Subwoofer Selection Criteria

[Soniclife]

SAM increases distortion? I thought it was supposed to lower it?

Damn Devialet and their mysteriousness....

You cannot have more extension without increasing distortion. Sam adds bass extension, lowers group delay, and adds protection.

 

[watchnerd]

My Sonus Faber guarneri evolution have similar sized everything to your speakers, I'm finding a sub is magic, and crossing them over is great, I have more work to do to get them perfect, and a proper sub purchase is on my list.

Always loved the looks of Sonus Faber and have found them easy to listen to forever on end. Just lovely.

What sub are you using with them, where is it located, and how are you crossing things over?

 

[watchnerd]

You cannot have more extension without increasing distortion. Sam adds bass extension, lowers group delay, and adds protection.

Makes total sense, of course.

 

[Soniclife]

What sub are you using with them, where is it located, and how are you crossing things over?

When my Devialets went off for an upgrade I borrowed my tiny Eve audio 2.1 system from my home office, I just plonked the system down in the lounge like a normal person would for a short term setup, it sounded great, in a crap this costs less than 5% of my main system and I really like it way. Anyway this sub is way too small for the room, I think the driver is slightly smaller than in the SF, and it only has one, but the bass was clearly better even if the rest of the range wasn't. So when my amps came back I've been playing with it where I initially dumped it, and adjusting delay and crossover in the configurator. It's all sub optimal but far too much heavy real life stuff is going on currently, plus the living room furniture is going to change layout, mainly to get a better listening position, so sub buying is on hold till everything is sorted.
I'm tempted by the small sub behind each main idea, as that's dead space anyway, and it frees up crossover point. I think 2 * 10" will be more than enough in my room, which is all solid walls.

 

[watchnerd]

When my Devialets went off for an upgrade I borrowed my tiny Eve audio 2.1 system from my home office, I just plonked the system down in the lounge like a normal person would for a short term setup, it sounded great, in a crap this costs less than 5% of my main system and I really like it way. Anyway this sub is way too small for the room, I think the driver is slightly smaller than in the SF, and it only has one, but the bass was clearly better even if the rest of the range wasn't. So when my amps came back I've been playing with it where I initially dumped it, and adjusting delay and crossover in the configurator. It's all sub optimal but far too much heavy real life stuff is going on currently, plus the living room furniture is going to change layout, mainly to get a better listening position, so sub buying is on hold till everything is sorted.
I'm tempted by the small sub behind each main idea, as that's dead space anyway, and it frees up crossover point. I think 2 * 10" will be more than enough in my room, which is all solid walls.

I basically did the same thing with the old M&K sub I had in the shed... just plonked it down between the speakers, right of center, roughed it in my ear and HP filtered the mains a tad...and was surprised that such crudity is actually an improvement, even if far from optimum.

 

[Soniclife]

I basically did the same thing with the old M&K sub I had in the shed... just plonked it down between the speakers, right of center, roughed it in my ear and HP filtered the mains a tad...and was surprised that such crudity is actually an improvement, even if far from optimum.

The most interesting thing for me is the biggest improvements are not in the obvious areas of grunt, but in all the other areas where the sub isn't doing anything. There are huge gains to be had from getting this right.

 

[dkinric]

I'll admit that my initial bias (probably from some combination of AV-centric thoughts on sub processing, plus how I run my nearfield monitors in the studio) is the former: sats should be high passed.

But....

1. My mains / sats (Dynaudio Contour 20) are pretty big for stand mounts, designed and voiced to be nearly full range, have a pretty high excursion driver for its (18cm) size, and my amp has lots of power to spare.

They also don't have the same kind of dynamic range constraints you get with a smallish 5" wall-mounted AV sat.

So that would seem to mean, I don't want subs because my sats have inherently weak bass; I want a sub because my room has issues.

2. Since I'm correcting for the room, and not trying to make up for dynamic limitations of the sats, I can see the case in hypothesizing that a high pass on the mains may do more harm (yet another circuit, albeit DSP, isn't the use case the speaker designers envisioned when they voiced the speaker) than good (hypothetically helping extremely rare cases of headroom limits).

3. Running full range would let me more properly use Devialet's SAM, which seems to make for lower Q / tighter bass.

Sounds like we have similar scenarios, although your speakers are very nice, in a different class than mine (the Contours are rated down to 39hz!). There are no issues running our speakers full range and we have enough power for it.
If my integrated had DSP or bass management, I would probably use it, but I see no reason to introduce a new component in the chain just for that. I opted for the sub with DSP instead.
I still think you may enjoy adding a sub, but a lot of it depends on your preferences and the type of music you listen to as to how much it comes into play and if it's worth it to you. You could always try one out, and be aware of the return policy.

 

[DonH56]

I'm going to try to ask this in a way that doesn't sound pedantic, as I mean it seriously:

For a given main, what qualifies as deep bass?

Because every driver has excursion limits / F3 / Fs resonances, what qualifies as "deep bass" will very from one main to another.

Would it be safe to assume:

=< F3: deep bass, high distortion >F3: gray zone, middle distortion >2 x F3 (an octave higher): safe, low distortion zone

I wasn't thinking of drivers but how we hear. Below about 80 Hz most folk cannot localize the bass and so that is what I consider "deep". That is where it changes from being perceived as "sound" to something we "feel". But, there are probably about N+1 definitions of deep bass for every N audiophiles...

Musically, bass is below C4 (260 Hz, middle C) and sub-bass is below 60 Hz (C2, deep C).

 

[andreasmaaan]

For best integration I like toplay the main speakers at full range and only use a low pass filter on the sub.

This approach gives up one of the main benefits of using a sub IMHO: reducing distortion-inducing stress on the speakers. I’d try this only as a last resort, ie after all reasonable attempts at integrating the sub(s) at a higher xo frequency have failed.

 

[watchnerd]

This approach gives up one of the main benefits of using a sub IMHO: reducing distortion-inducing stress on the speakers. I’d try this only as a last resort, ie after all reasonable attempts at integrating the sub(s) at a higher xo frequency have failed.

This is logical but I always struggle with how to actually objectively determine the crossover sweet spot for the benefits distortion reduction vs integration.

People seem to default to 80 hz for localization reasons, but I've never seen much methodology applied to empirically answer the question "if not 80 hz, what is right?"

 

[andreasmaaan]

This is logical but I always struggle with how to actually objectively determine the crossover sweet spot for the benefits distortion reduction vs integration.

People seem to default to 80 hz for localization reasons, but I've never seen much methodology applied to empirically answer the question "if not 80 hz, what is right?"

I think one reason for this is that the room is dominant in this frequency range, meaning that the answer will be room-specific. But as a general rule, the closer the subs are to the mains and the shorter the path length differences are at the listening position, the higher a frequency at which it will be acceptable to cross over and the shallower a filter slope it will be acceptable to use.

 

[watchnerd]

I think one reason for this is that the room is dominant in this frequency range, meaning that the answer will be room-specific. But as a general rule, the closer the subs are to the mains and the shorter the path length differences are at the listening position, the higher a frequency at which it will be acceptable to cross over and the shallower a filter slope it will be acceptable to use.

But to your original point:

High passing the mains reduces distortion.

Okay, makes sense. So how do I determine the optimal distortion reducing crossover point?

 

[DonH56]

This is logical but I always struggle with how to actually objectively determine the crossover sweet spot for the benefits distortion reduction vs integration.

People seem to default to 80 hz for localization reasons, but I've never seen much methodology applied to empirically answer the question "if not 80 hz, what is right?"

You can look at loudness curves to estimate where the power lies (integrate the curves) and balance the power loading in your system. I don;t recall the breakpoint off hand but I think it is around 300 Hz or so (half the energy above half below). But that says nothing about what the speaker's drivers can actually do... For many years my rule of thumb has been to use a crossover an octave above the -3 dB point since most loudspeakers distort heavily when large-amplitude signal are applied, and the largest amplitude tend to be in the bass region. sometimes I'll soften to half an octave depending upon the speaker and whether I think distortion or localization is more important. If your main speakers have a -3 dB point of 80 Hz then setting the crossover at 160 Hz is going to play havoc with the stereo image unless the subs are co-located with the mains, and that is not likely to be the best place to smooth the in-room bass response.

That is one reason I've always gone for larger mains -- I want to be able to set the crossover low enough that I cannot localize the sub(s) and thus can place them where they work best for the room. Since crossovers are not brick walls the mains must be able to handle significant power well below (say an octave assuming half-loudness ~10 dB and typical 12 dB/octave crossovers) the crossover frequency (and the subs an equal amount above assuming a symmetric crossover). That is also why I have preferred higher-order crossovers (e.g. the 24 dB/oct L-R crossover I mentioned before, and I am using the 24 dB/oct setting in my current processor).

To put some numbers on a very hand-waving example, say 40 Hz is played 20 dB louder to sound the same as a 1 kHz tone (see equal loudness curves). Then I am pumping 100x the power into my speakers at 40 Hz to match the volume of a 1 kHz (midrange) sound (all else equal, like speaker sensitivity and such). That means at 40 Hz my mains, if they are 20 dB down at that frequency due to the crossover, are getting the same power as the midrange drivers. That despite a 20 dB/oct crossover if the 40 Hz and 1 kHz tones are played so they sound the same loudness to me. The good news is that much less power is required in the midrange so the additional power is still fairly small; the bad news is that around and below the -3 dB point most speakers start distorting more than at higher frequencies so you don't want any extra power if you can avoid it. And of course that extra energy also makes it more difficult to integrate mains and subs since the sound is not independent through the overlap region when both mains and subs contribute. That is why integration is most critical right at the crossover frequency; above and below, the speakers are being rolled off, so there is less interaction among them.

So the methodology is there, but there are a lot of parameters/variables to consider, to answer what seems like a simple question.

FWIWFM/HTH/etc. - Don

 

[watchnerd]

You can look at loudness curves to estimate where the power lies (integrate the curves) and balance the power loading in your system. I don;t recall the breakpoint off hand but I think it is around 300 Hz or so (half the energy above half below). But that says nothing about what the speaker's drivers can actually do... For many years my rule of thumb has been to use a crossover an octave above the -3 dB point since most loudspeakers distort heavily when large-amplitude signal are applied, and the largest amplitude tend to be in the bass region. sometimes I'll soften to half an octave depending upon the speaker and whether I think distortion or localization is more important. If your main speakers have a -3 dB point of 80 Hz then setting the crossover at 160 Hz is going to play havoc with the stereo image unless the subs are co-located with the mains, and that is not likely to be the best place to smooth the in-room bass response.

That is one reason I've always gone for larger mains -- I want to be able to set the crossover low enough that I cannot localize the sub(s) and thus can place them where they work best for the room. Since crossovers are not brick walls the mains must be able to handle significant power well below (say an octave assuming half-loudness ~10 dB and typical 12 dB/octave crossovers) the crossover frequency (and the subs an equal amount above assuming a symmetric crossover). That is also why I have preferred higher-order crossovers (e.g. the 24 dB/oct L-R crossover I mentioned before, and I am using the 24 dB/oct setting in my current processor).

To put some numbers on a very hand-waving example, say 40 Hz is played 20 dB louder to sound the same as a 1 kHz tone (see equal loudness curves). Then I am pumping 100x the power into my speakers at 40 Hz to match the volume of a 1 kHz (midrange) sound (all else equal, like speaker sensitivity and such). That means at 40 Hz my mains, if they are 20 dB down at that frequency due to the crossover, are getting the same power as the midrange drivers. That despite a 20 dB/oct crossover if the 40 Hz and 1 kHz tones are played so they sound the same loudness to me. The good news is that much less power is required in the midrange so the additional power is still fairly small; the bad news is that around and below the -3 dB point most speakers start distorting more than at higher frequencies so you don't want any extra power if you can avoid it. And of course that extra energy also makes it more difficult to integrate mains and subs since the sound is not independent through the overlap region when both mains and subs contribute. That is why integration is most critical right at the crossover frequency; above and below, the speakers are being rolled off, so there is less interaction among them.

So the methodology is there, but there are a lot of parameters/variables to consider, to answer what seems like a simple question.

FWIWFM/HTH/etc. - Don

My mains have a spec for F3 of 39 Hz, so an octave higher would put the crossover at 78 Hz...right below the 80 Hz milestone.

But...with the deep bass removed, the power handling capabilities of the mains should go up. They're rated for 180W IEC power handling, but that's full range....so presumably it would be higher with the bass removed.

So if high passing the sats lowers distortion, and increases the dynamic range, won't that mean the maximum acoustic output of my sub(s) needs to also go up to match?

Intuitively, this would seem to be true, but absent a full test lab, I'm not sure how to even do the cocktail napkin math on it....

 

[Biblob]

Would I run into trouble if I were to crossover a bassreflex monitor which has the - 3dB at 53 hz, to a sub at 80hz?

 

[watchnerd]

Would I run into trouble if I were to crossover a bassreflex monitor which has the - 3dB at 53 hz, to a sub at 80hz?

Do you mean high pass the monitor at 80 Hz and low pass the sub at 80 Hz?

Or run the monitor full range + low pass the sub at 80 Hz?

 

[Willem]

The manual of my B&W PV1d sub suggests to choose a crossover frequency at the -6dB point of the main speakers and choose a 12dB 2nd order slope for sealed main speakers (and 180 degree phase) and 24 dB 4th order slope (and 0 degree phase) for ported main speakers. This is what I did , and it worked out exactly right, both with my Quad 2805s (with a 4th order slope on the stats to have them least affected by the dynamic sub) and in an experiment with my Harbeth P3ESRs. This is, of course, with the main speakers playing full range without any high pass filtering.

 

[watchnerd]

The manual of my B&W PV1d sub suggests to choose a crossover frequency at the -6dB point of the main speakers and choose a 12dB 2nd order slope for sealed main speakers (and 180 degree phase) and 24 dB 4th order slope (and 0 degree phase) for ported main speakers. This is what I did , and it worked out exactly right, both with my Quad 2805s (with a 4th order slope on the stats to have them least affected by the dynamic sub) and in an experiment with my Harbeth P3ESRs. This is, of course, with the main speakers playing full range without any high pass filtering.

Did you determine that -6 dB as measured in room?

Or just go by the spec sheet?

 

[Willem]

Just by the spec sheet, as was the B&W suggestion. The result was 33 Hz for the Quad 2805s and 45 Hz for the P3ESRs. In short, as per REL theory, only filling in the very lowest frequencies.
In addtion, the sub's manual writes: 'If the specification of your speakers only quotes a –3dB frequency, multiply this by 0.6 for closed-box systems and by 0.7 for vented-box (reflex) systems to get a close enough approximation to the –6dB frequency.'

 

[DonH56]

My mains have a spec for F3 of 39 Hz, so an octave higher would put the crossover at 78 Hz...right below the 80 Hz milestone.

But...with the deep bass removed, the power handling capabilities of the mains should go up. They're rated for 180W IEC power handling, but that's full range....so presumably it would be higher with the bass removed.

So if high passing the sats lowers distortion, and increases the dynamic range, won't that mean the maximum acoustic output of my sub(s) needs to also go up to match?

Intuitively, this would seem to be true, but absent a full test lab, I'm not sure how to even do the cocktail napkin math on it....

Power handling is tricky... Your speakers should be able to play louder with the bass rolled off, true, but I strongly suspect sending 180 W to the tweeter would quickly result in silence and bad smells. You'd have to ask the manufacturer the power limit over frequency, sort of a "safe operating area" curve for your speaker. The amplifier driving them should have to deal with less power so may allow you to play more loudly before it clips. Depending upon the spectrum of the signals (music or whatever). And assuming the speakers can handle it.

The output of your subs needs to be able to provide sufficient SPL for you across the frequency band they can cover. Same as the mains. Bass can require 10~1000x the power to sound as loud as midrange sounds so typically the majority of the power is going to the bass anyway. At 1 kHz, 1 W may be very loud coming from your mains, while it may 100 W at 40 Hz from the sub to sound as loud. Subs typically use higher-power amplifiers and drivers capable of greater output so I prefer to let them handle the load and let my mains handle the rest.

It is easy to overthink this and go deep down the rabbit hole. With your mains I would cross over to the sub at 80 Hz and be done with it unless you are more into tweaking than listening (some are, and I have been, but lately prefer to spend my limited free time just listening).