Subwoofer Comparison

[Keith_W]

Hi All, aside from SPL and frequency response, what is the most important spec for a subwoofer?

Distortion.

 

[PristineSound]

Distortion.

Aside from Perlisten, which sun has the lowest measured distortion?

 

[levimax]

Hi All, aside from SPL and frequency response, what is the most important spec for a subwoofer?

In addition to distortion I think low group delay is also important, not so much because you can hear it directly but many small subs have so much group delay they can cause cancellation when trying to integrate with other subs or even with mains. It is also a sign the subs are operating outside their linear range and are relying on DSP to squeeze out a few more Hz of extension.

 

[levimax]

Aside from Perlisten, which sun has the lowest measured distortion?

Big subs with big drivers https://shop.gsgad.com/

 

[staticV3]

Aside from Perlisten, which sun has the lowest measured distortion?

Here are the subwoofers with the highest measured distortion-free* SPL at 25Hz:


*according to CEA-2010-A

 

[nicolas-p64]

For a 12-inch woofer, with Xmax = 10 mm, the handly SPL calculator at https://www.baudline.com/erik/bass/xmaxer.html produces:
View attachment 559137
An Xmax = 15 mm gives the following SPL, which is about 3.5 dB higher:
View attachment 559138


What are the peak SPLs that you are after? With a sealed subwoofer, more than one subwoofer may be required in order to achieve them.

For a 12-inch woofer, with Xmax = 10 mm, the handly SPL calculator at https://www.baudline.com/erik/bass/xmaxer.html produces:
View attachment 559137
An Xmax = 15 mm gives the following SPL, which is about 3.5 dB higher:
View attachment 559138


What are the peak SPLs that you are after? With a sealed subwoofer, more than one subwoofer may be required in order to achieve them.

Thanks for your reply.

Here are a few more details:
I own a Focal subwoofer with an 11-inch driver (a monitoring model vented) with kef ls 50 wireless 2.
I am perfectly happy with it, but it is unfortunately starting to show signs of wear, which is why I am considering replacing it.

It can reach down to 30 Hz (-3 dB), which is sufficient for my needs.

I have no need to go down to 20 Hz or lower (I use it mainly for music, but also occasionally for movies).

I am therefore considering a replacement. It is capable of a maximum output of 115 dB and genuinely reaches 30 Hz with good volume and quality.

I am looking for a replacement subwoofer, but I have some concerns about the ability of a "sealed" (closed-box) design to deliver high-quality low-end performance—even with a 13- or 15-inch driver—in a fairly large, open-plan room (the space is at least 60 to 80 square meters).

My goal is to achieve 30 Hz performance with good volume and, above all, high quality:
real impact and minimal distortion.

The subwoofer needs to fit into a living room setting, so it should be aesthetically pleasing and not too bulky.

I have identified several potential candidates, though—like most people—I won't be able to audition them before buying...

I am trying my best to make the right choice for my specific situation.

My budget is €2,000.

The candidates I’ve spotted that—on paper—match what I’m looking for are:

SVS SB-3000 (sealed)

Dali K-14 F (vented)

Berlin 12 (sealed)

 

[witwald]

I am therefore considering a replacement. It is capable of a maximum output of 115 dB and genuinely reaches 30 Hz with good volume and quality.

But what is its maxixum SPL at 30 Hz? Is the requirement for 115 dB SPL at 30 Hz at 1 m?

 

[nicolas-p64]

He can ,i think, 105 db at 30hz with high-quality and
at a realistic level in my listening room.

It has scale and impact, with an extremely clean low-end performance that reveals every nuance of the instruments while extending very deep.

So i want the same capability in low frequency

 

[gabo4au]

He can ,i think, 105 db at 30hz with high-quality and
at a realistic level in my listening room.

It has scale and impact, with an extremely clean low-end performance that reveals every nuance of the instruments while extending very deep.

So i want the same capability in low frequency

It's an older design, but hard to beat the HSU ULS15 MK2. Very musical, relatively small (18"x18"x19"), 108db at 20hz, 115 or so at 30, attractive, reasonably priced. They don't advertise a lot, so not as well known as some of he bigger brands, but really good.

 

[nicolas-p64]

thank you!
Since I'm in France, I'll see if it's available (not sure).

 

[witwald]

In addition to distortion I think low group delay is also important, not so much because you can hear it directly but many small subs have so much group delay they can cause cancellation when trying to integrate with other subs or even with mains. It is also a sign the subs are operating outside their linear range and are relying on DSP to squeeze out a few more Hz of extension.

Isn't it the 5–10ms processing latency of subwoofer DSP systems that can lead to the cancellation effects that you refer to?

If the subwoofer and its DSP comprises a minimum-phase system, isn't group delay simply part and parcel of a subwoofer's frequency response? For example, vented subwoofers have about double the group delay of sealed subwoofers for the same low-frequency cut-off point. In any case, the group delay isn't all "bad", as it will be needed to match up with the group delay in the main speakers at low frequencies.

For example, below is the group delay of a passive vented box low-frequency alignment with a –3 dB point of 50 Hz, and the low-frequency group delay shown here is due entirely to the high-pass nature of the transfer function.

 

[witwald]

It's an older design, but hard to beat the HSU ULS15 MK2. Very musical, relatively small (18"x18"x19"), 108db at 20hz, 115 or so at 30, attractive, reasonably priced.

To achieve 108 dB at 20 Hz, a 15-inch woofer needs to have an Xmax of 31 mm. That's a tall order for a magnet motor system. For reasonably linear cone excursion, the maximum output would need to be at least 6 dB less.

 

[witwald]

I got some of the infos wrong in the my tech materials, sorry for that! The lead engineer sent me a correction about the subsonic filter "the Subsonic (Infrasonic) filter introduces group delay by design. It's a 12dB/oct filter. Its purpose is to reduce excessive pressure build-up below 10 Hz in small rooms, which can affect turntables and may be undesirable for some listeners."

Thank you for the clarification. It's important to accurately describe what's happening in the subwoofer. Turntable users will no doubt benefit from such a feature being present in the subwoofer.

Edit 2: Additional notes from the lead engineer:
"A word a about BERLIN Subs low frequency extension and subwoofer system integration. Room gain naturally boosts the lowest frequencies. For that reason, the target anechoic response of BERLIN Subs includes a gentle roll-off to -6 dB at 14 Hz, which we have found provides good in-room performance well into the infrasonic range. ... Extension into the infrasonic range is important because it minimizes phase shift and group delay effects around the crossover region, improving integration with the main speakers. A often overlooked problem.

Many subwoofer manufacturers have approached the problem in a similar manner. The assistance of room gain to achieve flat in-room response with sealed subwoofers is often mentioned in subwoofer product literature.

Many passive loudspeakers exhibit a Butterworth-style low-frequency roll-off, and BW filters are far from optimal for crossovers. SpeakerMatch address this and achieve more predictable integration between passive speakers and subwoofers.

Would you please present a number of examples to illustrate this particular point, as it's not a simple matter to visualise in one's head? That would be very helpful and instuctive.

It definitely does help to understand the need for "matching" of the subwoofer's low-pass filtered response with the high-pass transfer function of the main speakers. The differences between 2nd-order and 4th-order roll-offs in the main speakers is something that does need to be taken into account in the integration process. However, it can be surprising (at least to me) just how well Butterworth filters on the subwoofer can work to integrate with the Butterworth-like low-frequency responses of the main speakers.

 

[levimax]

Isn't it the 5–10ms processing latency of subwoofer DSP systems that can lead to the cancellation effects that you refer to?

If the subwoofer and its DSP comprises a minimum-phase system, isn't group delay simply part and parcel of a subwoofer's frequency response? For example, vented subwoofers have about double the group delay of sealed subwoofers for the same low-frequency cut-off point. In any case, the group delay isn't all "bad", as it will be needed to match up with the group delay in the main speakers at low frequencies.

For example, below is the group delay of a passive vented box low-frequency alignment with a –3 dB point of 50 Hz, and the low-frequency group delay shown here is due entirely to the high-pass nature of the transfer function.

View attachment 559314

DSP processing delay is a separate issue and relatively easy to allow for with DSP. Different group delay curves are harder. Below is a group delay chart for 2 different subwoofers I was trying to integrate together. The green line is an SVS SB3000 and the red line is a DIY sealed sub which had similar output but were larger enclosures. I was trying to use 4 of these (2 each) and all 4 were crossed to the mains at 80 Hz. I was surprised that 4 subs had much lower output at 30 Hz than 2 subs whether the SVS or DIY. Then I looked at the group delay and between 80 Hz and 30 Hz the SVS has ~15 ms more group delay which put it 180 degrees out of phase with the DIY subs which led to very strong cancellation. You can run into similar issues if the mains and the subs have very different group delay curves. To me a group delay curve like the SVS is a sign of excessive focus on small enclosures and advertised extension which require strong bass boost followed by a very steep "protective high pass filter" to stop the sub from self destructing. Basically trading better advertised specs against time domain behavior which few understand or ask about.

 

[smokeymackinsey]

Thank you for the clarification. It's important to accurately describe what's happening in the subwoofer. Turntable users will no doubt benefit from such a feature being present in the subwoofer.

Many subwoofer manufacturers have approached the problem in a similar manner. The assistance of room gain to achieve flat in-room response with sealed subwoofers is often mentioned in subwoofer product literature.

Would you please present a number of examples to illustrate this particular point, as it's not a simple matter to visualise in one's head? That would be very helpful and instuctive.

It definitely does help to understand the need for "matching" of the subwoofer's low-pass filtered response with the high-pass transfer function of the main speakers. The differences between 2nd-order and 4th-order roll-offs in the main speakers is something that does need to be taken into account in the integration process. However, it can be surprising (at least to me) just how well Butterworth filters on the subwoofer can work to integrate with the Butterworth-like low-frequency responses of the main speakers.

I am on a business trip for the around the next 10 days and I don't think I will have the time to find the fitting material in this time - sorry for the inconvenience. But as I said, our lead engineer is working on a white paper that goes into all these topics that should hopefully arrive in around 2-3 weeks. Maybe we can "pause" the discussion until then and re-start it when you guys get much more technical context and information !

On the point of max SPL for the Berlin subs: CTA-2010A is used by some manufacturers, but it has limitations and only paints a very broad picture of the performance, which is exactly why CTA-2010B was developed. That said, I'm not aware of many brands actually using CTA-2010B in their marketing material yet, so using it has very limited use for comparison for our customers. I of course know that a technical forum like ASR a) expects measurements like that and b) knows how to interpret them. As we come from a more consumer-centered view, we did not make either 2010A or B measurements, but we do see now that we need some kind of SPL measurements to go on our spec sheet. I will annoy my colleauges in product development for one !
Another point from our engineer: these standards are designed to find a product's excursion and amplifier power limits using test signals. BERLIN, on the other hand, uses a unique intelligent limiter - with test signals it shows the same limitations as other products, but it really shines with real-world, complex signals. And of course this can be interpreted as marketing talk and again, you have every right to be sceptical, but I hope the white paper will clear things up.
I will return to this threat once it has released!

 

[somebodyelse]

Thank you for the clarification. It's important to accurately describe what's happening in the subwoofer. Turntable users will no doubt benefit from such a feature being present in the subwoofer.

Only the ones who don't already have a subsonic filter in their phono stage, although to be fair that often isn't present on the cheaper ones, and even some of the more expensive ones. It only made it into the DIN standard recently.

 

[gabo4au]

To achieve 108 dB at 20 Hz, a 15-inch woofer needs to have an Xmax of 31 mm. That's a tall order for a magnet motor system. For reasonably linear cone excursion, the maximum output would need to be at least 6 dB less.

Yea, that's just the manufacturers published number, which we all know what that means. Here's a full review with all the measured details if you're interested.

Hsu Research ULS-15 mk2 Subwoofer Review

The new HSU Research ULS-15 mk2 subwoofer is 30% less expensive and 20lbs lighter than its predecessor but does that mean it takes a performance hit? Read on to find out.

www.audioholics.com

 

[rf1938]

I have two Sigberg 10D subs, each with two opposed 10” drivers, very compact sealed enclosure, meets your technical requirements. May be more than your budget, but the exceptional quality of this sub is worth it. Take a look, also made in Europe, so delivery shouldn’t be an issue.

 

[witwald]

That said, I'm not aware of many brands actually using CTA-2010B in their marketing material yet, so using it has very limited use for comparison for our customers.

Technical objective information based on test measurements will always be of importance to your customers. It definitely won't be of "very limited use"!

Another point from our engineer: these standards are designed to find a product's excursion and amplifier power limits using test signals.

And it would be very helpful if those limits were provided as information to consumers.

BERLIN, on the other hand, uses a unique intelligent limiter - with test signals it shows the same limitations as other products, but it really shines with real-world, complex signals.

Intelligent limiters don't seem to be that unique these days. After all, as far as I can tell, they are available in every commercial DAW. Of course, tuning the limiter for low-frequency signal content is part of delivering a good user experience, so no doubt some engineering work would have gone into that side of things. In any case, intelligent limiters are far from unique, as they are available in subwoofers like the KEF KC92 and KC62, the SVS SVS 3000 Micro R|Evolution, and B&W DB4S / DB Series, and probably a few others too. Then there are the sound reinforcement models such as the QSC KS118, RCF SUB 705-AS MK3 / 905-AS MK3, Yamaha DXS18XLF / DXS18mk3, and JBL LSR310S to name but a few.

 

[witwald]

To me a group delay curve like the SVS is a sign of excessive focus on small enclosures and advertised extension which require strong bass boost followed by a very steep "protective high pass filter" to stop the sub from self destructing. Basically trading better advertised specs against time domain behavior which few understand or ask about.

It's not entirely clear how the strong EQ applied to the SVS subwoofer is leading to the type of group delay bahviour that was shown. It is entirely possible to take a sealed subwoofer low-frequency transfer function with a shallow low-Qtc roll-off and heavily equalize it to produce a more Butterworth filter-like 2nd-order response with a –3 dB cut-off at 20 Hz. Of course, the equalized response has the same group delay as what a non-equalized response with the same transfer function would have.

Below is a simulation produced using VituixCAD of just such a heavily-equalized sealed subwoofer. The magenta dotted curve is the "natural" response of a Qtc = 0.50 sealed subwoofer with an Fc = 30 Hz. Its output at 20 Hz is –10.5 dB relative to the passband. The light blue dash-dot curve is the response of a sealed subwoofer with Qtc = 0.7071 and an Fc = 20 Hz, whose response at 20 Hz is –3 dB relative to the passband. The solid black line is the "heavily equalized" response of the original Qtc = 0.50 sealed system. It is evident that the magnitude reponse and group delay response of this equalized system is a very close match with the Qtc = 0.7071 system. Hence, this clearly demonstrates that heavy equalisation does not necessarily result in significantly different group delay characteristics. The heavily equalized subwoofer will also have the same time domain transient response characteristics, assuming of course that it is operating within the driver's linear displacement zone (i.e., not being overdriven).

Now, we can easily introduce the effect of a second-order high-pass filter to simulate those that might be used to "protect" this heavily equalized system. Below is the VituixCAD simulation when an 8-Hz 2nd-order Butterworth high-pass filter is added. Clearly, the group delay characteristics start being significantly affected below about 20 Hz.

Changing to an 8-Hz 4th-order Butterworth high-pass filter produces even more group delay. Of course, this is entirely as a expected. However, the group delay at 15 Hz is only 31 ms, which is a lot lower than the 120 ms in the group delay curve that was provided in an post #1354. Maybe that is indicative of something else?

It is also interesting to see the effect on group delay of an 80 Hz 4th-order Linkwitz-Riley low-pass filter (LPF). Below is the VituixCAD simulation of that particular scenario. Notice how the LPF seems to add an almost constant boost of 6–7 ms to the group delay over most of the operating frequency band of the subwoofer.