I measured 10 subwoofers.

[Tangband]

Very extensive measurements , thank you .

Soundquality from those subs can only be judged by the integration abilities - i.e their flexibility of their LP crossover filters.

As we all know - we never listen to the sound of a single subwoofer - we always listen to the whole system , thus good integration is mandatory for a good sound - more important than low distortion.

 

[Koeitje]

Very extensive measurements , thank you .

Soundquality from those subs can only be judged by the integration abilities - i.e their flexibility of their LP crossover filters.

As we all know - we never listen to the sound of a single subwoofer - we always listen to the whole system , thus good integration is mandatory for a good sound - more important than low distortion.

You can always use a minidsp.

 

[test1223]

Very nice measurements

Can someone please explain the modulation distortion figures and side conditions. What are typically hearing threshold? All subwoofers seems to don't provide good results. Especially the Neumann subwoofer has very high distortion here.

 

[Sokel]

Very nice measurements

Can someone please explain the modulation distortion figures and side conditions. What are typically hearing threshold? All subwoofers seems to don't provide good results. Especially the Neumann subwoofer has very high distortion here.

Is that the problem you see with Neumann?
My main worry is the fact that it stops where the bigger (but same price) Arendal starts in terms of SLP ability,the absolute single most important factor for a sub.

 

[Nuyes]

Very nice measurements

Can someone please explain the modulation distortion figures and side conditions. What are typically hearing threshold? All subwoofers seems to don't provide good results. Especially the Neumann subwoofer has very high distortion here.

The modulation distortion measurement represents the intermodulation distortion that occurs when a speaker produces bass and treble at the same time.

There are several ways to measure this, but the method we used in this event is to lock in one low frequency and measure the high frequency on top of it, which we call a "voice sweep".

When a speaker moves significantly to reproduce bass, there is modulation in both amplitude and frequency in the treble.

And it's dominated by the displacement of the speaker, so in principle, you can't give a one-size-fits-all measurement for all speakers.
In principle, you should set the resonant frequency of each speaker driver as a baseline.
And this is a measurement for driver developers, not for reviewers and consumers.

But at the end of the day, the bass that we universally expect from a subwoofer is at least down to 20Hz, and we expect good quality bass performance regardless of the product and the manufacturer, so...
I've taken measurements using the same criteria and enclosed the results.

So it could be an advantage or a disadvantage for each subwoofer.

Even if they're in the same weight class. If you have a subwoofer with a DSP that's set up to push hard and get loud around 20Hz, it's going to be at a disadvantage in IMD measurements.
Conversely, a subwoofer that is conservative for its size and aiming for a stable curve will have a cleaner IMD.

 

[test1223]

Is that the problem you see with Neumann?
My main worry is the fact that it stops where the bigger (but same price) Arendal starts in terms of SLP ability,the absolute single most important factor for a sub.

I honestly don't know how big a potential modulation distortion problem is in comparison with other abilities, that is why I asked the question. I guess that multi tone distortion is most important since in higher frequencies this is the case. With lower frequencies the distortion perception changes a lot so I am no expert in this regard.

 

[Sokel]

I honestly don't know how big a potential modulation distortion problem is in comparison with other abilities, that is why I asked the question. I guess that multi tone distortion is most important since in higher frequencies this is the case. With lower frequencies the distortion perception changes a lot so I am no expert in this regard.

I not talking about differences,I'm talking about sheer SPL.

Look at the numbers below:

Arendal 1961 1V

Newmann KH750

The difference is huge 5-10db.

 

[fzst]

I not talking about differences,I'm talking about sheer SPL.

Look at the numbers below:

View attachment 322120 Arendal 1961 1V View attachment 322122 Newmann KH750

The difference is huge 5-10db.

Well of course but that's to be excpected since the Arendal is a huge box (compared to the Neumann) and it's a ported design, hence providing a big boost in lower bass.
Furthermore it's only a10" driver compared to the 12" of the Arendal.
So it' pretty obvious that it will murder the Newmann across all bass frequencies in terms of sheer output.

If anything, we should compare the sealed version of the Arendal to the Newmann since they are about the same size.

 

[Sokel]

Well of course but that's to be excpected since the Arendal is a huge box (compared to the Neumann) and it's a ported design, hence providing a big boost in lower bass.
Furthermore it's only a10" driver compared to the 12" of the Arendal.
So it' pretty obvious that it will murder the Newmann across all bass frequencies in terms of sheer output.

If anything, we should compare the sealed version of the Arendal to the Newmann since they are about the same size.

You're right about all the above,but I just compared only what can be had with about the same money.
Not a difficult choice if I was out for a sub,considering the rest of the measurements are pretty decent too.

Edit:From 40Hz and up the sealed Arendal (1961 1S) also has a huge advantage in terms of SPL and is much cheaper too.

 

[fzst]

You're right about all the above,but I just compared only what can be had with about the same money.
Not a difficult choice if I was out for a sub,considering the rest of the measurements are pretty decent too.

I agree with that of course, it's always the best idea to buy a ported sub if you want to maximise output per dollar.

A lot of us (including me) don't want one or more huge boxes in our rooms though (or prefer the sound of sealed subs) be it because of looks or because of the more flexible placement options.
So me personally, I'm more interested in how much output can be had for a given size - even if that's more expensive.

That's why I bought an Arendal 1723 1S and I'm pretty happy with it in terms of output and depth in my room. It also sounds better than the two SVS SB 2000 Pros I had previously, much more punchy.
I still want to get a second one though to even out the bass response.

 

[test1223]

The modulation distortion measurement represents the intermodulation distortion that occurs when a speaker produces bass and treble at the same time.

There are several ways to measure this, but the method we used in this event is to lock in one low frequency and measure the high frequency on top of it, which we call a "voice sweep".

When a speaker moves significantly to reproduce bass, there is modulation in both amplitude and frequency in the treble.

And it's dominated by the displacement of the speaker, so in principle, you can't give a one-size-fits-all measurement for all speakers.
In principle, you should set the resonant frequency of each speaker driver as a baseline.
And this is a measurement for driver developers, not for reviewers and consumers.

But at the end of the day, the bass that we universally expect from a subwoofer is at least down to 20Hz, and we expect good quality bass performance regardless of the product and the manufacturer, so...
I've taken measurements using the same criteria and enclosed the results.

So it could be an advantage or a disadvantage for each subwoofer.

Even if they're in the same weight class. If you have a subwoofer with a DSP that's set up to push hard and get loud around 20Hz, it's going to be at a disadvantage in IMD measurements.
Conversely, a subwoofer that is conservative for its size and aiming for a stable curve will have a cleaner IMD.

Thank you for the explanation. So the spl of the 20Hz tone is different for every subwoofer in these measurements? How loud is the 20Hz tone?

Did anyone know a perception threshold for such distortion?

It would be very good to include a rough perception threshold in every none linear distortion figure. I wonder why nobody did that? Even the harmonic distortion figures would benefit a lot even if these measurements are much better understand by a lot of people...

 

[DJBonoBobo]

I think it's best to think of the Neumann as part of a system with DSP, Mic, software and other speakers and compare price/value of the whole thing with other systems.
It is reassuring to see the sub itself is obviously very well made for a small, sealed, 10 inch. But it is probably not the best choice for the money outside a complete Neumann system. Good to see constant good results from Arendal (also on EAC).

 

[Sokel]

Thank you for the explanation. So the spl of the 20Hz tone is different for every subwoofer in these measurements? How loud is the 20Hz tone?

Did anyone know a perception threshold for such distortion?

It would be very good to include a rough perception threshold in every none linear distortion figure. I wonder why nobody did that? Even the harmonic distortion figures would benefit a lot even if these measurements are much better understand by a lot of people...

You talk about this?

 

[Nuyes]

Thank you for the explanation. So the spl of the 20Hz tone is different for every subwoofer in these measurements? How loud is the 20Hz tone?

Did anyone know a perception threshold for such distortion?

It would be very good to include a rough perception threshold in every none linear distortion figure. I wonder why nobody did that? Even the harmonic distortion figures would benefit a lot even if these measurements are much better understand by a lot of people...

I can't find all the relevant research, but here's what I know, at least from what I've seen.
Finding an audible threshold for the IMD of a subwoofer would be very complicated.

First of all... Unlike THD, you need a specific situation.

If it's THD, it's relatively straightforward to quantify that frequency as the sum of the overtone distortion when it's played at a certain SPL, so it's possible, and there's a lot of research on audible thresholds.

However, we need at least two or more frequencies to reproduce IMD, and in real-world audio signals like music or movies, the ratio of treble to bass is not always the same.
Also, even if there is some distortion, masking can occur.

The second reason this is difficult has to do with crossovers.

THD is a form of canonical nonlinear distortion, meaning that as long as you play those frequencies at the same loudness, they will always appear distorted to the same degree.

However, with subwoofer IMD, unlike testing, you're usually leaving the front speakers to handle anything above 60-80Hz, so unless you have an LFE-dedicated subwoofer that has to carry it all the way to 120Hz, it's unlikely to be audible.

 

[Martigane]

Fantastic report, @Nuyes . Thank you!

What I was missing was a global overview, and since I did it for a friend who's looking for a subwoofer, I thought I'd share it with you:

5=best performance
1=worst performance

Difficult to establish a high level score when things are not comparable sometimes ; so I've filed the blanks using some of my experience on correlating measurements and subjective impressions.
Let me know if there are discrepancies with your observations, I can correct it. (some probably remain as it takes time to do this - hopefully small)
Here's the data if you want to comment/reuse:

ASR Subwoofer Comparison

Subwoofer Original thread: <a href="https://www.audiosciencereview.com/forum/index.php?threads/i-measured-10-subwoofers.49042/">https://www.audiosciencereview.com/forum/index.php?threads/i-measured-10-subwoofers.49042/</a> Low Extension,Max SPL,Harmonic distortion,Intermodulation distortion,Time...

docs.google.com

Next, if time allows, it'd be nice to have a table summarising the technology used for each subwoofers (closed/ported, woofer size , volume, etc..)

 

[fzst]

You talk about this?
View attachment 322142

No, that's the threshold for hearing certain frequencies. It's the basis for things like Denon's Dynamic EQ, which boosts certain frequencies at lower volumes because the human hearing doesn't perceive all frequencies linearly so the deeper bass frequencies and highs have to be boosted at lower volumes to be perceived at an equal loudness to the mids, to which we are most sensitive to.

 

[test1223]

You talk about this?
View attachment 322138

This is the perception threshold of a pure tone and the equal loudness curves.

There are other none linear curves for every single none linear distortion. Like one curve for the h2 harmonic distortion and so on.

I can't find all the relevant research, but here's what I know, at least from what I've seen.
Finding an audible threshold for the IMD of a subwoofer would be very complicated.

First of all... Unlike THD, you need a specific situation.

If it's THD, it's relatively straightforward to quantify that frequency as the sum of the overtone distortion when it's played at a certain SPL, so it's possible, and there's a lot of research on audible thresholds.

However, we need at least two or more frequencies to reproduce IMD, and in real-world audio signals like music or movies, the ratio of treble to bass is not always the same.
Also, even if there is some distortion, masking can occur.

The second reason this is difficult has to do with crossovers.

THD is a form of canonical nonlinear distortion, meaning that as long as you play those frequencies at the same loudness, they will always appear distorted to the same degree.

However, with subwoofer IMD, unlike testing, you're usually leaving the front speakers to handle anything above 60-80Hz, so unless you have an LFE-dedicated subwoofer that has to carry it all the way to 120Hz, it's unlikely to be audible.

I know for some distortions you have to calculate the specific threshold which isn't trivial in a lot of cases. And it isn't a tasks /issue one individual can solve. Nevertheless I want to address this problem here.

My thoughts are the following: most measurements aren't worth a lot without the perception threshold information. At least you have to have a rough distortion limit in mind to come to a meaningful interpretation of the measurement. So all these measurements are nice and some can easily be used but other measurements might only help a hand full of people who know more about such special thresholds and all other people might even come to wrong conclusions with these measurements. Maybe some members here who are more involved can put these things into the measurement programs or include some simpler harmonic distortion thresholds in their own measurements and start make the figures much more friendly for a correct interpretation.

 

[thulle]

The modulation distortion measurement represents the intermodulation distortion that occurs when a speaker produces bass and treble at the same time.

Just to clarify my thinking, and continue into a question that probably has been answered elsewhere already, but here we go: graph says "Bass tone = 20 Hz" and then we have Amplitude Modulation, 2nd-, 3rd order intermodulation, and a sum of all 3.
This in a graph over 20-200Hz where U2/U1 = 12dB.

Brain still asleep, but to explain how I interpret this, please correct me if I'm wrong.

U1 would be the 20 Hz tone, so we're playing the measured tone 12 dB louder than the 20 Hz one.
2nd order intermodulation distortion would be U2±U1, 2xU1 and 2xU2 Hz, so at 75 Hz measured we're looking at 55, 95, 40 and 160 Hz components. 40 and 160 also being harmonic distortion components I'm not sure if they're included here?
3rd order intermodulation would be 2xU1±U2 & 2xU2±U1", so at 75 Hz measured we're looking at (-35?),115, 130 and 170 Hz.

Personally I have two subwoofers to my desktop setup, one each at the walls to my sides. Setting the crossover to my desk speakers higher than ~80-90 Hz makes the directionality of where I perceive sounds to come from shift to the sides. I assume I get sounds coming from the subwoofers that are high enough in frequency that they're above the limit of where I lose ability to perceive direction.

The 95 and 160 Hz 2nd order IMD might be an issue for this, same with 115, 130 and 170 Hz 3rd order.

Checking the SB-1000pro we're at -10 dB when tested at @ 20+75 Hz tones, -23 dB for 3rd order IMD.
Checking your L26-ROY (excellent choice) we're at ~ -23 dB, and -30 dB for 3rd order. huge improvement.

Even if we remove some dB to ignore the components below crossover cutoff, wouldn't these levels make it hard to smoothly integrate SB-1000pro with the other speakers? Looking at the numbers, 70 dB SPL spray above 100 Hz with ~80 dB SPL at 95 Hz seems like something to improve.

I've probably read some things about the percievability of low-frequency IMD, but nothing comes to mind right now, but consider a worst case scenario of some music having a pause in everything non-bass, wouldn't the distortion from the subwoofer make its location in the room suddenly pop out as a source of noise way worse than the harmonic distortion?

I usually try to think these things through a bit more thoroughly before posting but haven't got the time right now, so I'm sorry if this is a bit uncooked.

 

[Ler0ck]

It must be a longtime works. Thanks

 

[ocinn]

thanks so much for this.

This post has me really mourning the loss of data-bass.com

FWIW in my professional and enthusiast experience, DIY for subwoofers is a no brainer. So many amazing plans out there, and you can easily beat the Neumann performance for 1/3rd the cost and an afternoon of wood cutting and assembly.