I measured 10 subwoofers.

[-Matt-]

Thanks for your hard work and for the measurements.

A small detail but... the first plots (80dB ones) for the Polk audio PSW111 and the Neumann KH750 DSP start at 10Hz whilst the others start at 20Hz. Consider making these consistent for easier direct comparison.

 

[radix]

That is an open baffle kit, which may not be to everyone’s liking. Open baffle costs a lot of bass (from rear wave cancellation).

Oh yes it is. I thought the sub enclosure was closed. I'll edit the post.

 

[Maiky76]

Hi, it's been a while!
I would like to thank the people who made this project possible on a scale never before imagined in South Korea. I measured a total of 10 subwoofers, including my own DIY subwoofer(L26RO4Y), and measured them to the CEA-2010 standard using a Klippel TBM module. Below are photos and a list of the subwoofers I measured.

View attachment 321682

(In order from left to right)

- Arendal 1961 1V

- Arendal 1961 1S

- Neumann KH750

- Bowers & Wilkins DB4S

- SVS SB-3000

- SVS SB-2000 pro

- SVS SB-1000 pro

- Polk audio PSW111

- My DIY Subwoofer

- SVS 3000 micro

This is officially the first time CEA-2010 has been measured in South Korea.
So there was a lot of trial and error, which took a lot of time.
And for the products that were not loaded enough, I did a second measurement under slightly harsher conditions with the owner's permission, and added data.

View attachment 321681
View attachment 321680

With that out of the way, let's get down to business, releasing data starting with the SVS SB-1000 pro model.
All subsequent data will be presented in the same format, with the only difference being the product name.

As I'll explain later, there are a lot of variables in active subwoofers, so it's not possible to compare products on a fixed basis.

SVS SB-1000 pro


View attachment 321683


-21.2dBu means that we measured with an input of 80dB SPL@2m (100Hz) and an output of 80dB SPL@2m (100Hz).
(I state this because it can vary depending on the volume and gain knob settings on the product).

THD
View attachment 321684View attachment 321685View attachment 321686View attachment 321687

Group delay
View attachment 321688View attachment 321689

ETC
View attachment 321690

Measure the attenuation level for 100ms after the peak (0dBFS).
Higher attenuation is considered ideal.

View attachment 321691

Now it's finally time to measure the TBM module. First, let's take a moment to explain how the TBM module works.

The TBM module measures a preset frequency, one cycle for each voltage, gradually increasing the voltage. If the THD value of the measured signal exceeds the threshold, it moves on to the next frequency. In other words, if you measure 20 Hz and the THD crosses the threshold, you move on to the next measured frequency, 25 Hz.

This is where the start voltage and maximum voltage settings become very important. The reason for this is that if you make the measurement too harsh, you risk damaging certain products, and if you make it too soft, you lose discrimination between products. Unfortunately, since I had to measure 10 subwoofers in one day with a lot of trial and error, I couldn't find the optimal measurement level for each product.

Therefore, I only set the minimum input gain and SPL for each product, and then replicated the rest of the conditions and measurement module settings.
So keep in mind that what may seem like an advantage for some products may not work for others.

However, please understand that we had to take this conservative approach because protecting the product is our number one priority.

View attachment 321695

The following introduces the THD thresholds for each frequency in this module.

According to the CEA-2010B standard, there are different THD thresholds for each frequency.

The two thresholds above indicate that,
If you don't intuitively understand what this means, it means that there are different values of acceptable THD for each measurement frequency!

In other words, at lower frequency measurements, we're more forgiving, but as we move to higher frequencies, we get stricter about THD.

And this applies as follows

View attachment 321696

If the THD of the measured signal is all below the threshold, the corresponding frequency measurement is PASS!

View attachment 321697

If any of them touch the threshold, that measurement will FAIL!

And if it fails, then we just move on to the next frequency measurement,
We have a threshold level, called [Neglect Threshold Below], that determines if we should try to resume the measurement with a higher voltage.

For this measurement, it doesn't matter much because I mostly set this value equal to the maximum voltage.
In other words, even if it fails, it tries to increase the level unconditionally up to the preset value.

Peak Value
View attachment 321692


If you've never seen this plot before, you might be confused.
Basically, the more closely spaced the frequencies, the better the effective bandwidth, and the straighter the line, the better the output.

However, each subwoofer has a different preset output value from the manufacturer, so it's not a perfect comparison.

View attachment 321693

View attachment 321694


Looking at this data, you should see some of the PASS! and FAIL! we discussed earlier.


The orange ones at relatively low voltages are the ones where background noise has crossed the THD threshold, so if you see green values at higher outputs, you can ignore anything below that.
(This is where the sophisticated 'Neglect Threshold below' setting comes in).


The first thing to look at is the top green value for each frequency.
That's the maximum output of that subwoofer, at least for the current measurement.

But that's not all.

Each measured signal increases by exactly 1 dB,
There are many instances where the actual measured SPL value is less than 1 dB, even at frequencies where a passing score is achieved.

In other words, the THD didn't touch the threshold in the 2010B standard, but it was already eating up compression in the lower range.

This is where you need to be proactive in your interpretation of each product.

View attachment 321698

View attachment 321699


The maximum voltage that the product can accept in the current measurement setup.
(More precisely, the voltage it can take within the limits of its normal output.)


View attachment 321700

THD% at the PASSed point for each frequency.

View attachment 321701



THD% by output.
Each frequency is color coded.

I hope you enjoy these measurements.
I'll leave you with the data for the remaining nine subwoofers in order, along with their names.



+++++

I apologize for not attaching a more detailed description and evaluation of the data and plots.

I measured 10 subwoofers.

Wahou ! thank you !

audiosciencereview.com

Great effort!

In case some are interested:

 

[sweetchaos]

I also added this data to my spreadsheet!

 

[Hellasärö]

Very nice work. Do you have the build documented somewhere I can read or watch?

Unfortunately not, because I had them custom made, I didn’t build them myself. The guy that built them (Harrin Kaiutin) is known here in Finland for his well measuring speakers and beautiful wooden finishes. Here’s a pic of his flagship model with a cardiod mid.

 

[thulle]

While I prefer baltic birch ply for my builds (both speakers and subs), many diy subs are mdf.

The point of using plywood, IIRC, is to make sure that the resonance frequencies end up above the range of the subwoofer. Might not be a non-issue if the subwoofer is small enough.

 

[DanielT]

While I prefer baltic birch ply for my builds (both speakers and subs), many diy subs are mdf.

MDF for subs. Of course that works too. I've even built a subwoofer out of a cardboard tube. It went well. Sounded surprisingly good, but that's what a chef always thinks about his own cooked soup. The sub at the top right of the image (cluttered with stuff in my in my living room but still).

EDIT: I thought the sub was enclosed, but it is also open baffle like the other half of the speaker. So, no back on the enclosure.

If someone in the US is looking for a DIY L26RO4Y with less "Y," you could try the kit from Madisound. It's a 13x13x10 flatpack enclosure for $64 and you can add the driver for another $400 ($470 for the whole thing with mounting hardware). You still need a DSP and amp.

Open Baffle Subwoofer Flat Pack (LXsub2)

Open Baffle Subwoofer Flat Pack for 10" Subwoofer (LXsub/LXmini+2) - EACHUsed as the +2 subwoofer option for the LXmini speaker kit, and cut for the Seas L26RO4Y. Order quantity 2 if building this design. A single, dual, or multi sub setup with these is a good choice for integration with other...

www.madisoundspeakerstore.com

I see you added an EDIT but generally about open baffle subs. Open baffle subwoofer sounds like a contradiction in itself. If we are talking about getting reasonably low sub-bass frequencies, that is. To show the difficulties of getting down in frequency see the thread below. There, Joseph Crowe has built an open baffle with a 15" driver and you can see for yourself how low (not so) in frequency it goes. He measured it:

Open baffle with SB Audience Bianco-15OB350. FR + distortion measurement outdoors and indoors.Plus measurements FR on IEC baffle.

I thought this was interesting. Might be of interest to someone who is thinking of building an open baffle speaker. Too big speakers for me to fit in my living room/ listening room (plus I have other projects going on) but for those who have room for such big ones, why not? They seem to require...

audiosciencereview.com

Open baffle costs a lot of bass (from rear wave cancellation).

Exactly.

 

[Robbo99999]

Great effort!

In case some are interested:

View attachment 322364

I think this is the most important comparison information regarding the subs tested! It took me a while upon reading the initial review that this graph @Nuyes did for each sub was the best way to visualise it's performance capabilities, unless my assumption is wrong! My understanding is that this graph is created by plotting the max recommended SPL for any given frequency whilst staying below the detectable distortion limits; however, how reliable are these distortion limits, how well have those distortion limits been determined, how valid are those distortion limits, is it conceivable that some people in some instances will be able to recognise or unknowingly experience the negative impacts of distortion levels below the recommended levels? (These are questions for anybody with detailed knowledge on this topic, not necessarily just you @Maiky76 )

(I'd say the Arendal 1961 1V is the best performing of the lot, based on that graph).

 

[Sokel]

(I'd say the Arendal 1961 1V is the best performing of the lot, based on that graph).

For the 20-32Hz area is alone,yes.
And for me this is the meaningful actual range of a sub with the very top being at 80's (unless you want to hear the baritone's voice coming spitted from a place far from your mains)

 

[Robbo99999]

For the 20-32Hz area is alone,yes.
And for me this is the meaningful actual range of a sub with the very top being at 80's (unless you want to hear the baritone's voice coming spitted from a place far from your mains)

I agree it performs the best right down there vs the other subs. (I've got my SVS SB1000 Pro crossed over at 120Hz, but it is right inbetween my speakers so I don't notice any sound splitting off. I tested it against a 90Hz crossover with open ports on the speakers and I preferred the 120Hz with closed ports subjectively, I noted that more detail in the bass and also in the areas above the bass - measured frequency response was virtually identical at listening position between the two )

 

[Sokel]

I agree it performs the best right down there vs the other subs. (I've got my SVS SB1000 Pro crossed over at 120Hz, but it is right inbetween my speakers so I don't notice any sound splitting off. I tested it against a 90Hz crossover with open ports on the speakers and I preferred the 120Hz with closed ports subjectively, I noted that more detail in the bass and also in the areas above the bass - measured frequency response was virtually identical at listening position between the two View attachment 322380)

Do you have a measurement with only mains at play?
I like your slope,i suppose this is before EQ.

 

[Hellasärö]

MDF for subs. Of course that works too. I've even built a subwoofer out of a cardboard tube. It went well. Sounded surprisingly good, but that's what a chef always thinks about his own cooked soup. The sub at the top right of the image (cluttered with stuff in my in my living room but still).
View attachment 322378


I see you added an EDIT but generally about open baffle subs. Open baffle subwoofer sounds like a contradiction in itself. If we are talking about getting reasonably low sub-bass frequencies, that is. To show the difficulties of getting down in frequency see the thread below. There, Joseph Crowe has built an open baffle with a 15" driver and you can see for yourself how low (not so) in frequency it goes. He measured it:

Open baffle with SB Audience Bianco-15OB350. FR + distortion measurement outdoors and indoors.Plus measurements FR on IEC baffle.

I thought this was interesting. Might be of interest to someone who is thinking of building an open baffle speaker. Too big speakers for me to fit in my living room/ listening room (plus I have other projects going on) but for those who have room for such big ones, why not? They seem to require...

audiosciencereview.com

Exactly.

Sorry for going OT a bit. But if one wants more sub-frequencies from a dipole, one should try a ripole-dipole where the drivers are facing each other. This lowers the tuning frequency. In my subs the tuning/resonance went from 31hz to 22hz. I have a pair with SB Audience Bianco 15” drivers and they work well, but need a lot of distance from the front wall.

Sorry for text being in Finnish, but the different colours are the different slot depths and how it affects the tuning and response.

 

[Willem]

Thanks for all this info. My remaining question is how much of all this is audible, apart from the obvious differences in spl at particular frequencies? I ask, because the logic behind the CEA metric is that all you need to know is how much undistorted output you get at particular frequencies.
I ask because I have three different subwoofers in my system, acquired at different moments in time but now playing together (main speakers are Quad 2805 electrostats high passed at 80 Hz): a B&W PV1d with dual opposed 8 inch drivers, a KEF Kube 8b, originally bought for my desk top system, and most recently a SVS SB2000 classic. The current set up uses MSO/2x4HD to equalize room response of the subs. When used individually the PV1d seemed to be tighest and cleanest, and I wondered why this could be, and if it was just an illusion.

 

[Miiksuli]

Diy subwoofers are worthy. I have also made two 10" passive, closed with passive high/low filter. It's decent but there was tiny problem with the filter. It didn't lower high frequencies enough as I hoped. More high frequencies were lowered the more it changed main speaker frequencies. Anyway it was fun project.

 

[Robbo99999]

Do you have a measurement with only mains at play?
I like your slope,i suppose this is before EQ.

That was after EQ in the bass to remove peaks (up to 200Hz if I recall). The speakers are JBL 308p Mkii, and they are EQ'd Anechoic Flat from Amir's measurements (he reviewed the 308p). With blocked ports I have to crossover at 120Hz if I recall correctly, because they start to roll off a bit lower than that. (With blocked ports nothing changes in the frequency response apart from the bass).

EDIT: found measurements of the 308p with blocked and open ports:

 

[Dave54321]

There is a consistent distortion peak at ~35hz for each subwoofer. Artifacts? If not, is the distortion significant?

Also someone mentioned wishing to see a Rythmik sub measured. Does anyone know has there been any third party measurements which demonstrate a significant benefit to a servo sub? Improved distortion?

 

[MAB]

Thanks for all this info. My remaining question is how much of all this is audible, apart from the obvious differences in spl at particular frequencies? I ask, because the logic behind the CEA metric is that all you need to know is how much undistorted output you get at particular frequencies.
I ask because I have three different subwoofers in my system, acquired at different moments in time but now playing together (main speakers are Quad 2805 electrostats high passed at 80 Hz): a B&W PV1d with dual opposed 8 inch drivers, a KEF Kube 8b, originally bought for my desk top system, and most recently a SVS SB2000 classic. The current set up uses MSO/2x4HD to equalize room response of the subs. When used individually the PV1d seemed to be tighest and cleanest, and I wondered why this could be, and if it was just an illusion.

I think it's that the CEA metric is about maximum undistorted output a system can produce across frequency, but it says nothing about how the speaker is EQ'ed in your room and therefore how it is actually sounding! I am guessing your PV1d is EQ'ed to a curve that matches your preference for bass. The CEA metric is just going to tell you about useable output, not how it is used!

 

[Willem]

I realize that in-room response is highly room dependent and hence I use MSO. However, rightly or wrongly I also had the distinct impression that the PV1d itself was sounding cleaner. The measurements posted here by Nuyes do show quite distinct differences other than CEA, so I am wondering if those differences are indeed audible.

 

[Left]

Interesting about the relatively poor time domain frequency performance of the SVS subwoofers. Any thoughts of why? I too naively thought the SVS Micro 3000 would do well here. Perhaps this will spur SVS to do better.

 

[Spocko]

I have a question to you as a Neumann user with two KH 750:
If I would combine a pair of KH 150 or a pair of the KH 120 II with two of these KH 750 Subwoofers (one for the left Channel and one for the right Channel) with a crossover Frequency of -let’s say- 100 Hz: is it possible to „tell“ the two subwoofers, that they have to handle a stereo signal with separate signals for left and right also below 100 Hz?
Or will they handle every signal below the crossover frequency as a Summn Signal - e. g. a Signal that is Mono (as it would be the case if you would just use one single KH 750 subwoofer)?
Are here (same Szenario) any differences between Neuman and Genelec and their dedicated subwoofers?

Do not chase the illusory unicorn of "stereo sub" because short of nearfield placement, if you place your subs at typical distances from the MLP, everything below your room's transition frequency are summed into wave interaction requiring multi-sub optimization and lacking any semblance of directionality.