Todd
I think that DBA being a special case of anybodies approach was the point that I was getting at earlier. IF one used eight subs, AND they put them on the opposing walls, THEN almost anybody would come up with a similar, or comparable solution. I just don't see this as practical. To me its not good engineering to throw all that you can at a problem. There is such a thing as simplicity and value in any solution and I don't see DBA as being that. SFM is closer, but I still think that I could do nearly as well with a whole lot less expense. But then saving the customer money has never been JBLs goal in life🙂
I think that DBA being a special case of anybodies approach was the point that I was getting at earlier. IF one used eight subs, AND they put them on the opposing walls, THEN almost anybody would come up with a similar, or comparable solution. I just don't see this as practical. To me its not good engineering to throw all that you can at a problem. There is such a thing as simplicity and value in any solution and I don't see DBA as being that. SFM is closer, but I still think that I could do nearly as well with a whole lot less expense. But then saving the customer money has never been JBLs goal in life🙂
Re: Re: Re: DC mode
Think of the room having very small height and width. Now at low frequency what we have is a plane wave propagating the length of the room. The delayed, out of phase drivers at the other end of the room are forming a perfect impedance match so there is no reflection for a perfectly rigid, sealed room. There is no comb filter effect. When losses are considered in the there is some comb filtering in the simulations because the wave arriving at the back wall is weaker than the wave launched from the back wall sources. I think this could be corrected by adjusting the level of the back wall sources. When the room has larger H and W dimensions the placement of the sources is as with your 4 x 4 array cancels or fails to excite lower order modes in the W x H plane (and some oblique modes too).
The rigid wall, no losses, L direction modes only simulation yields a response that goes at log (1/f), independent of listening position. If I omit the 0th order (DC) mode the response doesn't continue to rise to DC but starts to roll of at 6dB/octave at some point, like a 1st order band pass. The real case would be somewhere in between, maybe flat towards low frequency, then rolling off as DC is approached? When modes in the other directions are added back in they don't influence the result until the higher order modes affect the response (at higher frequency).
Sure, real room effect (wall admittance, losses, leaks...) will alter the results, just as they do all setups. My plan is to build a small "test room" say 4 x 4 x 6 and use arrays of 5 or 6" woofers to we how thing perform in a more realistic situation.
I am using these simulations to gain understanding of the problem and not to imply that they represent "real world results".
cap'n todd said:
Think of the room having very small height and width. Now at low frequency what we have is a plane wave propagating the length of the room. The delayed, out of phase drivers at the other end of the room are forming a perfect impedance match so there is no reflection for a perfectly rigid, sealed room. There is no comb filter effect. When losses are considered in the there is some comb filtering in the simulations because the wave arriving at the back wall is weaker than the wave launched from the back wall sources. I think this could be corrected by adjusting the level of the back wall sources. When the room has larger H and W dimensions the placement of the sources is as with your 4 x 4 array cancels or fails to excite lower order modes in the W x H plane (and some oblique modes too).
The rigid wall, no losses, L direction modes only simulation yields a response that goes at log (1/f), independent of listening position. If I omit the 0th order (DC) mode the response doesn't continue to rise to DC but starts to roll of at 6dB/octave at some point, like a 1st order band pass. The real case would be somewhere in between, maybe flat towards low frequency, then rolling off as DC is approached? When modes in the other directions are added back in they don't influence the result until the higher order modes affect the response (at higher frequency).
Sure, real room effect (wall admittance, losses, leaks...) will alter the results, just as they do all setups. My plan is to build a small "test room" say 4 x 4 x 6 and use arrays of 5 or 6" woofers to we how thing perform in a more realistic situation.
I am using these simulations to gain understanding of the problem and not to imply that they represent "real world results".
Don't think a DBA is that unpractical: double the front and rear wall and use the volume for all chassis mount directly into the wall. I think there are even AVRs with 2 sub outs that provide the necessary delay and phase invertion. 2 big amps and you're done. Then move on to frequencies higher than 100Hz 😉
Best, Markus
Best, Markus
What I forgot to mention is that SPL decreases only by 3dB per doubling of distance with a DBA.
Alternatively to using the front and rear wall for a DBA, one could also use floor and ceiling.
Best, Markus
Alternatively to using the front and rear wall for a DBA, one could also use floor and ceiling.
Best, Markus
markus76 said:
Markus - if its a plane wave then it should decrease with distance at all. The only losses will be at the edges. Its like John says, the wave starts out, propagates down the length of the room and gets actively absorbed by the sources at the other end. No wave spreading, no loss with distance.
Exactly as you say Earl. No significant reduction is SPL. Tomorrow I will do some simulations of Front wall, back wall DBA and floor formated, 4 woofer, 1/4 wave array and post the results.I am warning you all right now, 😉, these will be performed for rigid walls, no leakage and no damping. Thus they are academic in nature, but will serve to give some insight as to the behavior in this limiting case.
gedlee said:
Earl
My previous room had a professional full-circumference acoustic soffit w/ film maintaining reflectivity; current room has four subs following your advice & a deaf person could hear the performance improvement. I apparently mistakenly remember reading that your recommended NO bass absorption. Was that ever the case?
I have your book. What exactly is "some...good...bass absorption"?
It's hard to believe bass peformance might improve further but I'd welcome it. The floor & front wall are concrete. The room is symmetrical (HWD 7-7" x 16-7" x 25-9"). I don't understand how drapes could possibly absorb into the bass range. If recommend, proper floor-to-ceiling bass traps in three corners could be installed or possibly a partial soffit. I would insure all OC703 surfaces facing listeners are covered w/ a film to maintain mid-treble reflectivity.
gedlee said:
Now I was talking about a real room 🙂 A German guy measured that loss of SPL in his DBA.
Re: Re: Re: Re: DC mode
But there can BE no plane wave in a lumped parameter system (as the room is in the pressure zone), thus there IS no propagation of wave from front to back. Sound pressure is simply proportional to total displacement of the drivers (that is my understanding). Of course, sealed room and all that...
I know that in the pressure zone, the room should behave as a lumped parameter system. i also thought (but wasn't as sure) that this would be true for the 0th mode at all frequencies, in as much as one could separate it out from the rest (which really IS acedemic!).
john k... said:
But there can BE no plane wave in a lumped parameter system (as the room is in the pressure zone), thus there IS no propagation of wave from front to back. Sound pressure is simply proportional to total displacement of the drivers (that is my understanding). Of course, sealed room and all that...
I know that in the pressure zone, the room should behave as a lumped parameter system. i also thought (but wasn't as sure) that this would be true for the 0th mode at all frequencies, in as much as one could separate it out from the rest (which really IS acedemic!).
ro9397 said:
Bass absorption is a key problem because, yes, it is hard to achieve. The draps do provide some bass absorption, but not much, thats true. In one case we ended up filling the space behind the draps with anything that we could find that was absorptive. In my experience absorption behind the speakers always helps, both the bass and the imaging - the more the better. Toole even mentions this in his book.
True LF absorption can only come from a structural system like I use, where the walls actually are flexible and bend disipating the energy. This is very effective. A treatise on absorption is beyond the scope of web posts, but it is an extremely important aspect of the LF sound.
While I've never tried it, It would be possible to attach a dead wall onto an existing wall. Use RC-1 channel and then mount constrained layer damping onto that. This would yield a good amount of LF damping in a space of only about a 2" depth. Thats because its virtually all structural damping and not acoustic damping, which is a big difference. Structures can disipate a lot of energy while wave motion through some porous media is only marginally effective.
Re: Re: Re: Re: Re: DC mode
Todd
There is always some wave propagation, but as the wavelengths get longer and longer the spatial pressure gradient goes to zero - as you say, a pressure zone - but there is always a phase delay away from the source. Even in a port, which is assumed to be lumped parameter, there is some wave propagation. In the case of the DBA system its obvious since there is a delay from the front sources to the rear even at LFs. Of course the phase angle of this delay is going to zero at LFs.
The SPL being proportional to the sources is true only for a sealed room. In a real room with air exchange it depends on the loses - the Q of the Helmholtz mode. This is so commonly misunderstood. I studied this problem extensively back in my car audio days to understand how to get the most bass for the least inside of a vehicle. In smaller spaces (like cars) the Q of the Helmholtz mode can be fairly high, due to the small volume, and a lot of gain can be achieved. But in a real room this volume is much larger and the Q tends to go down and gets so low that no evidence of this mode is readily apparent.
cap'n todd said:
Todd
There is always some wave propagation, but as the wavelengths get longer and longer the spatial pressure gradient goes to zero - as you say, a pressure zone - but there is always a phase delay away from the source. Even in a port, which is assumed to be lumped parameter, there is some wave propagation. In the case of the DBA system its obvious since there is a delay from the front sources to the rear even at LFs. Of course the phase angle of this delay is going to zero at LFs.
The SPL being proportional to the sources is true only for a sealed room. In a real room with air exchange it depends on the loses - the Q of the Helmholtz mode. This is so commonly misunderstood. I studied this problem extensively back in my car audio days to understand how to get the most bass for the least inside of a vehicle. In smaller spaces (like cars) the Q of the Helmholtz mode can be fairly high, due to the small volume, and a lot of gain can be achieved. But in a real room this volume is much larger and the Q tends to go down and gets so low that no evidence of this mode is readily apparent.
Re: Re: Re: Re: Re: Re: DC mode
I already said "sealed room and all that..." so i get that.
yes, there is phase delay, but not a plane wave.
gedlee said:
I already said "sealed room and all that..." so i get that.
yes, there is phase delay, but not a plane wave.
markus76 said:
Funny, Markus, I had suggested exactly that very same product to Jimbo before. Is that what you have in your room? If so, can you tell me about your findings, perhaps in another thread (like your Abbey setup thread?) if it's too off-topic?
EG, thanks for answering my questions. As you point out, Toole discussed the beneficial effects of improved localization and reduced coloration with absorption behind the front speakers. I wonder whether much of the bass benefits you mention, as well as some of the reduced coloration, results from the absorption reducing the adjacent boundary effects like destructive acoustical interference off the front wall, which incidentally would change the off-axis spectral content due to varying path length. It had also seemed to me that the multiple subwoofer approach would not address a significant portion of the modal region below the transition region, assuming that the subwoofers would roll off somewhere between 100-200 Hz and a typical transition region between 200-300 Hz.
EG, TW, JK, I've enjoyed reading your discussions since I had idly wondered about pitting each approach against the other. Thanks, I'm learning a lot.
youngho said:
I would think that much too high in most rooms. I see this at more like half that number.
I'm very limited in what I can install in my living room. Everything has to be easily removable. Fortunately the walls flex a lot and do a pretty good job in dissipating low frequencies. So I'm focusing on reducing reverberation time which is pretty high. I'm using 10cm Basotect with a 10cm air gap to the walls behind the speakers and listening position. I'll add a porous absorber to the ceiling soon.
Best, Markus
Re: Re: Re: Re: Re: Re: Re: DC mode
I guess then that you would have to define what constitutes a plane wave, because to me if there is phase delay from one spatial point to another there is wave motion. Is it a "plane wave"? not necessarily I guess. But I was responding to your statement "But there can BE no plane wave in a lumped parameter system" which still does not seem correct to me.
cap'n todd said:
I guess then that you would have to define what constitutes a plane wave, because to me if there is phase delay from one spatial point to another there is wave motion. Is it a "plane wave"? not necessarily I guess. But I was responding to your statement "But there can BE no plane wave in a lumped parameter system" which still does not seem correct to me.
gedlee said:
I was going by figure 4.12 in Toole's book, but perhaps you are using the Schroeder frequency or are accustomed to larger rooms? Not a big deal. Markus's data at http://www.mehlau.net/audio/multisub_geddes/ shows some significant issues above 100 Hz that didn't seem to be smoothed significantly by the addition of the second or third subwoofers. It was this frequency region that I thought might benefit from bass treatments.