Seperate drivers for LF and ULF or one driver? - Page 8

Beast if you are down for hosting this that would be great. I know that Bosso lives fairly close. I could drive in. I can bring any number of different drivers. I have a large assortment of good pro drivers and subs. I know that Dave has quite a few as well. Wi also mentioned showing up and he has a bunch of drivers too so I don't think coming up with those will be a problem at all. Same deal with amps, cables and EQ units. I can bring whatever. Either myself or Dave can provide the measurement rig. If Dave wants to that would be fine with me as I wouldn't mind checking his out anyway. I could put all of the final results and impressions up on Data-Bass for posterity. We would need a screen to act as a blind to hide the subs from the listeners and probably 2 volunteers to move the subs in and out and hook them up that could still listen but would know what system was playing so couldn't take part. It would probably be best to limit the number of attendee's to a somewhat manageable number to help things flow smooth and keep things from getting over crowded. It's your place so just a suggestion. If we are serious about this I think it would be worthwhile to get a strong group of experienced members for this and take our time planning the best equipment, approach and time frame.

Name a date and time and I'll fly in.

 

 

I'd be willing to help out, very interested.  So long as its on a weekend of course.....  Let me know if I can help.

My guess is that if you overlay the GP response, you'll see that there is no suck out at 12.5 Hz, that it will be at or slightly above the outdoor number at 12.5 Hz. That shows a lack of room gain, not a suck out.

If your floor vibrated less than violently, say 1/16 inch peak-to-peak, or 1/32 inch each way, that would be approximately 180 liters of displacement. Hard to believe that would cause a suck out. In order to absorb your subs energy to affect a suck out, the floor would be barely moving, hardly perceptibly moving. Moving more than that (or violently), but still obviously below L360 structurally, it would have to show as an increase, not a suck out.

Some of your measurements show a drop in gain below 15 Hz, some show a leveling off below 15 Hz and the one you posted here shows huge gain with a pretty wide dip at 12.5 Hz. I would not conclude anything from that unless the "didn't matter where the subs were, how many or where the mic was" is qualified and accompanied by its pertinent graph.

That's a pretty wide notch to be from a resonance, IMO. As well, the resonance isn't causing a problem at any multiples of 12.5 Hz, which is, also IMO, unlikely.

I also find it most unlikely that the room is pressurized, but that the room is moving exactly 180 degrees out of phase with that pressurization at only one frequency, reducing PVG to a normal pressure wave, or something like that.

This is exactly what is shown in the second graph I posted right directly above your comment? The result is a net of no gain over outdoors at the suckout.

Why? It could be vibrating out of phase with the subs which would cause a reduction in energy.


I have no idea what you are talking about here.

The first graph posted is an equalized response curve but the suck out is there in the same basic shape regardless as I don't try to PEQ below 20Hz. The second graph is showing only the amount of gain seen at the listening position over groundplane which again shows a net zero gain at that point the same basic shape and very large amounts of gain everywhere below 30Hz otherwise. The drop off below 4Hz is probably loss of measurement resolution and equipment roll off.

This feature is prominent and can be seen in every measurement taken in this room regardless of system configuration, or which listening location had the mic placement. I also have the actual gain over outdoor groundplane seen at each listening position from each of 4 subwoofer placements in the room. It is present in all of those measurements as well. What more would be needed to reach some sort of conclusion?

I'm skeptical as well but if not this then what is it? If this does not make sense then what does to explain it? I've yet to hear a more sensical explanation for it.

Anyway I no longer live there so it is all moot at this point anyway, since I couldn't investigate anymore if I wanted to but I did find it highly annoying and curious at the same time. I am almost ready to start taking some measurements at the new place which will be a slightly smaller space with an attached hallway resulting in the dreaded L shape. Basement cement slab floor, 3 walls are exterior brick and or below grade and only the ceiling and one wall being wood frame. There is one window and a pair of glass doors to the back patio. If I were betting I would guess that this room will be much deader and offering a lot more low freq gain but much less tactile sensation. I would also bet that there will not be anything like the suck out seen in the old room. (Hopefully)
Edited by Ricci - 8/14/12 at 1:58pm

What does the group delay look like @12.5?

What about coincidence of seconday and tertiary room modes, ones possibly not easily calculated? Could that cause an isolated and broad dip like that?

Is a resonance necessarily additive? What about resonant absorbers? Its only additive if the resonant energy is confined to the system correct? What if the room possesses a resonance which while excited allows energy to escape the system?

What I meant is that I've saved most of the graphs you posted of various subs in that room.

Unfortunately, they aren't accompanied by where the mic was, where the sub was or where the subs were placed.

Neither did you ever measure multiple positions of subs and mic during the same session and study the outcome from the perspective of this discussion, below the so-called modal region.

And, neither do they all reflect the same pattern shown in the graph posted above.

That's what I was talking about there.

Regarding sympathetic resonances, they are always in phase and never 180 degrees out of phase. If the object absorbing the sound is gooey, it turns the sound into heat. But, if the object that's object absorbing the sound is rigid, and it begins to sympathetically resonate, the sound is not turned into heat, but radiated into the room. When the floor or a wall moves, it increases the output.

I asked LTD to harken back to the 6th grade or wherever he pulled the tuning fork example from and show how the sympathetic vibration of the 2nd tuning fork causes a suck out. Of course, it was rhetorical because it doesn't cause a suck out.

"I asked LTD to harken back to the 6th grade or wherever he pulled the tuning fork example from and show how the sympathetic vibration of the 2nd tuning fork causes a suck out. Of course, it was rhetorical because it doesn't cause a suck out."

it can and it does. resonant/absorptive room panels are a common method for turning acoustic energy into heat. ideally, the tuning frequency of these panels are centered on room resonances where the energy causes peaks. if they are not, they still work, and cause suckouts.

"Regarding sympathetic resonances, they are always in phase and never 180 degrees out of phase. If the object absorbing the sound is gooey, it turns the sound into heat. But, if the object that's object absorbing the sound is rigid, and it begins to sympathetically resonate, the sound is not turned into heat, but radiated into the room. When the floor or a wall moves, it increases the output."

no. if the walls are vibrating, you are losing acoustic energy to heat.
Edited by LTD02 - 8/14/12 at 3:56pm

I am pleased as punch with this prototype. It's only a few hours old so no charts yet, but I think having the ULF driver fire down helps. The ULF driver is a 15" Polk MoMo in a 3 cubic ft. sealed box. In my implementation, there is no back side on the LF box, it's operating as a dipole, which I feel also helps even out the performance. My first measurements with REW were markedly better than anything I've tried to date.

Ghetto Velo 1812


Edited by imagic - 8/14/12 at 7:05pm

Sympathetic vibrations below the modal region should always cause suckouts - conservation of energy would be violated otherwise.

The electrical circuit equivalent to a sympathetically vibrating wall is a parallel lossy bandstop filter. At resonance, the acoustic impedance is at a minimum, which effectively shorts out the acoustic pressure. Another way to think about it is that at the resonant frequency, the wall is easy to move. Because it is easy to move, the energy from the air molecules do work on the wall to create kinetic energy and heat. But because of the walls momentum, it takes some time to get the wall moving before it rebounds (just under 180 degrees) which causes any radiation to be out of phase with the incident source. The null isn't perfectly deep because some of that energy got converted into heat.

Another important thing to keep in mind is that when the wall is perfectly rigid, there is no sound transmission outside of the room. But when the wall vibrates, there will be sound radiating from both sides of the wall. That is even more energy leaving the room and is why you don't have a perfect 180 degree phase shift from the resonance.

As far as splitting the bandwidth versus sharing the bandwidth, I'm actually quite surprised there hasn't been any discussion about IMD.

There are two sources of amplitude modulation distortion: doppler effect, and motor non-linearities.

Doppler distortion is a direct calculation of bandwidth, and effective surface area. Double your bandwidth and IMD goes up 6dB. Cut it in half and it goes down 6dB. You get the same 6dB for doubling/halving the surface area. So if you have two 18" drivers covering 20Hz to 80Hz, you will get the same doppler distortion if instead you choose to have one cover 20-50Hz and let the other cover 50-80Hz. However, if you don't split the bandwidth equally, then the driver with more bandwidth will exhibit more doppler distortion unless you increase its surface area by the same amount. Btw, bandwidth in this case is calculated by linear frequency - not octaves....which is why horns sound so much better up top than down low (way different bandwidth requirements).

So then you've got motor non-linearities....at lower frequencies, distortion is usually dominated by the suspension stiffness and motor strength changing as a function of excursion. With the really nice drivers, there is a large region of linear excursion. First, you must ask yourself whether or not one driver will be able to linearly displace enough volume for your application. If the answer is no, then running two identical drivers in parallel will give you another 6dB of output - if that's not enough, then the question is totally a moot point.

But if one driver can displace enough volume with low enough distortion, then large excursion linearity is not a concern for the lower frequencies. However, at higher frequencies we have to deal with inductance modulation, suspension reflections, and port resonances (if its a vented alignment). So then the question is whether or not the driver in question exhibits any of these issues over the desired passband.

Generally speaking, your (good) pro drivers are going to have better high frequency performance because they'll be designed with stiffer suspensions (higher suspension reflections) and its easier to keep the sensitivity high when not digging as low (which also means less inductance modulation). The port resonances are also naturally higher because the diameters can be smaller and the lengths are much shorter thanks to the higher tuning point.

However, if your excursion monster doesn't exhibit any inductance modulation or suspension reflections within the passband, then it will sound every bit as clean as the pro driver at the higher frequencies. Now double up on the excursion monster and your low frequencies will have 6dB more headroom.

So all that to say, I would suggest buying the best driver you can find for the desired passband and then just go with two of them. Generally speaking, it's cheaper to achieve the same displacement with two drivers than it is with one driver. Really the only time it makes sense to split the passband is if you have excess displacement down low (which is almost never the case) and you can find a really cheap driver to cover the higher frequencies....or the scenario Mark Seaton describes where the room acoustics and aesthetics dictate alternate solutions.

Btw, the reason I bring up the doppler distortion is because it sets a bottom limit on the broadband distortion numbers....at some point, a more linear driver provides no benefit. There is also significant research showing that IMD is more audible than THD. Unfortunately, IMD is harder to quantify in a single number and requires way more complicated test setups (which usually means its harder to get comparative correlation). I'm wondering if it doesn't make sense to do IMD sweeps at every local excursion min/max (2 points for a sealed box, 4 points for a vented box, and probably 8 points for a tapped horn). Anyways, the research shows that IMD is one of the most dominant factors and yet it's almost never discussed. I believe strongly though that IMD would explain why the "pro drivers sounds better than the bottom feeder sub at higher frequencies".

Oh, I almost forgot to mention that if the single driver has enough displacement for the lower frequencies, then the IMD products are going to be cut in half since doubling up on the driver will also cut the excursion in half...we can see from the measurements that the distortion knees tend to be very steep, so if you can cut the excursion in half, then you're in a way more linear region of the driver's performance.

Also, xovers are never perfect and impart their own time domain artifacts as well (both spatial coupling and filter group delay).

"There is also significant research showing that IMD is more audible than THD."

do you happen to have a couple of links/resources on this one?

as you probably know, geddes has done some original work and has concluded that for practical purposes thd doesn't matter in the bass region and that harmonics can actually be preferred by untrained listeners, particularly the 2nd.

Not true.


Also not true.

could you explain the first "Not true."

"conservation of energy" is a technical term that means energy is not created, but can only be transformed from one form to another.

mbentz is saying that if a subwoofer puts energy into a room and that some of that energy goes to vibrating the wall at a certain frequency and that energy is lost to heat, that energy will come out of the room at that frequency, and there will be a suckout.

if it did not work this way, there would have to be another source of energy other than the subwoofer.

....

edit: after thinking about it again, it seems what may be happening with a resonating wall is that it is moving "out" and increasing the volume of the room nearly as much as the sealed woofer is moving "out" and trying to decrease the room volume.

if the wall and the sub are moving together in this way, there will be no net change in the volume of air in the room and the sound level will drop to wheree it would be if as measured outdoors or perhaps even lower.
Edited by LTD02 - 8/15/12 at 10:03am

Somebody asked for a waterfall plot.


Note the general ugliness near 12Hz. the strength actually appears to increase after a while.












Most people do not detail much of this in measurements posted. It is assumed that the mic is at the LP and the subs are wherever fits into their room. I know all of this information for my own measurements. Others brought the low bass suck out I measured up and it started being discussed so I threw out a couple of graphs that were handy that show it. If you want more detail all you have to do is ask.

Wrong. I did.

I have measurements from at least 6 or 7 different sub placements not counting combinations of multiples at these locations and from probably 10 or more mic positions throughout the room. Various different subs, types, etc. More importantly I also have measured all 15 possible combinations of operation for 4 corner placed subs at each of 6 different listening positions and can look at the changes in response and gains seen from adding the subs in any possible combination at any of those 6 listening positions. Any sub individually, any combination of 2, any combination of 3, all 4 operating for any of those 6 listening positions, etc. I also have the measurements of groundplane response at a known distance and drive level compared with and subtracted from the in room response at each of these same 6 listening positions at the same drive level and each of the 4 corner placements for the sub. I know how much gain over groundplane is seen at each of the placements or listening positions and I also know at each listening position what the response of each sub is and each combination of subs.

Here are the 15 measurements from 1 of the 6 listening positions. Note the suckout ever present 10-14Hz. The other 75 measurements are no different.



Here are the 25 ground plane versus in room measurements shown without the ground plane response subtracted. You can see the uptick below 4Hz in the outdoor measurement which is probably noise floor which results in the weird spike shown in the earlier chart of just the gain. I don't consider the gain below 5Hz as reliable due to this.



Here is how the response changes going from 1 to 2 to 3 to 4 subs operating at one listening position. I can start with any of the 4 subs individually and go to any combo of two from there followed by the measurements of any three I choose so there are quite a few combinations possible. Shown is the combination that results in the least amount of ULF gain.

Also note none of my measurements I have taken in any room show any reduction in ULF energy let alone a dramatic one that you have shown by adding subs. I'm curious as to what your theory to explain that is? Reversed electric polarity of a unit still makes the most sense to me as the easiest most likely explanation. Have you ever repeated this measurement?



Again wrong, they do. There is a suckout between 10-14Hz . Sure it shifts slightly in shape or up or down in frequency depending on sub placement and mic placement but it is there in its basic form at any of the places that a person would have listened from or I would have placed a subwoofer.
Edited by Ricci - 8/15/12 at 1:13pm

That's not what he said at all. He said:

He also said:

As I said earlier, if the wall is gooey, the sound energy is absorbed and turned into heat (suck out). If the wall is infinitely rigid (a prerequisite to your pressure pot theory) all of the sound energy is reflected back into the room (pressure pot increase).

Reality is somewhere in between those unreal extremes. When the sound energy causes the wall to move at the resonant frequency of the wall (sympathetic vibration), there is a formula to predict what happens.

that is exactly what was said:



so i will repeat:

could you explain the first "Not true."
"conservation of energy" is a technical term that means energy is not created, but can only be transformed from one form to another.
mbentz is saying that if a subwoofer puts energy into a room and that some of that energy goes to vibrating the wall at a certain frequency and that energy is lost to heat, that energy will come out of the room at that frequency, and there will be a suckout.
if it did not work this way, there would have to be another source of energy other than the subwoofer.

First things first:

You and others call the dip a suck out, I don't see it that way at all. IOW, the wall or the floor is not sucking out a notch at 12 Hz. Rather, the floor is adding SPL below the dip, especially if the subs were run naked (no L/T boost), and, I beg to differ, the mic placement is all that if the floor is having and influence, either way.

In measurements I have saved, there is a drop off and the response is level from there down as opposed to the above response which shows a large peak below the dip. At +12dB, that makes it a very large difference to me regardless of how anyone else reads it.

As far as the billion traces on a single graph (which I have not seen until this post), as I said, there is no explanation attached other than you telling me you have lots of combinations of measurements. At a glance, there are HUGE differences in those traces, so I'm not sure how critically you viewed them to conclude they're all the same save a bit of shift in the dips center?

The graph I copied in this post is easier to get to the bottom of. Is the mic in the same position for all 4 traces? Are the subs stacked, co-located or placed separately, and if separately, where for each trace? That's what I mean by saying you haven't done a post with that info that I ever saw. If so, and it's not too much hassle, just link me to it and I'll go from there.

As far as explaining the loss of ULF in the "both" graph I posted, it was positively not polarity. I just ran a R, L, R+L W/Reversed Polarity on R last night to show that the drop is from cross down and much more of a drop when polarity is reversed. There's a lot to be learned from this exercise as well, BTW. I've been dropping pieces of what I believe happens down low in this thread but there's a lot of noise burying it in irrelevant and unrelated and/or partial truths debates so far. I'll sum it up when all is said and done, FWIW,, YMMV, etc., if I hold out that long.

Not true.

See the above answer.

well, if you are going to quote wikipedia, at least get it right.

continue reading the article and you will get to:

"The damper, instead of storing energy, dissipates energy. Since the damping force is proportional to the velocity, the more the motion, the more the damper dissipates the energy. Therefore a point will come when the energy dissipated by the damper will equal the energy being fed in by the force."

every wall is naturally damped, or again to quote your resource, "If no damping exists, there is nothing to dissipate the energy and therefore theoretically the motion will continue to grow on into infinity."

http://en.wikipedia.org/wiki/Vibration

Unfortunately, the damping behavior of the mass on a spring mentioned above is not representative of how a sympathetically vibrating surface affects the acoustic output in the room.

The mass on a spring partially describes the motion of the wall, but ignores the acoustic coupling (ignores the fact that the wall and the air pressure causing it to move are interacting with each other). Let's say the wall is completely undamped - its amplitude would approach infinity if you stimulated it with an imaginary force. But in the acoustics analogy, that force is generated by air pressure acting on the wall. When a positive air pressure hits the wall, it causes the wall to move outward, and because it's undamped, it wants to continue moving outward. However, once it moves far enough that the air pressure is no longer positive, then the air molecules start pulling back on the wall and effectively damp its motion. In the process, the work being done to dampen the motion of the wall results in a reduction of the air molecule kinetic energy inside the room. In other words, the pressure at the wall gets cancelled by the motion of the wall. That is what the 180 degree phase shift represents and why it can be described as an acoustic short-circuit. The same thing happens in the opposite direction when a negative pressure hits the wall surface - in that case, the wall gets pulled into the room, but once the air molecules start compressing, they push back on the wall and lose energy by damping its motion.

The reason it is called a suckout is because the dip wouldn't exist if the wall were rigid. The total acoustic pressure inside the room has been reduced by the sympathetic vibration.

I would suggest that you are making a semantic argument by trying to define a different reference point that adds no further insight, and more importantly doesn't change how the problem would be addressed. Or maybe we're all missing the point that you are trying to offer a better solution? Care to elaborate on the practical implications of your redefinitions?

Bosso.... care to expand on your thoughts... other than arguing that "you're wrong and I'm right.... "

We went 'round this before on Patrick's thread, you said on

post #248 of 505
8/30/10

"Agree, except you're mixing 2 of my thoughts here.

Just looking at the many graphs that progressed to the final FR, I imagine the number of filters used to get there. Usually, EQing a single signal takes less EQ.

The other part is the loss of everything below 20Hz. Of course, we can't see what's happening below 15Hz on the graph, but assuming there is a wholesale loss of <20Hz, It's my opinion that it is NOT sympathetic vibration of the floor system. If the floor is vibrating, there should be a point where the output JUMPS due to hitting the resonant freq of the floor.

Here's a graph of a) one sub playing solo in right corner, b) another identical sub playing solo in left corner and c) both subs playing at once.

NOTE: Left corner is 1M further from LP than right corner.

Above PVG frequency, the expected boost in output is evident (+4dB), but beginning at around PVG frequency and below is a loss of output. The total swing of 12dB from what it should be vs what it is is significant and wholesale.

So, I know for a fact that the dual sub loss isn't from sympathetic vibration, but rather is a phase cancellation issue.

The DCX has separate output delay settings, which is the sole reason I want to try one, but...

BTW, didn't you send me traces of a loopback of the DCX? Please refresh my memory... where is the roll off?

Bosso"

Just like when you advised him to use '12 db of boost' in their multi-sub system, when measurements show that there was basically no room gain to be had, close to an outdoor condition.

I'd like to quote again:

Quote:
Originally Posted by bossobass View Post


Get the bassis, set it according to Ilk's FR and start with 12dB boost setting. No driver will be blown. Crank the sweep level to around 100dB and expand the graph to 5Hz.

Size, schmize... 5Hz is 220 feet long. The room size is irrelevant. If the room isn't boosting it, it ain't there to be boosted, as Josh suggests, it might be signal chain roll off or inadvertent filter in line, but it's not the floor eating it and it's not because the room's too big.

Or, whatever.

Bosso

Me: "I will say my condolences now for the passing of 3 x Maelstrom-X drivers, one TC 3000 15" with a delaminated coil former and one TC LMS 5400 that's no longer in service.

12 db boost would not be my recommendation."

Your 'recommended' boost + party volumes = 6 nicely dented LMS 5400's.... but according to you, no driver would have been blown. Signal levels were set throughout the system properly.

I start with a tapped horn sub that runs from 15 - 40 Hz. The lowest room mode is at 44 Hz which has a big impact on the way I'm doing this. The sub runs below it and above I use sealed boxes for Magnum 18" drivers. The mode helps extend the bass down and in a sense acts like a port. So I can make 60L boxes work. The woofers were required for the Synergy horns and I already had the Rythmik subs. All things considered there is a "synergy" in this approach for this particular system and room, considering the parts on hand and what the room is doing.

If the subwoofer system is designed properly (excursion control and not overtly overpowered) no driver would be damaged with boost. Could be 48dB of boost that matter. It's simple signal shaping. Obviously those with damaged drivers were either overpowering or there was no excursion control or both. It's nearly impossible to design the 5400/LMS-U with excursion control though cuz of it's insane motor. Many others have bottomed it out and will continue to in the future.
Edited by Scott Simonian - 8/16/12 at 2:57am

You claim that sympathetic vibrations "always" cause a suck out. I'm pretty sure I didn't misread that statement.

It seems to me that you guys are suggesting that Ricci's room only has a single resonant frequency below 30 Hz, that all resonant frequencies in a room excited by the sub cause a suck out and the rest of the profile is gain from the cones movement in and out, changing the rooms volume, to which I reply; baloney... prove it.

You've offered zero proof and no data because there exists none. Just errant interpretations of what forced sympathetic vibrations do to the response.

R is not "always" >>1. To say outright that the opposite is the case is simply not true, as in not true. In fact, in the context of this discussion, the better argument would be to say R is "never" >>1.

In Ricci's case, the so-called suck out is painfully obviously not the result of 180 degrees opposing vibrations. I would suggest that you aren't able to interpret a graph.

My floor is of the same construction and similar size to Ricci's. Where's this 180 degrees suck out that "always" occurs when the subs force a sympathetic vibration in my floor (which is easily felt and occurs around 6 Hz)?



Where is this suck out at any frequency? Is it that you believe my room has no resonances below 100 Hz that are forced into sympathetic vibration by the subs?

I didn't quote Wikipedia, I pulled the graph from Wikipedia and posted it here. Which part of that process did I not get right? It's you who quoted Wikipedia, not me.

Read the quoted section in my post. If it's your contention that Bentz is correct in declaring that R is "always" >>1, then we have nothing to discuss from here. You've added nothing to the discussion. Your stance is that rooms in homes are pressurized by subwoofers drivers cones below some formula-derived frequency, that when you open a door or a window you have instantly transformed your room into the inside of a bandpass subwoofer and that forced sympathetic vibrations of the resonant frequencies of the rooms walls, floor and ceiling are always destructive to those phenomena.

I got it, I get it and I've had it.

Problem is, in actual posted data, your predicted magic frequency is "always" wrong, the suck out that you now claim "always" occurs, seems rarely to occur, if ever, and why I should continue to debate this subject with someone who, if he even owned a subwoofer at all, would prefer a pro sound sub that rolls off and dies well above the region we're discussing is causing me to consider committing myself for observation.
Edited by bossobass - 8/16/12 at 6:57am

Not sure what I'm supposed to be commenting on, but I assume it's the bolded text.

Exactly correct: The suck out is NOT caused by any of a rooms forced sympathetic vibration at any of the resonant frequencies of the rooms boundaries.

Exactly correct: Unless the signal chain roll off and other specifications of the measurement system and methodology are posted, any dip or alleged suck out below 10 Hz is beyond dictates from observers of a graph.

The final point seems to be some sort of axe to grind regarding my L/T boost suggestion.

Get a Bassis... check.
Start with Ilk's GP FR... check.
Start with +12dB of L/T boost... check.
Sweep level at 100dB... check.

How did ^^this^^ suggestion blow a driver?

WTF is "party volumes"??? and what does it have to do with ME?

Sympathetic vibrations of natural structures necessarily dissipate energy as a result of their inherent damping and also typically via lossy transmission through the structure. Net effect is reduced total energy within the room, which would only be fully defined on integration of the response over frequency, time and spatial position. Resonances however store energy, and can return that energy into the room concentrated into a narrow bandwidth. Depending on mic location and frequency, you may see a peak or a suckout as the structure is simply behaving as another acoustic source and sums and interferes as all do. Thus the net effect is largely a redistribution of energy with a component of averaged energy reduction as well.