I disagree. Pretty much any competently built, low-pass filtered sub made from parts of reasonable quality* will produce close enough to "pure bass tones" for this purpose. Especially when there's HF content above.
*For all this to hold, the drive-units don't need to be at the Aurasound level of price/quality. There are plenty of inexpensive drive-units that are low enough in distortion and free of mechanical noise. The Peerless SLS jobs, for example.
OK. Show me a sub with crossover that has NO audible content above whatever frequency you contend is omni-directional. If it has none, then maybe it isn't possible to locate. Until then, it remains fairy dust.
Pallas. I own some Peerless SLS. They are indeed clean, but that's not the whole story.
Pallas. I own some Peerless SLS. They are indeed clean, but that's not the whole story.
You chose to participate in a thread on theory- as many threads are here. There are also build threads, and they're frequently labeled as such, and threads with a bit of each. I'm mostly a sawdust guy myself- but theory necessarily informs the solder slinging, or else our projects will never be anything but clones of products by those who know theory.
If there's any place for the theory, it's in this, a fantastic conglomerate of multinational nerds, hobbyists, craftsmen, and professionals. That doesn't preclude the practical side- I'll happily have my cake and eat it too.
I've advocated nearfield subwoofing (and multisub where it works) for some of my clients and friends for quite some time, for example. I never would have tried either without having a rudimentary understanding of modal behavior in rooms.
There's a fellow who starts a lot of threads in these pages who does a huge amount of prototyping but never seems to get any closer to his goal. The gap is theory.
Oh and- relatively clean LF isn't so far from reality as to make this navel-gazing.
Why get over it? If theory doesn't get you good results you should beware of it. If people want to believe that bass is omni-directional and therefore you can simply do whatever you want below ~100 Hz, that's sad. Because it's not true. I've not seen it proved on paper or in person.
And FWIW - I was banned from the Flat Earth Society. Really. So no need to go there.
And FWIW - I was banned from the Flat Earth Society. Really. So no need to go there.
Part of the rest of the story. In real applications whatever bass localization cues (harmonics etc) might be coming from the subwoofer will be masked by those same cues coming from the mains at higher level . . . and even in live performance it is sometimes difficult to localize the string bass because of masking from the violas and, perhaps, bassoons. Not to mention the "ambiguating" effect of hall reflection/reverberation.
In any case in small rooms subwoofer performance is dominated by the room itself and (far too commonly) by phasing errors around crossover (one of the potential problems with subs that are not located near the mains). Multiple (distributed) subs can have that problem big time, especially if the crossover is too high and you end up with summation nulls as bad as the room modes you're trying to correct.
The whole discussion is near to pointless without specifying the room, the mix, and the "kind" of bass in question . . . dealing with movie thuds and bumps is quite different from reproducing music.
OK - just had a glimpse at that post. I guess if your DBA really works well and gives a good improvement of precision in bass reproduction it would be worth the probably small sacrifice in localisation from the mono summing. As already mentioned the localisation of bass instruments is mostly done via their overtones and not their fundamentals.
I can only talk about my experiences with a small PA system that I once built and that I also used to listen to music quite a few times.
It is consistung of two satellites with a 10" mid-woofer and a HF horn. These are completed with a pair of bandpass boxes covering the range from the 40ies up to about 100 Hz. In Hifi terms these would be called woofers and in PA terms they would often be called subwoofers. When I built the prototype for the woofer modules and did some casual listening (i.e. not a scientific test at all) I had the feeling that the bass sounded "disconnected" somehow and that localisation of bass is drawn to the side where I used the woofer module as stand for the satellite speaker. The fact that the transient response of a bandpass isn't the best and the unavaidable resonant peaks in the midrange would have been exagerating this effect. I decided to build two of these modules because I expeced to get the localisation better and also to profit form the 6 dB increase of headroom.
Even tough both subs were then still fed the same mono bass signal the whole bass reproduction gained precision when the two bass modules were put under each one of the satellites. And with bass signals (in terms of bass instruments containing harmonics - not just low pure frequency tones) I had the feeling that localisation would be definitley better now. I.e I had better "sub" integration now in terms of temporal and localisation behaviour. I guess this is due to several factors like better masking of midrange frequencies coming form the subs (which would be lower for a high-quality subwoofer system), better integration regarding path lengths and improved excitation of room modes.
While it is difficult to draw conclusiuons from theses experiences I would say that your planned DBA wouldn't bee detrimental to localisation and gain a lot of precision elsewhere.
Hope this post was helpful to some degree.
Regards
Charles
Agreed- and then some- even quick and dirty multisub works much better than single in every instance I've tried, so long as you do at least basic effort towards getting them aligned.
Well here's the thing.
It's all a weak approximation to begin with.
As far as I can determine, Dr. Geddes' approach is to use multiple subs in order to minimize the effects of room modes. In effect creating an average in terms of amplitude ("flat response") in the room. (I apologize if this is not an accurate summary of the method, in which case please correct)
Others prefer the "corner" placement or the "wall floor" boundary.
These methods exploit the change in "pi space" which effects the LF response by lifting the lower freqs. Of course the results will differ from room too room and in position in a given room rather significantly.
Some prefer what we can call a "time aligned" (actually this is trademarked name) approach. In this situation the acoustic centers of the subs + whatever is running higher in frequency is aligned either by physical setbacks or electronic delays, or both.
The latter two methods can be used mono, multiple mono or 2 channel stereo - the subs we're talking about.
Personally I am a strong advocate of the last method. I find it produces the best harmony in terms of the spatial presentation. Admittedly the results of doing this properly are akin to getting the ripple out of a gloss (piano) finish or perhaps a car paint job. It makes a difference when you are paying close attention and/or you can compare to *not* having done this. Maybe it is more like having a projector lens really sharpened up, or the difference between an average and a really good lens on a camera.
One can rely upon the higher frequency cues to "localize" bass, and in fact this is what *will* happen under normal listening conditions. But when the subs are not "time aligned" AND the rest of the system *is* the absolute "focus" in the ability to hear and resolve subtle spatial aspects are *reduced* in my experience.
Having said all this, I think that most speaker systems are not set up or designed with the "time alignment" idea at all. Frequency response is usually the primary consideration.
It seems to me that a fast transient, like a drum hit, ideally would be reproduced, looking like the original (as picked up by a single optimally placed microphone - for example) when viewed on a scope or other test gear. When looking at the impulse response using a real(ish) time FFT box and a decent mic the results ought to show the drivers output adding to make the highest peak. As you physically move the drivers WRT each other, or alter the delays electronically one can see the peak drop and the contributions of each driver become discrete and separate. This shouldn't be news to most.
Regardless, when the mic is moved for multiple driver speakers, the results change. But even so, this is a useful process and seems cause a general improvement in spatial imagine and tonality when followed, as compared to merely flattening frequency response.
In addition, this is at least part of the reason that many prefer so-called "full range" speakers, including but not limited to ESLs, big ribbons, large wide-range horns and things like the smaller "full range" drivers. The reason being that they inherently come with the "time alignment" built in to a great extent.
So, this is my best case for keeping the subs in close and proper relationship to the main speakers. The separate "localization" of subs by themselves turns out to be something of a non-issue in my view.
It's all a weak approximation to begin with.
As far as I can determine, Dr. Geddes' approach is to use multiple subs in order to minimize the effects of room modes. In effect creating an average in terms of amplitude ("flat response") in the room. (I apologize if this is not an accurate summary of the method, in which case please correct)
Others prefer the "corner" placement or the "wall floor" boundary.
These methods exploit the change in "pi space" which effects the LF response by lifting the lower freqs. Of course the results will differ from room too room and in position in a given room rather significantly.
Some prefer what we can call a "time aligned" (actually this is trademarked name) approach. In this situation the acoustic centers of the subs + whatever is running higher in frequency is aligned either by physical setbacks or electronic delays, or both.
The latter two methods can be used mono, multiple mono or 2 channel stereo - the subs we're talking about.
Personally I am a strong advocate of the last method. I find it produces the best harmony in terms of the spatial presentation. Admittedly the results of doing this properly are akin to getting the ripple out of a gloss (piano) finish or perhaps a car paint job. It makes a difference when you are paying close attention and/or you can compare to *not* having done this. Maybe it is more like having a projector lens really sharpened up, or the difference between an average and a really good lens on a camera.
One can rely upon the higher frequency cues to "localize" bass, and in fact this is what *will* happen under normal listening conditions. But when the subs are not "time aligned" AND the rest of the system *is* the absolute "focus" in the ability to hear and resolve subtle spatial aspects are *reduced* in my experience.
Having said all this, I think that most speaker systems are not set up or designed with the "time alignment" idea at all. Frequency response is usually the primary consideration.
It seems to me that a fast transient, like a drum hit, ideally would be reproduced, looking like the original (as picked up by a single optimally placed microphone - for example) when viewed on a scope or other test gear. When looking at the impulse response using a real(ish) time FFT box and a decent mic the results ought to show the drivers output adding to make the highest peak. As you physically move the drivers WRT each other, or alter the delays electronically one can see the peak drop and the contributions of each driver become discrete and separate. This shouldn't be news to most.
Regardless, when the mic is moved for multiple driver speakers, the results change. But even so, this is a useful process and seems cause a general improvement in spatial imagine and tonality when followed, as compared to merely flattening frequency response.
In addition, this is at least part of the reason that many prefer so-called "full range" speakers, including but not limited to ESLs, big ribbons, large wide-range horns and things like the smaller "full range" drivers. The reason being that they inherently come with the "time alignment" built in to a great extent.
So, this is my best case for keeping the subs in close and proper relationship to the main speakers. The separate "localization" of subs by themselves turns out to be something of a non-issue in my view.
Bear
I don't see how your description makes sense in the modal region of a room - i.e. VLFs. Everything that you are saying applies only to frequencies where one can actually find a "fast transient". Can you measure or determine such a thing looking only at frequencies below 100 Hz? I don't see how. What you will see is a myriad of reflections etcetera, within a single cycle of the sound. It simply does not make sense to talk about transients and impulse responses in the modal region because neither your instrumentation or your ears can sort out what is initial sound from what is a reflection. This is, of course, easy to do above say a few hundred Hz, but not below 100 Hz. Below 100 Hz in a small room and all one can talk about is steady state - that's what we hear and that's what we can measure.
Above a few hundred Hz and I can accept what you are saying. Time alignment yields a compact impulse response, something that I strive to achieve, but below 100 Hz is an entirely different story.
I don't see how your description makes sense in the modal region of a room - i.e. VLFs. Everything that you are saying applies only to frequencies where one can actually find a "fast transient". Can you measure or determine such a thing looking only at frequencies below 100 Hz? I don't see how. What you will see is a myriad of reflections etcetera, within a single cycle of the sound. It simply does not make sense to talk about transients and impulse responses in the modal region because neither your instrumentation or your ears can sort out what is initial sound from what is a reflection. This is, of course, easy to do above say a few hundred Hz, but not below 100 Hz. Below 100 Hz in a small room and all one can talk about is steady state - that's what we hear and that's what we can measure.
Above a few hundred Hz and I can accept what you are saying. Time alignment yields a compact impulse response, something that I strive to achieve, but below 100 Hz is an entirely different story.
There's "the room" again (and what "small" is). Mine's 24 feet long (25 counting the bay window on the back wall), which makes the longitudinal mode almost a non-issue (and not as bad for the neighbors as it might seem, as there are no windows on their facing wall). The ceiling is lathe and plaster on 24 foot 2x4 joists (those were the days) with a floating sister in the middle to reduce sag, so the bass is almost as loud in the attic as it is in the listening room. The floor is wood on joist (granted shorter spans) and pretty transparent into the basement. Side walls have windows on one side and a fireplace and a large opening to the dining room on the other. Getting the picture? Not much in the way of "mode" problems, and those easy to fix.
So I'm with Bear on this . . . "time aligning" the subs (a line of them on the front wall behind the dipoles) to keep everything in phase at crossover makes a substantial difference, while additional subs around the perimeter not only doesn't improve the bass but degrades it. Even a single sub centered between the mains and crossing as high as 80 Hz. sounds better than any combination of multiple subs scattered around the room that I've tried (and without any "localization issues" at all). What works below 100 Hz. depends almost entirely on the room you're working in/with . . .
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.