xpert said:
Why should that be the case? Do you want to create an anechoic chamber? Is there a certain percentage of a room's surface that needs to be treated by default? What would be the "barely acceptable panel size"?
I feel confident that every room is different and needs to be treated like that.
Best, Markus
Never said anything bad about your site besides that it needs a redesign.
Marcus
They imply that higher modes also act as absorbtion and that's just not true. Since the wavelength in air is so much greater than in the panel ONLY the lowest mode is excited by the sound and hence ONLY the lowest mode can contribute to the absorption. They are correct in stating that HOW the panel is supported has a big effect on this lowest mode (but that is well known), and their method of support does, in fact, add a lot of absorption because it allows a lot of displacement for a gven panel size. Thats a good idea. But the higher modes do virtually nothing. That part is wrong.
Think about it, At 80 Hz the wavelength in air is about 16 feet, give or take. The panel is about 4 feet wide. That means that the entire panel can fit into 1/4 wavelength of the sound field - virtually completely uniform pressure excitation across the entire panel. This large mismatch will not excite a mode that is only 1/4 the wavelength - period.
The situation IS more complex than that, because the wavelengths in the panel do go down a lot faster in the panel than they do in air, and at some frequency they will match. This is called "coincidence", and the panels excitation will jump way up - very high - and then slowly drop in amplitude due to its mass -6dB/oct. But below coincidence it is very well know that only the lowest mode is a major contribution. The guys who did that site know this too, I guarantee it. But the facts don't make good "copy".
They imply that higher modes also act as absorbtion and that's just not true. Since the wavelength in air is so much greater than in the panel ONLY the lowest mode is excited by the sound and hence ONLY the lowest mode can contribute to the absorption. They are correct in stating that HOW the panel is supported has a big effect on this lowest mode (but that is well known), and their method of support does, in fact, add a lot of absorption because it allows a lot of displacement for a gven panel size. Thats a good idea. But the higher modes do virtually nothing. That part is wrong.
Think about it, At 80 Hz the wavelength in air is about 16 feet, give or take. The panel is about 4 feet wide. That means that the entire panel can fit into 1/4 wavelength of the sound field - virtually completely uniform pressure excitation across the entire panel. This large mismatch will not excite a mode that is only 1/4 the wavelength - period.
The situation IS more complex than that, because the wavelengths in the panel do go down a lot faster in the panel than they do in air, and at some frequency they will match. This is called "coincidence", and the panels excitation will jump way up - very high - and then slowly drop in amplitude due to its mass -6dB/oct. But below coincidence it is very well know that only the lowest mode is a major contribution. The guys who did that site know this too, I guarantee it. But the facts don't make good "copy".
markus76 said:
Hi,
O/k, room surface area == of boundary == walls/ground/ceiling ~100..200sqm. Treated area may be 5sqm - and that's a lot if rooms purpose is for living, meals, guests, games, telli ... and that is only 5..3% of the total area. Let it luckily have absorption of 100%. The total absorption coefficient of the room then would be 0.95 to 0.97. Anything You don't agree with?
so long
markus76 said:
Hi,
If You treat Your room for sports only, have fun!
I was trying to say (citation from above):
"Markus,
The caveat is the relation between the area of panels (any kind) and the total boundary area of the room. The effect of even barely acceptable panel size is to low to mention.
my 2 cc"
You asked me following to that (citation again):
"Why should that be the case? Do you want to create an anechoic chamber?"
The ratio between the area that can be treated within the room and the grand total of the room boundaries (walls) is frustrating. That is and has been my very declaration. Damping is not to effective. That for back to the multiple sub approach (already well known long before Geddes)?
Just for the record: why not use (but must be massive) spanish walls to seperate the room into smaller semi open compartments to dense the (non) modal structure?
Thank You
markus76 said:
Hi,
how much of a room can be treated? If it should remain a living room. May be to live together with regarding audio less enthusiastic personalities. Typically the surface area of a room is in total 100sqm to 200sqm. What a percentage would You cover with treatening?
It's on You
What do you mean by "It's on you"–the "Bringschuld" is on me? I don't think so.
But to answer your question: I would evaluate what potential problems there are and how to quantify them. Maybe everything sounds good (to you) without any treatment, so why change anything?
Best, Markus
P.S. Every room has a ceiling–pretty huge area.
But to answer your question: I would evaluate what potential problems there are and how to quantify them. Maybe everything sounds good (to you) without any treatment, so why change anything?
Best, Markus
P.S. Every room has a ceiling–pretty huge area.
O/k,markus76 said:
this board is dedicated by itself not to the enthusiastic but to the fanatics (disclaimer). It's on You what to enjoy the most - a partner with maybe more common interests or a not ever really perfect stereo set.
The quintessence is in my understanding that treatening 4 square meters (10 sq. feet) is prone to redicule. I expect at least the threefold to be minimally effective. That becomes a cost issue too rather fast.
so long
I once visited a room where aproximately half the room was full of diffusers, speakers at different positions, plant's, etc. "Trying to get to the equipment and wall sockets was like walking through a jungle. Similar arrangement was used when on display at High End Audio shows. If the speakers could not generate enough bass, then it really sounds like small speakers.xpert said:
Group Delay
Can we revisit the group delay issue again. I'm still confused on this issue. and it's been on my mind.
Using one of the many subwoofer simulator software's such as WinISD, or even my acoustics calculator, I can calculate graphs of the group delay with various boxes and various filters added in. If these are accurate or not I am not sure. You can not specify the type of filter, IIR or FIR for instance, but can mimic them if you are creative I think. I'm sure much more accurate models of group delay could be modeled in something like Matlab for those who better understand the math, and I'm curious just how serious an issue it is.
Now going with what we get from something like WinISD, and assuming it's results are at least close to accurate, the more filters you add, the greater the group delay. Steeper filters create steeper group delay increases with frequency. Highpass filters create group delay above their cutoff filter and low pass filters create group delay below. Again, this is just, simplistically, what something like WinISD shows when you add filter.
Now for the real world problem. All plate amps that I have seen or used have non-defeatable high and low pass filters. Many claim that the LFE input defeats the low pass filter, but my own measurements and those shown by companies like parts express indicate it simply sets it to a highest frequency point, but that a low pass filter still exists. They also include highpass filters, often of the peaking 2nd order type.
If you model all of these filters on a subwoofer, you find that group delay cumulatively builds up quite seriously. I know that there is no definitive research indicating anything legitimate about group delay below 100hz. As I understand things, we don't know how large group delay needs to be to be audible, nor do we know what it would sound like. However, if we extrapolate from the research below what we do know, it would seem that group delay would have to become progressively larger to have an audible impact, as we go down in frequency. John Murphy has some information on his website that takes the Blaurt and Laws paper, turns the findings into time cycles, and then uses 2 time cycles to roughly equal the audibility threshold. This then coincides well with the delay numbers I've heard as being audible in low frequencies, something like 50ms. This would indicate that even the very high group delay numbers we see with all these filters is largely inaudible, but I don't know if that is true or not.
Again, this is just my take on all the information floating around on the web, I have no idea if any of this is accurate. I'm looking for some opinions here. If the WinISD models are accurate, it would seem that all of the high and low pass filters (including the acoustic filters of the box design) would cause group delay to be in the 10+ms range at 50hz and 30+ms range at 20hz. The worst case scenario I could model was a sub-optimum 8th order bandpass box with high and low pass active filters added with 6th order slopes. All of this together created group delay in excess of 20ms below 100hz, and exceeding 50ms by 30hz. That required a spice model, and I used pspice and used a series of equivalent circuits, with my speaker model coming from Klippel's article, which I believe to be a more complete and accurate circuit. Again, no idea if it would be correct to cascade these equivalent circuits to get what the total system group delay is. It also doesn't include the impact of the room. Is the room another variable that needs to be considered? It has filtering effects right?
Can we revisit the group delay issue again. I'm still confused on this issue. and it's been on my mind.
Using one of the many subwoofer simulator software's such as WinISD, or even my acoustics calculator, I can calculate graphs of the group delay with various boxes and various filters added in. If these are accurate or not I am not sure. You can not specify the type of filter, IIR or FIR for instance, but can mimic them if you are creative I think. I'm sure much more accurate models of group delay could be modeled in something like Matlab for those who better understand the math, and I'm curious just how serious an issue it is.
Now going with what we get from something like WinISD, and assuming it's results are at least close to accurate, the more filters you add, the greater the group delay. Steeper filters create steeper group delay increases with frequency. Highpass filters create group delay above their cutoff filter and low pass filters create group delay below. Again, this is just, simplistically, what something like WinISD shows when you add filter.
Now for the real world problem. All plate amps that I have seen or used have non-defeatable high and low pass filters. Many claim that the LFE input defeats the low pass filter, but my own measurements and those shown by companies like parts express indicate it simply sets it to a highest frequency point, but that a low pass filter still exists. They also include highpass filters, often of the peaking 2nd order type.
If you model all of these filters on a subwoofer, you find that group delay cumulatively builds up quite seriously. I know that there is no definitive research indicating anything legitimate about group delay below 100hz. As I understand things, we don't know how large group delay needs to be to be audible, nor do we know what it would sound like. However, if we extrapolate from the research below what we do know, it would seem that group delay would have to become progressively larger to have an audible impact, as we go down in frequency. John Murphy has some information on his website that takes the Blaurt and Laws paper, turns the findings into time cycles, and then uses 2 time cycles to roughly equal the audibility threshold. This then coincides well with the delay numbers I've heard as being audible in low frequencies, something like 50ms. This would indicate that even the very high group delay numbers we see with all these filters is largely inaudible, but I don't know if that is true or not.
Again, this is just my take on all the information floating around on the web, I have no idea if any of this is accurate. I'm looking for some opinions here. If the WinISD models are accurate, it would seem that all of the high and low pass filters (including the acoustic filters of the box design) would cause group delay to be in the 10+ms range at 50hz and 30+ms range at 20hz. The worst case scenario I could model was a sub-optimum 8th order bandpass box with high and low pass active filters added with 6th order slopes. All of this together created group delay in excess of 20ms below 100hz, and exceeding 50ms by 30hz. That required a spice model, and I used pspice and used a series of equivalent circuits, with my speaker model coming from Klippel's article, which I believe to be a more complete and accurate circuit. Again, no idea if it would be correct to cascade these equivalent circuits to get what the total system group delay is. It also doesn't include the impact of the room. Is the room another variable that needs to be considered? It has filtering effects right?
I guess I interpreted your response to be that if it was only a few ms's, it wouldn't be an issue. My question was, how big does the delay need to become before it matters. Is it correct to say that, it would have to become so big as to be an unlikely or even impossible situation. I mean, if it needed to become greater than 50ms to be audible, then I would think that is an unlikely if not impossible situation. If on the other hand, at 50hz, greater than 20ms was audibly a problem, I can model situations where this is true. While unlikely, I see people do this sort of thing (8th order low pass filters for instance).
markus76 said:
It's my understanding that the any type of panel resonate trap works by reducing the frequency response in an inverse relationship to the resonance of that panel. In other words, the panel will resonate with a specific center frequency and Q, and this is the fashion in which energy is going to be absorbed. If you then dampen the panel, such as with foam, the Q is lowered, but not the frequency, and thus the peakiness of the absorption is reduced, but so is the absolute amount of absorption. The amount of dampening then creates a trade off between total amount of absorption the panel will have, vs bandwidth. From this standpoint, the foam doesn't really add absorption, in fact it reduces the absorption. However, you could argue it increase absorption if you consider that you now have more absorption at frequencies above and below the center (relatively speaking). Since the absorption level of a panel trap with no dampening would be very peaky, meaning very high at the resonant frequency, but with very little to the sides, it's very desirable to add dampening and broaden that, then use multiple traps to bring back up the absorption coefficient.
Now, I don't see how this is contradictory to what is presented on that web page you showed. It appears to be the same basic thing, no?
As to a comparison between RPG and Hunecke's designs, I don't think they are identical. They work on the same principles, as do all panel traps, but are implemented differently, I think. I've never owned or taken apart an RPG Modex, so I'm basing this on their description, and my talks with Peter. I believe that the RPG traps use a metal panel that is un-dampened directly, i.e. no foam is touching the actual panel, and suspended (clamped) at it's edges. The only dampening is inside the chamber. According to RPG, the metal resonating panel converts the high pressure low velocity energy into low pressure and high velocity, which can then be absorbed by the foam or fiberglass dampening. As I understand this, it's no different than if you were to build a big box and put a passive radiator in it. There would be a specific resonant frequency at which the passive radiator resonates at, and adding box dampening into the box would work just like adding it to any speaker box, but in this case, would also act to lower the Q of the tuning frequency, reducing peak amplitude, but increasing bandwidth some.