Getting subwoofers for the first time

[h.g.]

I recently spent a cuple of months with two subs for my Kef R100's. There was bass, no problem, and one could go much louder, no problem. I have measurement capabillities and aren't afraid of using an EQ (digital), lots of tweaking was done. But it never sounded right. Even if measurements didn't say so, it was always boooomy and uncoherent. I gave up.

Sub's where the bass sections of my JBL 4331B's (15" and very large), used full range they have stunning, fast and full bass.

Conclusion (not for the first time): If you want bass, use full range speakers. But feel free to experiment!

I suspect you may be lumping together a number of factors rather than separating them.

I too have some KEF coaxials but, like KEF themselves in their higher quality models, wouldn't consider using a 5" driver (equivalent to a 4" due to lost area for the tweeter) below about 400 Hz because not only does it have insufficient area for clean transients at standard levels the large cone deflections will lead to raised levels of distortion and more importantly modulate the coaxial tweeter output. High quality in the 80-400 Hz range requires adequate cone area in the main speakers. At a distance of 3-4 m in a room at standard levels this means the clean reproduction of around 112-115 dB transients (at 1m) and 4" drivers tend to be a good 10 dB shy of this. Close at a desk they may be OK(ish) but not as mains in a room. It requires a 10-12" driver or more commonly these days 2 x 7-8". This is what allows low frequency percussive sounds like drums to sound reasonably correct. Not "dynamic", "fast bass" or whatever but reasonably correct. It has nothing to do with subwoofers or low frequency extensions of the mains just cone area.

With a pair of subs and a pair of R100/LS50s the best approach isn't clear to me. Sitting the subs under the "tops" and crossing at 200-400 Hz to create mains with adequate cone area and then living with whatever careful EQ can do to improve the room response or crossing at 80-100 Hz and distributing the subs to improve the quality of the room response below 80 Hz and then living with the consequences of inadequate cone area in the mains. If the subs were capable of crossing high I suspect I would lean towards the former due to preferring to listen at standard levels and favouring clean percussion but would be a bit fidgety rather than happy and content. For those that tend to listen at quieter levels and with a different set of preferences it is easy to see a different setup being preferred. If the subs are individually DSP controlled and setup competently a preference for no subs makes little sense to me given they could be arranged to be effectively inaudible apart from absorbing some of the room boom. Not competently setup and emphasising the room boom then off may well be preferred.

 

[Robert]

I recently spent a cuple of months with two subs for my Kef R100's. There was bass, no problem, and one could go much louder, no problem. I have measurement capabillities and aren't afraid of using an EQ (digital), lots of tweaking was done. But it never sounded right. Even if measurements didn't say so, it was always boooomy and uncoherent. I gave up.

Sub's where the bass sections of my JBL 4331B's (15" and very large), used full range they have stunning, fast and full bass.

Conclusion (not for the first time): If you want bass, use full range speakers. But feel free to experiment!

I would agree with the conclusion in many instances. However the use of multi-sub is to create multiple LF sources in order to tackle room modes.
In smaller rooms these cluster in the 35-60Hz region and can really mess with bass evenness and perceived timing. Larger rooms, which on the whole tend to house larger, full range speaker systems place the modes <30Hz where there is far less signal on most recordings.
I likely need a third to do this properly in my room but using two is definitely very beneficial.

As for timing issues, I'm not convinced that <100Hz it matters a great deal where the subs are positioned, within reason.
Live music is often amplified through systems with bass stacks position well away from the primary sources but don't seem to suffer from a timing POV.

@Robert, when you EQ'd the frequency response to achieve +/- 3dB below 80Hz I presume you applied minimal smoothing to the measurements?

Yes for the initial setup. Though once things start to take shape I switch to 1/6 smoothing which far better represents what we hear in room. In fact the sub plots above are at 1/12dB Oct.
I've just installed REW which seems to be the commonly used software rather than Arta. I'll see how that goes.

 

[Robert]

I suspect you may be lumping together a number of factors rather than separating them.

I too have some KEF coaxials but, like KEF themselves in their higher quality models, wouldn't consider using a 5" driver (equivalent to a 4" due to lost area for the tweeter) below about 400 Hz because not only does it have insufficient area for clean transients at standard levels the large cone deflections will lead to raised levels of distortion and more importantly modulate the coaxial tweeter output. High quality in the 80-400 Hz range requires adequate cone area in the main speakers. At a distance of 3-4 m in a room at standard levels this means the clean reproduction of around 112-115 dB transients (at 1m) and 4" drivers tend to be a good 10 dB shy of this. Close at a desk they may be OK(ish) but not as mains in a room. It requires a 10-12" driver or more commonly these days 2 x 7-8". This is what allows low frequency percussive sounds like drums to sound reasonably correct. Not "dynamic", "fast bass" or whatever but reasonably correct. It has nothing to do with subwoofers or low frequency extensions of the mains just cone area.

With a pair of subs and a pair of R100/LS50s the best approach isn't clear to me. Sitting the subs under the "tops" and crossing at 200-400 Hz to create mains with adequate cone area and then living with whatever careful EQ can do to improve the room response or crossing at 80-100 Hz and distributing the subs to improve the quality of the room response below 80 Hz and then living with the consequences of inadequate cone area in the mains. If the subs were capable of crossing high I suspect I would lean towards the former due to preferring to listen at standard levels and favouring clean percussion but would be a bit fidgety rather than happy and content. For those that tend to listen at quieter levels and with a different set of preferences it is easy to see a different setup being preferred. If the subs are individually DSP controlled and setup competently a preference for no subs makes little sense to me given they could be arranged to be effectively inaudible apart from absorbing some of the room boom. Not competently setup and emphasising the room boom then off may well be preferred.

I still have a pair of R3s here which combine the 5" coaxial with a 6'5" bass unit. I found them a bit bright compared to the LS50 (original, not Meta) but I'll swap them in and see how they perform.
I measured them again recently and discovered that the 'shadow flare' which sits around the coaxial unit need to be pressed well into the mount (it can work loose) - if not you get a 2-3 dB presence dip which emphasises the tweeter.

I listen neafield though in a small room so don't push the LS50 particularly hard even full range.

 

[ToTo Man]

Yes for the initial setup. Though once things start to take shape I switch to 1/6 smoothing which far better represents what we hear in room.
I've just installed REW which seems to be the commonly used software rather than Arta. I'll see how that goes.

Don't be surprised if the smoothing in REW looks different to the smoothing in ARTA. I only have experience of REW but I remember someone on the forum recently posted a graph of their in-room FR with 1/24 smoothing and it looked somewhere in between the 1/12 and 1/6 I'm used to seeing in REW.

PS - I'm not convinced 1/6 does better represent what we hear in-room, especially in the bass. IME even 1/12 smoothing can mask the severity of low frequency nulls too much if they have narrow Q. For example, there's a certain distance from my rear wall where I have a deep but narrow null at 77Hz, which essentially eliminates the D# from the bass guitarist's fretboard. If I viewed my in-room FR at this listening position with 1/6 smoothing, the null would (wrongly) appear much less of an issue than it is in reality.

 

[Tony L]

Live music is often amplified through systems with bass stacks position well away from the primary sources but don't seem to suffer from a timing POV.

Most live pop/rock via PA systems sounds absolutely awful to my ears. Enjoyable from an event/social perspective, but just shockingly bad from a hi-fi/audio viewpoint IMHO. I far prefer seeing bands at the very start of their careers with unamplified drums and simple backline. I’ve always viewed hi-fi as a means to recreate the studio sound. The very last thing I want in my house is the sound of a rock band playing Wembley, Glastonbury or whatever!

PS Top sub evaluation tip: don’t focus on the bass at all! Listen to the midband, the emotional connection, and especially soundstage depth which is often a very easy ‘tell’ something is not right (also factor that having the undriven subs in the room has a detrimental effect, so if you prefer it with the subs off there is still another fairly substantial gain to be had). I remember going through this with someone who had just bought a very fancy sub with correction etc and even when we turned it down to the point it didn’t annoy me (pretty much off) it was easy to tell if it was on/off as the soundstage just collapsed when on. This using good jazz recordings that don’t have hugely deep bass at all.

 

[Robert]

Don't be surprised if the smoothing in REW looks different to the smoothing in ARTA. I only have experience of REW but I remember someone on the forum recently posted a graph of their in-room FR with 1/24 smoothing and it looked somewhere in between the 1/12 and 1/6 I'm used to seeing in REW.

PS - I'm not convinced 1/6 does better represent what we hear in-room, especially in the bass. IME even 1/12 smoothing can mask the severity of low frequency nulls too much if they have narrow Q. For example, there's a certain distance from my rear wall where I have a deep but narrow null at 77Hz, which essentially eliminates the D# from the bass guitarist's fretboard. If I viewed my in-room FR at this listening position with 1/6 smoothing, the null would (wrongly) appear much less of an issue than it is in reality.

Good advice - I've certainly found REW and ARTA to give differing results in the past.
I've always preferred the latter as it runs continuous pink noise and you can walk the room in real time and observe changes.
REW fires out a sweep of course but I've no ideas if the different stimulus methods should impact results, smoothing aside.

 

[Tony L]

I think you can get REW to do the pink noise RTA thing, but I can never remember how!

 

[Robert]

Most live pop/rock via PA systems sounds absolutely awful to my ears. Enjoyable from an event/social perspective, but just shockingly bad from a hi-fi/audio viewpoint IMHO. I far prefer seeing bands at the very start of their careers with unamplified drums and simple backline. I’ve always viewed hi-fi as a means to recreate the studio sound. The very last thing I want in my house is the sound of a rock band playing Wembley, Glastonbury or whatever!

PS Top sub evaluation tip: don’t focus on the bass at all! Listen to the midband, the emotional connection, and especially soundstage depth which is often a very easy ‘tell’ something is not right (also factor that having the undriven subs in the room has a detrimental effect, so if you prefer it with the subs off there is still another fairly substantial gain to be had). I remember going through this with someone who had just bought a very fancy sub with correction etc and even when we turned it down to the point it didn’t annoy me (pretty much off) it was easy to tell if it was on/off as the soundstage just collapsed when on. This using good jazz recordings that don’t have hugely deep bass at all.

Not hearing any change to the mids. I am crossing at 70Hz 24/dB Oct slope though.

When discussing the impact of subs we are of course dealing with a moving target and something very room/main specific.
With LS50s or any small 2 way, relieving the 'speaker of large excursion bass duties can also improve the mids by reducing intermod, especially with a small coaxial as h.g. notes above.
Swings and roundabouts as usually the case with hi-fi!

it will be interesting to see what happens when with the LS50s electrically high passed.

 

[ToTo Man]

Good advice - I've certainly found REW and ARTA to give differing results in the past.
I've always preferred the latter as it runs continuous pink noise and you can walk the room in real time and observe changes.
REW fires out a sweep of course but I've no ideas if the different stimulus methods should impact results, smoothing aside.

I think you can get REW to do the pink noise RTA thing, but I can never remember how!

Yes, REW has a Generator that can play a plethora of continuous tones including pink noise, and an RTA that allows you to observe the output in real time. I personally prefer to play Periodic Pink Noise as I find the RTA readings are more stable and don't 'dance around' as much as with standard pink noise (I don't know the technical reason for this).

 

[Robert]

Yes, REW has a Generator that can play a plethora of continuous tones including pink noise, and an RTA that allows you to observe the output in real time. I personally prefer to play Periodic Noise as I find the RTA readings are more stable and don't 'dance around' as much as with standard pink noise. I don't know the technical reason for this.

That's good to know as I've spent little time with REW.
ARTA is powerful and I find it good for full range measurement but I've also noticed how it jumps around especially at LF - a pain if trying to measure subs!

 

[ToTo Man]

That's good to know as I've spend little time with REW.
ARTA is powerful and I find it good for full range measurement but I've also noticed how it jumps around especially at LF - a pain of trying measured subs!

Another tip I just remembered, - be sure that the RTA in REW is set to 'RTA Mode' and not ''Spectrum Mode', as the latter is calibrated for a White Noise signal.

 

[Tony L]

I am crossing at 70Hz 24/dB Oct slope though.

I’d actually try lower than that. The LS50s are a remarkably capable speaker and I know your listening levels won’t be getting close to their limits. There may be an advantage to removing the real deep flappy stuff on electronica etc, so I’d be tempted to go for 40 Hz or so, but you may find the HPF is more detrimental to the sound than any logical benefit it may bring (for others reading Rob is a friend, we know each others rooms!). As you say all scenarios are different, but I’m factoring your room and listening levels here. My view is the LS50 is capable of far higher clean levels as-is than you will ever get close to. I’m the same, I’m not stressing even the far less volume-capable LS3/5A or JR149 the way I use them.

 

[sonddek]

Well.... That's Robert converted then. Just Tony L to go. Join us on the dark side Tony....... the bass awaits you. Consume you it will.

At this point I should admit my dirty little secret: I bought a slightly iffy used Linn Akurate 221 for ~£400 a few months ago to try alongside my active 242iis as an experiment. My 12'x20' room has a terrible trampoline creaky floor, very large double-glazed windows, and noisy passing lorry-traffic, so I was very skeptical that things would improve. I put the sub between my speakers, set its roll-off at the recommended frequency and the volume a bit lower than default, and have left it like that for months. I'm enjoying hearing a few deeper notes, slightly more full-fleshed timbre, and no particularly discernible down side in coherence or timing.

No DSP, no adjusting main speaker eq or output, no crawling around on the floor, no trouble.

In fact, I like it enough that I will try to pick up a second sub to run stereo. Maybe I've just been lucky, or maybe I'm deaf.

 

[Robert]

I’d actually try lower than that. The LS50s are a remarkably capable speaker and I know your listening levels won’t be getting close to their limits. There may be an advantage to removing the real deep flappy stuff on electronica etc, so I’d be tempted to go for 40 Hz or so, but you may find the HPF is more detrimental to the sound than any logical benefit it may bring. As you say all scenarios are different, but I’m factoring your room and listening levels here. My view is the LS50 is capable of far higher clean levels as-is than you will ever get close to. I’m the same, I’m not stressing even the far less volume-capable LS3/5A or JR149 the way I use them.

True, they do play loud and clean for a small box.

40 Hz would make sense if simply wanting to augment the LS50 extension, but I'm also trying to actively suppress room modes which brings its own benefits for timing.

This is only day 2 of real testing - weeks to go yet

 

[Alastair]

I tried duel NHT subs with my LS50's (and a single Sunfire sub), both with and without an active crossover, they had more dynamic head room with the subs. Ultimately i thought the LS50 just sounded more satisfying on their own.
Glad you like what they do Rob and you've got them to work well .

I think Rodney one of the South African members here, had a similar set up for a while then moved onto Dickie designed Giya's, i'm not sure if he's about anymore?

Sadly Rodney died a couple of years back now. He certainly experimented a huge amount with subs and dsp options, I recall with Meridian DSP6ks before switching to the Giya’s. Top chap.

 

[ToTo Man]

Sadly Rodney died a couple of years back now. He certainly experimented a huge amount with subs and dsp options, I recall with Meridian DSP6ks before switching to the Giya’s. Top chap.

I didn't know this. I always loved seeing photos of his listening room, it was something to be proud of. Belated RIP.

 

[James]

The last octave of bass doesn’t matter to me that much to be honest, I’d certainly not compromise coherence, timing or soundstage depth/resolution to achieve it. The thing I like about proper big speakers (15” Tannoys, huge JBL monitors, Maggies, Khorns etc) is that real ease, effortlessness, scale and heft, and adding a sub to a small speaker never achieves that to my ears. It is a different thing, and one so far I don’t feel any real need for.

A strict definition of full range is 20Hz to 20kHz flat (in-room). Very few so-called full range loudspeakers can wear that tag with integrity. That being said, I agree with you that the last octave of bass is not all that important unless you happen to be a pipe organ fan.

There is also a distinct difference between bass extension and gravitas. It is entirely possible, within the realms of power and excursion limits, to EQ a small loudspeaker to dig deeper. But a much larger loudspeaker with the exact same extension (F3) will sound altogether more visceral. That is why you and I like big loudspeakers, and little loudspeakers that don't try too hard in the LF department have their own charms.

 

[tuga]

As @h.g. said the LS50 is a small box with a tiny midwoofer; measured THD (96dB at 1m here) shoots up below 200Hz and IMD should follow suit.
Even if listening distance is 5-6ft I would still rather cross above 100Hz and position the subs under high-passed mains, but this will require the use of something like a miniDSP or a HT pre with crossover capabilities (and defeat the purpose of using sub positioning flexibility to achieve a flatter low/ and sub-bass response at the listening spot.
Personally I’d rather go for a pair of 3-ways like the R11s but the OP setup may sound nice too considering the listening distance and max levels involved.

 

[Discopants]

That last octave has the harmonics of many more instruments than pipe organ lower registers.

If you get the sub(s) well integrated you get a lot of spatial info from the recording in this octave too.

 

[Robert]

As @h.g. said the LS50 is a small box with a tiny midwoofer; measured THD (96dB at 1m here) shoots up below 200Hz and IMD should follow suit.
Even if listening distance is 5-6ft I would still rather cross above 100Hz and position the subs under high-passed mains, but this will require the use of something like a miniDSP or a HT pre with crossover capabilities (and defeat the purpose of using sub positioning flexibility to achieve a flatter low/ and sub-bass response at the listening spot.
Personally I’d rather go for a pair of 3-ways like the R11s but the OP setup may sound nice too considering the listening distance and max levels involved.

I'm listening at about 1.5m and the LS50s are up on 28" stands. More like the situation you'd see on a mixing console than a typical listening room.
I can play the LS50 way louder than I find comfortable without obvious driver distress.
Remember that we have far more tolerance to LF distortion than at mid frequencies.

This is another example where listening in a defined setting is essential as the distortion graphs alone don't tell anything like the full story.