Seems like I haven't saved most of them. These were done 5 years ago as answer to kimmosto on a Finnish forum. No answer... no reference, so forgotten. Doppler and IMD is supposed to be a problem with both multiway and widerange speakers... I can't understand, what is reference?
New revisions of REW have this multitone test that you have used - I haven't done those. Perhaps I'll try single fullrange vs. 2- way vs. 3-way vs. 4-way? Different spl high-low? Distance? In-room artefacts?
Those look really clean! Can I suggest we move the multitone discussion to the new thread I created? https://www.diyaudio.com/community/...n-measurement-on-drivers.391826/#post-7163145
Distortion of a speaker/driver should go down with volume. If not (and this happens with PA drivers more often) you have some effects happening (unlinearities in the mechanics for e.g.) - I don't want that in my speakers.
But of course you need a useful measurement microphone and a very silent environment to do these measurements! I'm happy to have both ;-) (real measurement mics start at 1000,-. For a budget project I bought a bunch of MicW M215, they have a low noise floor).
When measureing 0,1% you need a noise floor significant lower as 60dB under your source signal. (percentage and dB are easy to recalculate if you are used to it). 94dBSpl is the standard -> your noisfloor from microphone and room should be in the 25dBSpl range. When you start measureing with easy 80dBSpl in your living room ... don't think about measureing THD (unless your drivers are crap ;-)).
Always make a noise floor measurement! (just unplug the speaker and perform a measurement)
@Rallyfinnen Did you try listening to the 15NMB420 yet? Did you like it? That driver is on my short list for a sealed bass leg for my existing 10NMB420 mids and NSD1095N highs, so we have some similarities in our preferences 
I've read a top tip from multiple sources: buy the best you can afford. But just how do I know what's best? The price is not a absolute indicator. There is also the 16 Ohm version of the 15ND930, which seems to be slightly better suited for sealed cabinets. You spoke of the Faital Pro 15PR400, do you know if it really would be better?
I've read a top tip from multiple sources: buy the best you can afford. But just how do I know what's best? The price is not a absolute indicator. There is also the 16 Ohm version of the 15ND930, which seems to be slightly better suited for sealed cabinets. You spoke of the Faital Pro 15PR400, do you know if it really would be better?
I get most of that, but it seems to me that averaging can remove a lot of the background noise based on what I was able to measure with multitone in my garage. I guess averaging could be used if you do distortion measurements with stepped sine too, not sure about sweeps? What is your opinion on the multitone tests? https://www.diyaudio.com/community/threads/multitone-distortion-measurement-on-drivers.391826/
Not 'really'. I did some impedance measurements and listened to one 'wideband' without any stuffing in the box. I'm waiting for stuffing material, DSP etc to arrive, so it will take some time before I even listen to them in the garage, and even more before they are in my living room. This is my 'winter project', so I don't want to rush it, the winter is long!@Rallyfinnen Did you try listening to the 15NMB420 yet? Did you like it? That driver is on my short list for a sealed bass leg for my existing 10NMB420 mids and NSD1095N highs, so we have some similarities in our preferences
I've read a top tip from multiple sources: buy the best you can afford. But just how do I know what's best? The price is not a absolute indicator. There is also the 16 Ohm version of the 15ND930, which seems to be slightly better suited for sealed cabinets. You spoke of the Faital Pro 15PR400, do you know if it really would be better?
I have no clue if other drivers are better or worse. All I can say is that I did not like the basket resonance on these around 250Hz, other than that they seem like fine woofers. My guess is that the frame was made for ferro-magnets, and that the mass-loading form a ferro magnet would have moved the resonance far lower in frequency and made it less likely to be audible.
I was aiming for the 15PR400 just because other builders seemed to like them, and the price was good, but they were out of stock then. I see they are available now. This is my first 15" build, so I can't really say I have any experience with 15" drivers
Parameters for this 18sound woofer are quite close, and I found them used, so I just went for them.
How about using home-hifi 15" woofers instead of PA "woofers"? PA stuff have typically high Fs, low Xmax and very stiff suspension. They seem to have high sensitivity, but that comes with limited low bass capacity.
Ten years ago I ended up with cheapish Dayton DCS380 15" for HT subwoofers I didn't test them at all because they are dsp-subwoofers in my case
https://www.diyaudio.com/community/threads/a-bench-style-cabinet-subwoofer-2x15-closed.223960/page-4
More tests https://data-bass.com/#/drivers?_k=ne2kvi
Ten years ago I ended up with cheapish Dayton DCS380 15" for HT subwoofers I didn't test them at all because they are dsp-subwoofers in my case
https://www.diyaudio.com/community/threads/a-bench-style-cabinet-subwoofer-2x15-closed.223960/page-4
More tests https://data-bass.com/#/drivers?_k=ne2kvi
I think I must have been sold the 'fast bass' argument
I have had the thought of actually trying sub drivers too, but that will be for later. What I can say in my defense is that the room is not that big, so playing low in the room should not require much, and I'm not aiming for high SPL, and most of my amps are below 100W. Hoping that these will match the room gain to some degree, but on the other hand I will have DSP do do whatever corrections needed. I have no idea if a corrected PA driver with a relatively high Fs will sound different than a corrected sub driver with low Fs. I have even simulated adding 10g to these drivers, and the total Q should still be under 0.6 but with a bit lower Fs. Possibilities for future experiments
For the same reason I want the upper baffle to be screwed, in case I want to try other drivers.
In summary, many uncertainties and therefor some flexibility in the design.
As a side note, I received some boxes with damping material today. Some of it to put inside the boxes, and some outside Will put some foam on the backside of the speakers, and also planning to do some room treatment. Denim and hemp-based insulation, I don't want to use rockwool or fiberglass in the room or in the boxes.
I have had the thought of actually trying sub drivers too, but that will be for later. What I can say in my defense is that the room is not that big, so playing low in the room should not require much, and I'm not aiming for high SPL, and most of my amps are below 100W. Hoping that these will match the room gain to some degree, but on the other hand I will have DSP do do whatever corrections needed. I have no idea if a corrected PA driver with a relatively high Fs will sound different than a corrected sub driver with low Fs. I have even simulated adding 10g to these drivers, and the total Q should still be under 0.6 but with a bit lower Fs. Possibilities for future experiments
For the same reason I want the upper baffle to be screwed, in case I want to try other drivers.
In summary, many uncertainties and therefor some flexibility in the design.
As a side note, I received some boxes with damping material today. Some of it to put inside the boxes, and some outside
Will put some foam on the backside of the speakers, and also planning to do some room treatment. Denim and hemp-based insulation, I don't want to use rockwool or fiberglass in the room or in the boxes.
Last edited:
I have simulated a bunch of 15" woofers (🐶) with dsp filters. When EQed to similar frequency response, all woofers behaved more or less the same. You are mostly limited by excursion and power handling for SPL. Or I have simulated only similar woofers, who knows.. I'm not sure what that means in reality though, simulators don't show distortion and what else I'm forgetting. Anyway I'd be amazed if you can't squeeze a decent output from any 15" woofer in a moderate box for home use.
Luckily for me, I'm not looking for SPL, a pair of 8" woofers would probably be enough for my listening levels I'm more concerned about bass quality than quantity. Room treatment is on the agenda too, since that is a major problem since I moved.
How about BR vs DSP EQ when it comes to group delay?
I'm more concerned about bass quality than quantity. Room treatment is on the agenda too, since that is a major problem since I moved.How about BR vs DSP EQ when it comes to group delay?
Averaging helps against noise and is very welcome to be used when needed.
A THD measurement gives way more information in my opinion and is easy to do - so I stick with it. When you have high THD you alo have high multitone distortions.
Audix TM1 is a good cheaper measurement mic. Roughly $300 and comes with a cal file. It even sounds great used for recording.
I have an earthworks M40, but rarely use it anymore because the TM1 goes past 25k, which is adequate in most situations.
Very similar, but the closed is better by some 20-30%. The port adds group delay peak of a few ms, but nothing major for a 15" driver. I simulated the Eighteen Sound 15ND930 four different ways. It's made for vented boxes, but works fine in closed too. Each is limited to excursion at 10 Hz.Luckily for me, I'm not looking for SPL, a pair of 8" woofers would probably be enough for my listening levels
How about BR vs DSP EQ when it comes to group delay?
Closed box with dsp filter boosting bass.
I used just one Linkwitz transform filter and tried to limit the extension to match the others. The result became easily the loudest, but also draws most power. NB no high-pass filter, which would be a good idea to limit power draw and excursion for subwoofer range. ~63 litre box. With dsp you can very easily EQ it flat even wider with just one filter.
I'm not a pro when it comes to BR tuning or passive filters, but here is the BR tuned to ~39 Hz, 70 litre box, one low-pass filter at 70 Hz.
I guess it could be tuned lower, but I'm no expert in how and why. It draws a lot less power than closed, but extension is a lot worse. Note the peak in the GD. This too would benefit from from infra-filter.
And finally the same BR with passive filters.
Again, I'm not very good with analog components, but this filter has a coil, a cap and two resistors. It's essentially the same for most likely intended frequency range. With passive crossover you are not going to make an infra-filter, you'd need a power plants worth of caps.
There are also the passive radiators, but optimally that would need a radiator unit that's as large as the main driver and I've got no data of such things. However I put two 12" radiators with a dsp filter into my simulator.
Box size matches sealed. PR is supposed to combine the best of sealed and vented, but to me it looks like the worst of both. SPL dropped and GD rocketed, albeit at subwoofer frequencies.
-EDIT- Tuning lower would require 100-200 cm reflex tube if you want to avoid port noises and it dramatically increases GD. Again I'm no expert in this.
Last edited:
The increase in group delay comes from an increase in the slope of the highpass function. Adding an extra highpass to control excursion in a vented box will cause the group delay to rise but only at quite low frequencies, this is a small price to pay compared to having the cone flapping uncontrolled if the track contains enough very low frequency content.
Using higher slope minimum phase crossovers will cause the overall GD to rise too.
With an acausal FIR filter you can remove any excess group delay, but at a cost of latency, filter taps and an increased level of pre-ringing if it's a single sided non complementary filter.
Excuse me asking a dumb question about Excess group delay, but
I've learned that a reflex radiator (and bass horns?) radiates sound one full cycle late at tuning F (x ms related to frequency). We often see summed simulated response of BR compared to sealed woofer., and if one compares these responses eq'd similar and does mathematics, EGD might look the same. Instead we shoud compare measured responses, but of course that is very difficult, most of us don't have access to huge chambers or cotton fields. In-room measurements are plaqued by room - reflections and modes.
https://data-bass.com/#/systems?_k=s5r19y and https://data-bass.com/#/articles/5cc0bc36a75a260004255c88?_k=wk7ktt
This is how a BR radiates IRL https://www.stereophile.com/content/mission-770-loudspeaker-measurements (funny to notice that now JA cuts refex tube response at 400Hz before resonances start happening!)
I've learned that a reflex radiator (and bass horns?) radiates sound one full cycle late at tuning F (x ms related to frequency). We often see summed simulated response of BR compared to sealed woofer., and if one compares these responses eq'd similar and does mathematics, EGD might look the same. Instead we shoud compare measured responses, but of course that is very difficult, most of us don't have access to huge chambers or cotton fields. In-room measurements are plaqued by room - reflections and modes.
https://data-bass.com/#/systems?_k=s5r19y and https://data-bass.com/#/articles/5cc0bc36a75a260004255c88?_k=wk7ktt
This is how a BR radiates IRL https://www.stereophile.com/content/mission-770-loudspeaker-measurements (funny to notice that now JA cuts refex tube response at 400Hz before resonances start happening!)
Last edited:
Something I found on BR:
https://rmsacoustics.nl/papers/whitepaperbassreflex.pdf
I'm not a fan of BR, so this paper supports my beliefs I'm not aiming to play 20Hz at thigh levels, so I think DSP on a sealed box is the way to go for me.
I listened to one of the woofers yesterday, box is damped (and woofer was mounted cone facing outwards this time). I ran a wideband signal again, and listened to some music. Firstly the upper mid was screaming at me, so I EQ'd it down a bit. Then I raised the low end a bit and listened some more. However, I was not really 'impressed' in any way, but one speaker standing in the middle of the room playing one channel with some quick EQ might not be much to go by either. I did some close mic measurements, and an impedance sweep. Low end starts dropping around 100Hz, impedance looks much nicer now with damping in the box, but the basket resonance at 250Hz is clearly visible. There was also some wiggle in the FR around 50Hz. For some reason I have problems with my impedance measurements, odd curve with one sound card, and the one that worked last time is now showing abt double the actual resistance. Not sure what is going on with that. I don't have the usb stick with me now to post any graphs.
https://rmsacoustics.nl/papers/whitepaperbassreflex.pdf
I'm not a fan of BR, so this paper supports my beliefs
I'm not aiming to play 20Hz at thigh levels, so I think DSP on a sealed box is the way to go for me.I listened to one of the woofers yesterday, box is damped (and woofer was mounted cone facing outwards this time). I ran a wideband signal again, and listened to some music. Firstly the upper mid was screaming at me, so I EQ'd it down a bit. Then I raised the low end a bit and listened some more. However, I was not really 'impressed' in any way, but one speaker standing in the middle of the room playing one channel with some quick EQ might not be much to go by either. I did some close mic measurements, and an impedance sweep. Low end starts dropping around 100Hz, impedance looks much nicer now with damping in the box, but the basket resonance at 250Hz is clearly visible. There was also some wiggle in the FR around 50Hz. For some reason I have problems with my impedance measurements, odd curve with one sound card, and the one that worked last time is now showing abt double the actual resistance. Not sure what is going on with that. I don't have the usb stick with me now to post any graphs.
Is it not half a cycle at res frequency? Anyhow me too do not fully understand group delay, but i do not like the sound of breflex anyhow, i prefer well damped sealed boxes.
Here is the nearfield FR and impedance curve of the woofer in the damped box. I used xsim to get the graphs, I'm not familiar enough with vcad yet to easily plot driver measurements. The scaling of the impedance is abt double, but it gives the general idea of the shape and resonances.
The big hump in the midrange has something to do with how I placed the mic, a bit further away it smooths out. The wiggle around 55Hz I mentioned must be from the room, it became stronger the further the mic was from the speaker, so I used the measurement with the mic next to the cone.
Last is a pic of how I ended up 'smoothing' the shelf under the round overs. Not perfect, but a fairly straight forward cut that improved it IMO.
The big hump in the midrange has something to do with how I placed the mic, a bit further away it smooths out. The wiggle around 55Hz I mentioned must be from the room, it became stronger the further the mic was from the speaker, so I used the measurement with the mic next to the cone.
Last is a pic of how I ended up 'smoothing' the shelf under the round overs. Not perfect, but a fairly straight forward cut that improved it IMO.
You can use WinISD to take a look at Group Delay. Create two projects with the same driver, one BR and one sealed. The are two main ways to get the sealed box response to approximate the -3 dB frequency of the BR box: a Linkwitz Transform or a PEQ. Both are in the WinISD filters selection box. You will find it's a series of trade offs between frequency response, group delay, cone excursion and required amplifier power. The seldom used WinISD "Amplifier power apparent load power (VA)" option will show how much additional power a Linkwitz Transform or a PEQ requires.