid of what that will do to distortion? Not anymore, here's an oldie...
Again, as measured at the listening position.
The only thing that does matter is to start with a valid good performing subwoofer with decent efficiency, damp the back wave of the sub so it "sees" a bigger enclosure but also shows the smoothest impedance plot I can get in that tiny sealed enclosure. The rest is up to sheer amplifier power to drive the speaker up to it's potential. Making sure not to step over it's obvious boundaries.
Is it going to work out to improve on this? Apart for a possible financial loss, it wouldn't matter. It's performance is outsane as is (I.M.H.O.). But it is a challenge.
Think of this thread as a DIY way to get some of that Beolab 90 theory in a living room. Sure it's not the same, but the goals aren't that different. Only I work with the room, by placing damping panels at key points and choosing this type of speaker, that does have certain advantages over most "normal" setups.
Anyone with a Geddes multi-sub setup willing to show APL plots at the listening position? Sure it will smooth out room modes, but what does it do for timing?
I know the Geddes theory and I've tried to get Geddes try mine. Look at the group delay out in my room:
Spreading subs throughout the room will not get you this. Look at the millisecond scale.
Geddes is convinced group delay below 70 Hz does not matter. I don't see it that way. Combine that plot above, a nice visual of the time alignment with this FR plot:
Frequency and Phase as measured from a stereo pair at the listening spot. (6 cycle FDW view)
And know this is the IR/STEP:
Now that's what I want to improve on. As separately the left and right channel "do" show some irregularities below 70 Hz. Where the left is helped out by the right channel and vice versa.
My girlfriend has given me strict orders not to add anything visual. So all I have available are two tiny feet bench that could be turned into subwoofers.
Once you talk about Linkwitz transform, PEQ and FIR, the box Q isn't that relevant anymore, it wasn't on my arrays either. Each of the TC9 drivers has just over 2 litre, the impedance bump shows at ~145 Hz. Look at that frequency curve again, what Q were we talking about? Afraid of what that will do to distortion? Not anymore, here's an oldie...
Again, as measured at the listening position.
The only thing that does matter is to start with a valid good performing subwoofer with decent efficiency, damp the back wave of the sub so it "sees" a bigger enclosure but also shows the smoothest impedance plot I can get in that tiny sealed enclosure. The rest is up to sheer amplifier power to drive the speaker up to it's potential. Making sure not to step over it's obvious boundaries.
Is it going to work out to improve on this? Apart for a possible financial loss, it wouldn't matter. It's performance is outsane as is (I.M.H.O.). But it is a challenge.
Think of this thread as a DIY way to get some of that Beolab 90 theory in a living room. Sure it's not the same, but the goals aren't that different. Only I work with the room, by placing damping panels at key points and choosing this type of speaker, that does have certain advantages over most "normal" setups.
Anyone with a Geddes multi-sub setup willing to show APL plots at the listening position? Sure it will smooth out room modes, but what does it do for timing?
Last edited:
Ron,
Earl is a really smart guy but very convinced his solution is the only one to work. Look around though and tell me how many of those systems are really out there, not many from what I can tell. He doesn't believe that phase alignment is important at anything but the crossover point, this is back up by his own research but many don't seem to agree with his conclusions about all of this.
I am sending you a PM.
Earl is a really smart guy but very convinced his solution is the only one to work. Look around though and tell me how many of those systems are really out there, not many from what I can tell. He doesn't believe that phase alignment is important at anything but the crossover point, this is back up by his own research but many don't seem to agree with his conclusions about all of this.
I am sending you a PM.
Wesayso, how are two array speakers any different from two subs (say) when you think about it from the prespective of a dumb microphone? They are just two sources of sound placed in the room and equalized for time and level. I am not sure where this notion that the arrays are somehow superior to distributed subs comes from. If you applied the same techniques you did to your arrays to distributed subs, you would get the same results.
I am going to try to measure over the coming days and report back. It will obviously not look as good as your measurements because I haven't put in the kind of heroic efforts in my room in trying to eliminate reflections or time my subs perfectly. I could spend hours doing that like you did but I just don't have the time. Though, more importantly, I don't think it is audible. At 20 Hz, 360 degrees is equal to 50 milliseconds and 56.25 feet. Any fine tuning you do on the millisecond level to get that APL TDA graph to look sweet is unlikely to be audible.
All your efforts are definitely laudable and better measurements are always good, but with the DSP power available, we must be careful about chasing ghosts and things that unlikely to be audible.
Another point about how the arrays work should make you think. At any given frequency and measuring distance in the near field, each driver is a certain distance away from its neighbors. If at say 1000 Hz, two drivers have a phase difference of 360 degrees, they will sum perfectly and in REW it will show up as an SPL gain, not as a notch. The same way, it does not matter if there is some delay between the sub and the mains as long as they sum correctly. Of course if the delay gets too long, it sounds bad.
I don't think I am going to convince you otherwise. But you are probably in a better position to run different tests because you have a perfectly aligned system. You can test imperfect time alignment but same FR and report back. Of course, the test must be blind.
I am going to try to measure over the coming days and report back. It will obviously not look as good as your measurements because I haven't put in the kind of heroic efforts in my room in trying to eliminate reflections or time my subs perfectly. I could spend hours doing that like you did but I just don't have the time. Though, more importantly, I don't think it is audible. At 20 Hz, 360 degrees is equal to 50 milliseconds and 56.25 feet. Any fine tuning you do on the millisecond level to get that APL TDA graph to look sweet is unlikely to be audible.
All your efforts are definitely laudable and better measurements are always good, but with the DSP power available, we must be careful about chasing ghosts and things that unlikely to be audible.
Another point about how the arrays work should make you think. At any given frequency and measuring distance in the near field, each driver is a certain distance away from its neighbors. If at say 1000 Hz, two drivers have a phase difference of 360 degrees, they will sum perfectly and in REW it will show up as an SPL gain, not as a notch. The same way, it does not matter if there is some delay between the sub and the mains as long as they sum correctly. Of course if the delay gets too long, it sounds bad.
I don't think I am going to convince you otherwise. But you are probably in a better position to run different tests because you have a perfectly aligned system. You can test imperfect time alignment but same FR and report back. Of course, the test must be blind.