Increasing distance to the source of sound can, of course, attenuate SPL. e.g. a doubling of the distance might lose 6dB depending on the environment.
However, I've frequently seen this principle applied to subwoofers in small rooms, and I'm not sure this is correct - certainly it doesn't seem to tally with my past experience. I presume this is because at low frequencies, the room's modes start to dominate; the louder areas are more random and may not even be nearer the subwoofers. Even with multiple distributed subwoofers, the peaks and troughs may be less pronounced but the aim is for more evenness throughout the room.
Am I correct in my belief that listening distance in a small room (say in the order of 5m/15feet square) is largely unimportant, or should I still factor in some distance attenuation when choosing what SPL to design for?
Thanks,
Kev
However, I've frequently seen this principle applied to subwoofers in small rooms, and I'm not sure this is correct - certainly it doesn't seem to tally with my past experience. I presume this is because at low frequencies, the room's modes start to dominate; the louder areas are more random and may not even be nearer the subwoofers. Even with multiple distributed subwoofers, the peaks and troughs may be less pronounced but the aim is for more evenness throughout the room.
Am I correct in my belief that listening distance in a small room (say in the order of 5m/15feet square) is largely unimportant, or should I still factor in some distance attenuation when choosing what SPL to design for?
Thanks,
Kev
Yeah, there might be 15db dips and peaks at your listening spot on modal region in a room. To get some sense on the mess, you'd likely want to design flat anechoic frequency response with your system, perhaps some particular house curve for in room sound. You could relax on bass if your listening room and positioning helps it, or perhaps you need more bass than you though if not. Best to think flat system response while planning, and then adjust in situ. If you notice you now had to throw away 10db of bass with EQ not to overwhelm the room, you could build smaller ones next time to sacrifice the excess for better something else.
Bass is usually something that separates impressive system from wimpy one, so as long as you have space for big basses then it's the good side to err on, have bit too much on reserve than too little, assuming you have EQ to adjust it. If you do not have EQ to adjust bass in situ, then perhaps err on lean bass unless you like the boom boom
Bass is usually something that separates impressive system from wimpy one, so as long as you have space for big basses then it's the good side to err on, have bit too much on reserve than too little, assuming you have EQ to adjust it. If you do not have EQ to adjust bass in situ, then perhaps err on lean bass unless you like the boom boom
Thanks very much for your thoughts! I'll probably not factor in the distance very much, then - if there were much effect then probably it'll be small enough that room gain will likely compensate.
There will definitely be EQ available, so more SPL than needed is not technically a problem. But I don't want to over-specify by 'too' much, as the size, weight, cost and power can get out of hand pretty quickly. It is probable that a couple of sealed 12" drivers would suffice for my needs, but if another 10dB were needed (for a 3m/10-foot listening distance) they would not, and things would start to spiral.
I do intend to add one or two more in future; if, as and when budget and space allow. That would mostly be for evenness, as they would see diminishing returns in terms of added SPL, though it would still offer a small margin for future tweaking.
Thanks again,
Kev
There will definitely be EQ available, so more SPL than needed is not technically a problem. But I don't want to over-specify by 'too' much, as the size, weight, cost and power can get out of hand pretty quickly. It is probable that a couple of sealed 12" drivers would suffice for my needs, but if another 10dB were needed (for a 3m/10-foot listening distance) they would not, and things would start to spiral.
I do intend to add one or two more in future; if, as and when budget and space allow. That would mostly be for evenness, as they would see diminishing returns in terms of added SPL, though it would still offer a small margin for future tweaking.
Thanks again,
Kev
Yes it's bit of a shady target. A 12" per side as subs/bass are big enough you likely get smile on your face. I had Eminence LAB12 in very small 30litre closed boxes with few hundred watts of power and DSP and they got windows rattling, certainly enough in that sense. Fine for home party, but more people and it's not enough Then got some 15" and and it's still smile on face, although I rarely listen window rattling loud. I do not comment what the distortion is or anything, likely quite much, but that doesn't take the smile from the face It really makes quite a different feel when there is enough capacity, there is no return to less capable system once you do.
So, if you want to be able to listen average 85db on your seat, then I think it is good and straight forward to design around that for the whole bandwidth + peaks + distance for peak capacity, for example about 105db/1m max output would make 85db average around 2 meters with 15db peaks. If you need the 105db down to say 30Hz, average 15" might be fine and you could put it into a relatively small box if amp power. Since you've got DSP all you need is enough capacity, which you can then form into almost any shape.
Now that is said, 85db is quite loud and not sure how many really listens that loud, so this all depends quite a lot. There are people who have multiple 18" or bigger, in big reflex boxes, so this is not single truth, just some of my experience so far. Also, the room modes really seem to dominate bass. Positioning would be more important than accurate box size, so, if you calculate you need huge system but like it smaller, just go as big as you feel fine and it's gonna be fine. If possible, make cheap prototypes until you find out what is good for you. Have fun
Then got some 15" and and it's still smile on face, although I rarely listen window rattling loud. I do not comment what the distortion is or anything, likely quite much, but that doesn't take the smile from the face It really makes quite a different feel when there is enough capacity, there is no return to less capable system once you do.So, if you want to be able to listen average 85db on your seat, then I think it is good and straight forward to design around that for the whole bandwidth + peaks + distance for peak capacity, for example about 105db/1m max output would make 85db average around 2 meters with 15db peaks. If you need the 105db down to say 30Hz, average 15" might be fine and you could put it into a relatively small box if amp power. Since you've got DSP all you need is enough capacity, which you can then form into almost any shape.
Now that is said, 85db is quite loud and not sure how many really listens that loud, so this all depends quite a lot. There are people who have multiple 18" or bigger, in big reflex boxes, so this is not single truth, just some of my experience so far. Also, the room modes really seem to dominate bass. Positioning would be more important than accurate box size, so, if you calculate you need huge system but like it smaller, just go as big as you feel fine and it's gonna be fine. If possible, make cheap prototypes until you find out what is good for you. Have fun
Thank you once again! By great coincidence, my drivers are very much like the LAB12 that you mentioned; they're from BK-electronics over here, called the FAB12. 12" is quite modest in diameter (by subwoofer standards), but they have low Fs, a decent throw and work in small sealed boxes, which suits my small room nicely.
As I understand it, 'reference level' can (should?) be de-rated for small rooms, as perceptually things sound louder in this situation. The guidance for rooms of my size seems to be for low-frequency peaks of 108db (instead of 115dB in the cinema). But yes that is still very loud; it was IMO (when I last had a decent system) so your 105dB coincides well with what I was envisioning this time. 25-30hz would also be low enough, as I mostly listen to music and don't want to encourage unnecessary infrasonic resonances, so I'd probably roll it off sharply below there.
Actually 105dB could work out nicely. It looks like I can get 102dB at 1m at 25hz from a single Fab12 at Xmax. So a pair should be at least +3db=105db. Though it might be more where it counts, as they'd be within a quarter wavelength of each other at the bottom end of the frequency range, so might couple/sum more efficiently.
However, you can see why this wouldn't be enough if they then lost 10dB due to 3m/10' listening distance, which is why I was keen to check my reasoning. As Art pointed out recently in another thread, we perceive things a bit differently at low frequency; IIRC only 5dB can sound like half/twice as loud down there so -10dB would have sounded like a substantial loss. (EDIT: though would have been offset to some degree by room gain)
Thanks,
Kev
As I understand it, 'reference level' can (should?) be de-rated for small rooms, as perceptually things sound louder in this situation. The guidance for rooms of my size seems to be for low-frequency peaks of 108db (instead of 115dB in the cinema). But yes that is still very loud; it was IMO (when I last had a decent system) so your 105dB coincides well with what I was envisioning this time. 25-30hz would also be low enough, as I mostly listen to music and don't want to encourage unnecessary infrasonic resonances, so I'd probably roll it off sharply below there.
Actually 105dB could work out nicely. It looks like I can get 102dB at 1m at 25hz from a single Fab12 at Xmax. So a pair should be at least +3db=105db. Though it might be more where it counts, as they'd be within a quarter wavelength of each other at the bottom end of the frequency range, so might couple/sum more efficiently.
However, you can see why this wouldn't be enough if they then lost 10dB due to 3m/10' listening distance, which is why I was keen to check my reasoning. As Art pointed out recently in another thread, we perceive things a bit differently at low frequency; IIRC only 5dB can sound like half/twice as loud down there so -10dB would have sounded like a substantial loss. (EDIT: though would have been offset to some degree by room gain)
Thanks,
Kev
Last edited:
Coincidentally, in Josh Ricci's "Room Gain?" tests, a frequency range near your "low enough" suffered losses as much as -30dB compared to the outdoor 2meter reference (blue line starting at -40dB@4Hz:
https://data-bass.com/#/articles/5cb5fb285389a80004c7e58a?_k=ebqbeu
Adding a second subwoofer may result in more than a 6dB gain at the listening position at some frequencies and a loss at others, depending on the room dimensions, sub placement, and listening position.
Since the room response may result in gain at various low frequencies and loss at others, very easy to find response can sound half/twice as loud over a small position change.
Art
Thank you, Art. Gosh that -30dB is huge! In a small room, it certainly puts distance attenuation into perspective. Thank you.
The chart really shows how much things vary with location. Thankfully I shall only need to optimise for one modestly sized listening area, so will try to remain flexible on subwoofer position; to be guided by the acoustics (rather than preconceived ideas of room aesthetics). I'm very tempted to follow the Earl Geddes distributed sub method too; I'm already intending two subwoofers so three wouldn't be a big step up.
GM yes, this is another challenge. I've currently got a suspended timber floor with solid brick plastered walls, but I can move home every few years so there is no knowing what will be encountered. There are many concrete floors around in the UK, or newer constructions can have concrete beam and block floors, and/or plasterboard (drywall) walls. About the only likelihood is that the rooms will remain pretty small, larger rooms tend to be less common over here and out of my budget.
Thanks again,
Kev
Room gain is not the same as room modes. In practice, rooms generally aren't sealed enough to produce room gain, and if there appears to be any it is usually a combination of other things. Therefore it's difficult to make a blanket statement about rooms in general.
Thanks Allen,
Yes I'm treating room gain as a separate factor (to room modes), certainly as having different practical results to consider. Though in all honesty my knowledge on it is lacking, so instead to trying to calculate I've tended to tentitively apply figures gleaned from tables for room size and gain at different frequencies. None of which are for my specific room or even necessarily say much other than size about the room (theoretical or real) that they themselves were derived from.
But perhaps that doesn't matter now. Given the relatively bigger scale of room modes, and the room changing from time to time, I'm tempted to just not plan for any particular room gain; if any occurs in practice then it'll easily be attenuated by EQ. (If my living room ever got as small as (say) a car, then I'd have more pressing problems). So what you said above about insufficient room sealing leads me even more to not bank on getting any/much.
Thanks again,
Kev
Yes I'm treating room gain as a separate factor (to room modes), certainly as having different practical results to consider. Though in all honesty my knowledge on it is lacking, so instead to trying to calculate I've tended to tentitively apply figures gleaned from tables for room size and gain at different frequencies. None of which are for my specific room or even necessarily say much other than size about the room (theoretical or real) that they themselves were derived from.
But perhaps that doesn't matter now. Given the relatively bigger scale of room modes, and the room changing from time to time, I'm tempted to just not plan for any particular room gain; if any occurs in practice then it'll easily be attenuated by EQ. (If my living room ever got as small as (say) a car, then I'd have more pressing problems). So what you said above about insufficient room sealing leads me even more to not bank on getting any/much.
Thanks again,
Kev
Last edited:
Well I had intended dual-opposed (push-push) drivers for my subwoofers, but now (for similar cost/size/weight) I'm convinced that splitting them into two or probably three separate distributed subs would be better use of resources. I suppose there's nothing preventing distributed subs also being dual-opposed but the cost and size would almost double.
One criticism/concern that people make is being able to localise the subs as being in different locations to the main speakers. Though IMO, certainly with my ears in a small room and low-ish crossover, this this only occurs due to the subs making noises they technically shouldn't - like higher frequency resonances or other audible byproducts. I've had drivers that were better than others (and have made boxes that were better than others); some were not localisable at all and some just needed a steeper or lower crossover but some were never free of problems.
That said, Earl Geddes apparently goes a step further and uses bandpass boxes to acoustically filter out-of-band noise. Not sure if the LAB12/FAB12 is good in a small bandpass, though I can get the curve to look OK in winisd. It needs some combo of a bigger box and/or more power and/or hefty EQ though, none of which are entirely ideal when talking multiples of them. Maybe I'll have to build a sealed test box to see if it is warrented.
One criticism/concern that people make is being able to localise the subs as being in different locations to the main speakers. Though IMO, certainly with my ears in a small room and low-ish crossover, this this only occurs due to the subs making noises they technically shouldn't - like higher frequency resonances or other audible byproducts. I've had drivers that were better than others (and have made boxes that were better than others); some were not localisable at all and some just needed a steeper or lower crossover but some were never free of problems.
That said, Earl Geddes apparently goes a step further and uses bandpass boxes to acoustically filter out-of-band noise. Not sure if the LAB12/FAB12 is good in a small bandpass, though I can get the curve to look OK in winisd. It needs some combo of a bigger box and/or more power and/or hefty EQ though, none of which are entirely ideal when talking multiples of them. Maybe I'll have to build a sealed test box to see if it is warrented.
Take care.. look at the plot shown in #6. The lower region is approximately 18dB/oct, whereas room gain is theoretically 12. What you might have there is a helmholz resonance, which would drop back to nothing towards zero Hz because it isn't sealed.. as explained here..
Very useful, thanks again Allen. And thank you for putting it so kindly, too. I've heard room gain stated as a fact since quite early days in my audio hobby (probably from around when the internet started to host forums and websites). So it is upsetting to realise that I've been naive about the effect for this long; I'll no doubt have perpetuated it several times over the years myself..
Though it is also quite easy to accept, since I've rarely been able to measure (in my own rooms) the amount of room gain that was supposedly expected. To be fair there often was less attenuation due to distance than expected, so an apparent net gain. But I now realise this would simply have been unevenness from room modes.
Getting back to room modes then...
Looking at Art's chart, there is (to my eye) a trend that follows the outdoor reference over my intended frequency range (and above), it really only loses correlation towards infrasonic frequencies, outside of my range. The correlation is quite sloppy because the modes are so big that they clearly dominate, but smoothing from multiple subs would presumably also bring things somewhat closer to the trend. Whether distance-related attenuation would then be (relatively) more apparent I don't know, but I'd guess perhaps not by 'very' much because of the distributed nature of the subwoofers. There would also be one or two more of them to somewhat increase available SPL, and one or more might also be placed closer to the listening position.
So I think this means I'm still content with the SPL available from my chosen drivers, albeit for slightly different reasons. I'm also happier with their (12") size; they look good but seemed quite modestly sized (for subwoofers), but now I'm envisioning multiples, any larger would seem excessive for my requirements.
Cheers,
Kev
Though it is also quite easy to accept, since I've rarely been able to measure (in my own rooms) the amount of room gain that was supposedly expected. To be fair there often was less attenuation due to distance than expected, so an apparent net gain. But I now realise this would simply have been unevenness from room modes.
Getting back to room modes then...
Looking at Art's chart, there is (to my eye) a trend that follows the outdoor reference over my intended frequency range (and above), it really only loses correlation towards infrasonic frequencies, outside of my range. The correlation is quite sloppy because the modes are so big that they clearly dominate, but smoothing from multiple subs would presumably also bring things somewhat closer to the trend. Whether distance-related attenuation would then be (relatively) more apparent I don't know, but I'd guess perhaps not by 'very' much because of the distributed nature of the subwoofers. There would also be one or two more of them to somewhat increase available SPL, and one or more might also be placed closer to the listening position.
So I think this means I'm still content with the SPL available from my chosen drivers, albeit for slightly different reasons. I'm also happier with their (12") size; they look good but seemed quite modestly sized (for subwoofers), but now I'm envisioning multiples, any larger would seem excessive for my requirements.
Cheers,
Kev
It occurred to me there 'might' be another cause of gain at low frequencies.. If dual/multiple subwoofers are placed close enough to acoustically couple they might gain (up to) 6dB instead of 3dB. I believe they couple best/most if placed within 1/4-wavelength, and then less as the distance increases. So at 28hz any pairings placed within about 3m of each other will sum very strongly, but by even 80hz (let alone higher) any pairings spaced over about 1m will achieve less gain. In small rooms, I'd guess quite a lot of subwoofers would be close to 3m separation, or at least much closer to that than 1m.
Also... another implication on this might be my intention for using distributed subwoofers to smooth room modes. If one can't keep the subs further (or even much further) apart than 1/4 wavelength... in a real acoustical sense are they actually separate or distributed at all?
Also... another implication on this might be my intention for using distributed subwoofers to smooth room modes. If one can't keep the subs further (or even much further) apart than 1/4 wavelength... in a real acoustical sense are they actually separate or distributed at all?
After some thought, this is exactly what I've decided to do. So, then any extra LF gain, from whatever source, will just be a bonus - not a problem if it doesn't happen. The room modes will still dominate, but things can be optimised for quite a small listening position; I don't need to cover many seats evenly (unlike in some home theater type arrangements). And the intention is for multiple/distributed subs, so hopefully I could counter dips by placement, rather than needing lots of extra SPL.
I had considered going for THX levels, but even if down-rated for a small room it is louder than I'd personally want or need, and perhaps more useful for sound effects in films/movies. It is also getting quite big and costly, especially if one doesn't factor in getting help from things like alleged room gain or convenient spikes in room modes.
For music, an average of 85dB with +15dB headroom for peaks (within linear cone excursion) would indeed be fine for me; a maximum of 100dB at the listening position. If that is at 3m then at worst for distance attenuation I'd only need 110dB at 1m. Even sealed, my 12" drivers will each do 106dB at 30hz within Xmax, so (even if not closely coupled) two would achieve 109dB and three would reach 111dB. In terms of SPL adding more subs gains increasing less, though I might eventually go to four if it aids placement and room modes (I'm also likely to use 2-channel amplifiers, so two or four subs would be more fitting than three).
Cheers,
Kev
