We could argue all day long about the woofer speed issue but the only thing that could/should resolve it would be a double blind test. Can you hear the difference between the two subs when they are eqed to the same freq response?
Consider that those long throw woofers with heavy cones and thus require strong motors, which results in high Le. That high Le leads to an early HF roll off making that kind of woofer less well suited for operation above 100 Hz.
At the same, I've seen slot loaded sub designs where the slot is designed to function as a horn (edit: and bandpass chamber) and provide gain just where the high Le is causing the response to start to roll off, thus extending the response without boosting the high end. That might be a problem at high output where the cone doesn't move enough to cool the voice coil.
So you really can go round and round on this. But if you want a true woofer instead of a pure subwoofer and select based on smoothness and flatness of response up to 500 Hz or 1 Khz or so, you will likely end up with a low Le woofer which correlates with having a light weight cone.
Consider that those long throw woofers with heavy cones and thus require strong motors, which results in high Le. That high Le leads to an early HF roll off making that kind of woofer less well suited for operation above 100 Hz.
At the same, I've seen slot loaded sub designs where the slot is designed to function as a horn (edit: and bandpass chamber) and provide gain just where the high Le is causing the response to start to roll off, thus extending the response without boosting the high end. That might be a problem at high output where the cone doesn't move enough to cool the voice coil.
So you really can go round and round on this. But if you want a true woofer instead of a pure subwoofer and select based on smoothness and flatness of response up to 500 Hz or 1 Khz or so, you will likely end up with a low Le woofer which correlates with having a light weight cone.
Last edited:
Think the speed problem is the relative difference in inductance data L(e) per band pass, so say woofer and tweeter band had same L(e) data the problem wasn't there or we can set XO point for woofer a tad low enough before its inductance start the climbing on the graph, low L(e) is one reason JoeRasmussen in his Helsinore build use those 7 inch SBA woofers to integrate with his tweeters because their L(e) is as low as 0,13mH.
L(e) is a distortion the higher the frq the bigger the time delay, but not shure our hearing can reconize the short time distortion listening to say one transducer at a time, but when summing two drivers with different L(e) harmonics as frq goes up can get out of timing no matter time sum is perfect in the low end so get it crossed below that point, most often transducers with advanced motors have improved L(e) numbers probably to help integrate with tweeters.
Think not wesayso should get any L(e) coursed speed problems as seen below : )
L(e) is a distortion the higher the frq the bigger the time delay, but not shure our hearing can reconize the short time distortion listening to say one transducer at a time, but when summing two drivers with different L(e) harmonics as frq goes up can get out of timing no matter time sum is perfect in the low end so get it crossed below that point, most often transducers with advanced motors have improved L(e) numbers probably to help integrate with tweeters.
Think not wesayso should get any L(e) coursed speed problems as seen below : )
Attachments
Last edited:
Clever idea! Using the damping for 3 purposes (add HD reduction from sub) is fantastic design.
Funny. I'm stealing that when I get the chance.
I'm not a "fast bass" proponent, but there is another way of looking at it. Suppose you play a 50Hz sine at 83dB. Now turn it up to 95dB. How fast is the cone moving (cm/sec) in both scenarios? Could a lower Mms cone respond better in a dynamic situation, when the SPL isn't static?
As I see/hear it, this is what most people attribute the "slow bass" moniker.
The multiple booming reflections arriving during a period of time takes the snappiness of it away.
So, most of the time, it is the room which is the culprit.
Adding a couple of bass traps will improve on that.
Of course, if one gets a very cheap sub that has no cone control and is flapping endlessly, then that distortion will also contribute to the perceived slow bass, but with any of the nice subs or woofers out there, without needing to spend outrageous amounts of money, it is very possible to get a very nice and tight sounding bass with placement, adjusting delays, some EQ because we usually don't have the chance to have the perfect room, and adding some form of bass traps to deal with that room.
Back when I knew even less of loudspeakers than now, I used to place my subwoofer at LP and then look for the place in the room it sounded 'best'. Then placed subwoofer at that spot. I had the idea that it worked, however the space beind 3rd dimensional, it could make for some very impractical subwoofer placement locations 
To the contrary, if the room is very long, it might also open up some new subwoofer locations at the other end of the room. I'll be laughing out loud if this leads to new pictures of random stuff taking up space for his other half to get used to the idea
On a more serious note,
'Fast' bass, I think has to do with spectral decay/impuls response.
Hypothesis: At <25Hz, the 40ms(?) window to perception of these things as a seperate quality opens up...
To the contrary, if the room is very long, it might also open up some new subwoofer locations at the other end of the room. I'll be laughing out loud if this leads to new pictures of random stuff taking up space for his other half to get used to the idea
On a more serious note,
'Fast' bass, I think has to do with spectral decay/impuls response.
Hypothesis: At <25Hz, the 40ms(?) window to perception of these things as a seperate quality opens up...
Last edited:
For people looking for drivers: Keep in mind, that not every woofer needs the same excursion to achieve the same SPL. Xmax therefore is a number, but only one of many.
More about Xmax: value is defined as the maximum excursion that is within "reasonable" limits of distortion. These "reasonable" limits could be and often are: 10% THD. And that is, if the manufacturer adheres to the "standard" lol. This can not be considered reasonable for anyone that's serious with audio.
Klippel reports are valuable.
For high fidelity, I would prefer a driver which shows an absolutely flat klippel curve over 10mm, if it has a 12mm Xmax, over a driver that has 30mm Xmax but only shows a straight klippel response over 8mm.
When listening to music, I stay withing limits of the straightest part of the curve.
Under 15Hz, it becomes more difficult to hear with the ears alone. This is where I don't experience non-linear excursion as problematic. Movie effects can still be had while designing a system for music is what I'm trying to convey.
Also, it's usefull to keep output voltages in mind too when designing your boxes, subwoofer or not.
Black= 2.83v, Blue = 33V. The lower two curves we are interested in (closed box).
I wonder if Klippel curves would differ too? I heard DUMAX might have answers to this.
@wesayso: Excursion linearity, power linearity, box size, phase and frequency response... MFB can lessen the strain on requirements, enabling more freedom in driver and design choices for subwoofers a lot. Take a look on MFBlabs.nl. When using MFB regard the lowering of distortion as a nice bonus ;-)
Also, I saw a couple of posts on ambient tweeters a few pages back.
Have you considered using a low shelving filter instead of low pass filters in your ambient speakers?
More about Xmax: value is defined as the maximum excursion that is within "reasonable" limits of distortion. These "reasonable" limits could be and often are: 10% THD. And that is, if the manufacturer adheres to the "standard" lol. This can not be considered reasonable for anyone that's serious with audio.
Klippel reports are valuable.
For high fidelity, I would prefer a driver which shows an absolutely flat klippel curve over 10mm, if it has a 12mm Xmax, over a driver that has 30mm Xmax but only shows a straight klippel response over 8mm.
When listening to music, I stay withing limits of the straightest part of the curve.
Under 15Hz, it becomes more difficult to hear with the ears alone. This is where I don't experience non-linear excursion as problematic. Movie effects can still be had while designing a system for music is what I'm trying to convey.
Also, it's usefull to keep output voltages in mind too when designing your boxes, subwoofer or not.
Black= 2.83v, Blue = 33V. The lower two curves we are interested in (closed box).
I wonder if Klippel curves would differ too? I heard DUMAX might have answers to this.
@wesayso: Excursion linearity, power linearity, box size, phase and frequency response... MFB can lessen the strain on requirements, enabling more freedom in driver and design choices for subwoofers a lot. Take a look on MFBlabs.nl. When using MFB regard the lowering of distortion as a nice bonus ;-)
Also, I saw a couple of posts on ambient tweeters a few pages back.
Have you considered using a low shelving filter instead of low pass filters in your ambient speakers?
Indeed a clever idea. Have you seen this implemented anywhere? I tried to dig a little around for some experience with such an arrangement, but could't find any. I might steel that idea, when I finally do my line array subs..
Haven't seen it anywhere else. Straightforward analysis and thinking shows that it "should: work. Boundary interference calculator will tell you where the boundary nulls are for the HF array. Passive absorber calculator will tell you thicknesss of absorber to kill those nulls and then how much of the LF will pass through that thickness.
Interesting to read all the good posts here
And darn if I know what is right or wrong .... some of my previous posts were not so much my facts but merely statements from people in the business that worked with this for ages.... still, there is no difficulty to find people who argue the other way around
Pretty expensive to do experiments though if they are based on €500+ drivers
Sometimes I think just: to h#&@ with it and just buy two Dayton 18” ultimax, and see what happens
And darn if I know what is right or wrong .... some of my previous posts were not so much my facts but merely statements from people in the business that worked with this for ages.... still, there is no difficulty to find people who argue the other way around
Pretty expensive to do experiments though if they are based on €500+ drivers
Sometimes I think just: to h#&@ with it and just buy two Dayton 18” ultimax, and see what happens
Last edited:
One point in a bass line array is beam control - in vertical. However that doesn't happen down low...
I have 2x15" Dayton DC380 drivers in a a slot loaded HT subwoofer and I remember I was amazed how they played fullrage when testing... 18" Ultimax should be ok as hifi woofer too, crossed below 500Hz. I bet they are easy to sell if you get disappointed. I think I have heard some postive comments of JBL M2 which has a single 15" woofer crossed around 1kHz to a horn, no complaints of sloppy or slow bass...
I have 2x15" Dayton DC380 drivers in a a slot loaded HT subwoofer and I remember I was amazed how they played fullrage when testing... 18" Ultimax should be ok as hifi woofer too, crossed below 500Hz. I bet they are easy to sell if you get disappointed. I think I have heard some postive comments of JBL M2 which has a single 15" woofer crossed around 1kHz to a horn, no complaints of sloppy or slow bass...
Attachments
Last edited:
Interesting to read all the good posts here
And darn if I know what is right or wrong .... some of my previous posts were not so much my facts but merely statements from people in the business that worked with this for ages.... still, there is no difficulty to find people who argue the other way around
Pretty expensive to do experiments though if they are based on €500+ drivers
Sometimes I think just: to h#&@ with it and just buy two Dayton 18” ultimax, and see what happens
I was a bit curious, but at 665 gram mms this sure isn't a low cone weight
.None of the Dayton subs seem to fall into that category.
There certainly is an advantage of using multiple drivers like your plan. But you've got to trust in such a project to bite that bullet.
Lots of people were telling me the arrays weren't going to work. More so on another forum I posted than over here, probably because I wasn't the first here to do so. However I had faith in it and some emergency backup plans to still make it work if i were disappointed somehow.
I think you can build the sub you want. Take your time to research possible drivers. Run some sims and draw up a model (or just some paper sketches still work for that). Sim it to see where you end up with fixed boundaries like x-max limited SPL levels (if we can trust the specs). Figure out a good amp to feed them etc.
I'll take a look! My processing is sort of a feed forward already
. Measured, correcting the measurement and applying that before the amplifier. However it's not the same. I have looked into threads about motion feed back in the past, but have not been seriously planning to use it yet.I have enough experiments as it is, as I'm not planning a normal crossover between subs and arrays, that alone is going to take me some time to setup.
On the topic of fast bass, I usually post this graph:
Taken at the listening spot in my room, this is what fast bass means to me. This is the Stereo sum at the listening spot. Left nor right channel looks like that by itself. The sub project is a mission for me to keep the above result, but make each channel behave that much better by itself as well. Slightly higher headroom and reduced needs for boost for the arrays (low end) are hopefully useful byproducts.
That idea about a low shelving instead of low-pass is interesting for the ambient channels. Ambient tweeters were on my mind because I could make them coming from the front. But I'm not thrilled about adding even more hardware. So far I have 2 Stereo amps and 2 mono amps running. A bit much for the average full range project (lol). But if it works!
I was a bit curious, but at 665 gram mms this sure isn't a low cone weight
None of the Dayton subs seem to fall into that category.
There certainly is an advantage of using multiple drivers like your plan. But you've got to trust in such a project to bite that bullet.
Lots of people were telling me the arrays weren't going to work. More so on another forum I posted than over here, probably because I wasn't the first here to do so. However I had faith in it and some emergency backup plans to still make it work if i were disappointed somehow.
I think you can build the sub you want. Take your time to research possible drivers. Run some sims and draw up a model (or just some paper sketches still work for that). Sim it to see where you end up with fixed boundaries like x-max limited SPL levels (if we can trust the specs). Figure out a good amp to feed them etc.
I believe that going with a sealed design and a setup that takes ideas from a commercial design, using simulations... it can’t go that bad.... and even so using great ScanSpeak drivers that proved their way in other diy designs... I don’t think the risk is that large .... I reckon the question is how far-out we can risk going on our designs
That graph is awesome, and I thank you for having the courage for being so complete in your documentation. The initial response curve alone, is food for thought already.
The graph reminds me of mapping ECU's a bit in this 3D view.
I hope that setup of the subwoofers will provide you both with the challenge and the results you are hoping for. Can you explain which one of the towers requires more energy right now, in which freq ranges and how much energy difference is there?
The graph reminds me of mapping ECU's a bit in this 3D view.
I hope that setup of the subwoofers will provide you both with the challenge and the results you are hoping for. Can you explain which one of the towers requires more energy right now, in which freq ranges and how much energy difference is there?
Last edited:
I believe that going with a sealed design and a setup that takes ideas from a commercial design, using simulations... it can’t go that bad.... and even so using great ScanSpeak drivers that proved their way in other diy designs... I don’t think the risk is that large .... I reckon the question is how far-out we can risk going on our designs
How much are you able and willing to spend for some cool project to learn from?
If it works out, that's always a great feeling! If it doesn't, figure out the "why" and make it work! That's what works for me.
I think if one goes sealed he will get what he simulates - anechoically. How it performs in the room is another story. That other story is the critical part and where persistence comes in. That is why I would hesitate to put a huge investment into a single bass source when several lesser sources distributed around the room might work better. Of course if you can afford several uber subs, then go for it!
I'm betting it will help me out to have more drivers to play with, here's an old graph (december 2016) that shows how left and right combine to a good overall Stereo balance:
This above graph is a 5 cycle Frequency Dependent Window (FDW). If we look at an ungated response, filtered with 1/12 it looks like this:
(forget about room curves etc, that might have changed, my left/right processing still looks largely the same as shown here)
I've always encouraged using EQ on the direct wave and not much else. As can be seen, the ungated results hold up pretty good, even though very short windows were used for correction. A big part of that is finding the biggest flaws (of the room) and fix those with treatment.
The couple of dips in the right channel (140-220) are from the wall behind the listener. That makes it hard to repair that, without adding treatment. You can't EQ a "0". All of the trouble higher than about 80-100 Hz should be fixed in the room anyway. Below that we can shift some energy and still get away with it. If it's within reason.
The hole in the left (gray color in top graph) channel at 72 Hz is compensated for in the right channel. I even remove energy from the standard processing at the left, and add a similar amount of boost to the right. This is about 4 dB max to get a good overall Stereo sum.
The left array is placed near a corner and even though that creates the dip at ~72 Hz, it also supports that channel (room gain) below ~45 Hz. The right channel struggles there in comparison as it does not have a true corner to help it out.
Both channels have exactly the same amount of boost applied around 25-30 Hz. They sum real well after finding the best balance. As you see it never takes more than about 4-5 dB to fix. Due to the array nature (which needs help with parallel planes, I used absorption on first reflection parallel planes) that can average out many unwanted little stuff. Each driver is at a slightly different position than the next, that's the secret of arrays. They hardly ever share the same reflection, happening at the same frequencies, aside from those parallel planes. So very little treatment goes a long long way.
The right sub should help me get a true bottom end on the right side, while the other sub would help straighten out the dip at 72 Hz on the left. Both subs will make it possible to carry the first octave and a half so those small drivers won't have to. I'll try and help the right channel around 100-200 Hz if I can, but don't expect much improvement. Placing of subs it should be possible to get it a little better, as the damping panel behind the listening spot is a little off centre when viewed from speaker placement points. It's possible the sub does play nice there and will loan/share the load there. We're still talking about wavelengths surpassing 1.5 m so the placement of sub/array close together (as seen from the listener) should make it possible.
There's hope that distortion levels will improve too, here's what they look like from a 2015 measurement:
This thread is filled with similar results. It's even quite difficult not to get passing cars outside disrupt the true results. Here the subs should provide headroom as the room curve I have would not work for loud bass notes as found in Home Theatre. However most music doesn't really hit that low and loud. The arrays work like they do due to one reason only. Placement near the wall gives it a boost (room gain and boundary gain) in the bottom end. They are way too tiny, even in these large numbers, to be able to do this, if there wasn't a nearby wall to help out. If my room were bigger, it wouldn't work either. My room is 8m long by slightly over 4m width. The stereo is on one side of that room along the long wall, hence the asymmetrical placement. That asymmetrical placement is both a blessing and a curse
.
(those black benches beside the arrays will be replaced by the subs, the only place I'm allowed to use them at, no discussion possible. So I'll have to make those spots work!)
What is your secret? If that were my room you would see lots more treatment and troublesome room modes despite the treatment. One thing which seems to be the same for both of us, or so I infer from your comments, is that the wall behind the speakers is the least troublesome. I think one of the things that helps you is aiming the speakers across the width of the room; I have a roughly 25' square room which moves the modes down lower where they are harder to absorb and square is known to be problematic.