Leob...The length of the impulse response doesn't necessarily correlate with the frequency content. The ringing is affecting the time domain mostly, so you can have a dampened plate with smooth FR as well as a ringing plate with smooth FR.
...
Frequency and time are the both side of the same coin. They are not affected independently. Ringing in the time domain => some peaks in the frequency domain. I don't think it is possible to have a ringing plate (a "ringing time domain") with a smooth FR. It is the advantage of the spectrogram view to see frequency and time at the same time.
Christian
+ @VelericAs I understand it that graph is the actual filter response, not the FR of the panel. But yes, that shows around 5dB difference, and should mean that the actual raw response of the panel is off by the same amount.
I have in mind the same document. the good remark is to point that the maximum correction is less than -5dB.
Note that the FR is a "power" response = the averaging of 1349 FR. I have no idea how it change things compare to a single point FR. Most probably in the way of smoothness...There is no mention of the kind of smoothing of the FR neither of time windowing. Doing some tests with REW, 1/3 of octave smoothing seems possible.
By the way, Leob, the graphs from Steve are picture of a Behringer display. In horizontal the frequency, in vertical the dB. It displays the level in 1/3 of octave.
I had a DEQ2496 some years ago. Very handy as you can access very easily to filters.
I said the ringing mostly affects the time domain, not that frequency domain is unaffected. But the main signal is stronger than the ringing, and even if the ringing happens in certain band, getting rid of it only affects overall FR marginally.Leob
Frequency and time are the both side of the same coin. They are not affected independently. Ringing in the time domain => some peaks in the frequency domain. I don't think it is possible to have a ringing plate (a "ringing time domain") with a smooth FR. It is the advantage of the spectrogram view to see frequency and time at the same time.
Christian
Eric.
You have said it yourself marketing BS !!
It is only DML if we say it is.
Before NXT re invented DML it was called a soundboard.
As I have said before, get a life !!!
Steve.
You have said it yourself marketing BS !!
It is only DML if we say it is.
Before NXT re invented DML it was called a soundboard.
As I have said before, get a life !!!
Steve.
Hello EricI would never have said it, but Tectonic don't think your panels are DML's, Spedge. Flat panel loudspeakers yes, but DML no.
View attachment 1076736
View attachment 1076737
Now that part seems vaguely familiar...hmmm...where did I read that before??
oh, that's where!
About the optimization of the modes, I found this paper James Angus DISTRIBUTED MODE LOUDSPEAKER RESONANCE
STRUCTURES. The explanation of the different types of waves is interesting (general information). There is an original approach about the boundaries, the suspension. I made a quick test this afternoon on a panel tranforming it from a full surronding suspension to a random one (I removed some portion of foam according to a MLS sequence)... but the result doesn't convince me. The advantage of a "random" suspension seems more theoretical than demonstrated in practice (at least my test).
You might find some new ideas in the paper...
Christian
Thanks Christian,Hello Eric
About the optimization of the modes, I found this paper James Angus DISTRIBUTED MODE LOUDSPEAKER RESONANCE
STRUCTURES. The explanation of the different types of waves is interesting (general information). There is an original approach about the boundaries, the suspension. I made a quick test this afternoon on a panel tranforming it from a full surronding suspension to a random one (I removed some portion of foam according to a MLS sequence)... but the result doesn't convince me. The advantage of a "random" suspension seems more theoretical than demonstrated in practice (at least my test).
You might find some new ideas in the paper...
Christian
I did a quick read and it does sound very interesting. In the past I also tried some "random-ish" boundary conditions. My thinking/approach was pretty simple: I wondered if by creating less symmetrical mode shapes there might be less "self cancellation" of antiphase areas of the plate. But I never found anything particularly useful. But also I had no particular strategy for choosing points of suspension, so it felt a bit like complete trial and error and hence was a bit unsatisfying. Also, when trying to use FEM to analyze such suspension systems, it was impossible to guess which modes might be effective ones, and which would not. So I really felt like I was just hunting around in the dark.
But maybe it's worth another try.
Thanks,
Eric
The paper doesn't explain how the sequence was chosen. I even haven't recognize it... The full perimeter seems divided in 26. the maximum elements for a Barker is 13. It is why I tried a 5bits MLS to get 32 elements leading to 6.5cm on a 29x80cm panel.Thanks Christian,
I did a quick read and it does sound very interesting. In the past I also tried some "random-ish" boundary conditions. My thinking/approach was pretty simple: I wondered if by creating less symmetrical mode shapes there might be less "self cancellation" of antiphase areas of the plate. But I never found anything particularly useful. But also I had no particular strategy for choosing points of suspension, so it felt a bit like complete trial and error and hence was a bit unsatisfying. Also, when trying to use FEM to analyze such suspension systems, it was impossible to guess which modes might be effective ones, and which would not. So I really felt like I was just hunting around in the dark.
But maybe it's worth another try.
Thanks,
Eric
In the interesting point of the paper, there is the criteria of 3 modes overlapping and the mode bandwidth.
hkguy6;I bought and tried eps from 11 to 18k, they all have sound different. (heavier better)
The heavier I can purchase here they "marked" 25k but actually just around 18kg/m3.
The best one is an air-conditioner packing board I pick which at 41kg/m3 (pic2). It sounds so GREAT but there're no way to get this in a bigger side. It's original size is 59x38x1cm. I have to cut it in two to be L and R. Fortunately the sound signiture image size don't change just loss some low end. Smaller size also make the sound stage a little deeper. It's my default panel now.
I also want to try the graphite one but they're all comes with 5cm thick.
Tried seveal thickness of xps too but all sounds like toy even apply a "deep" sanding to remove the skin totally. Give up.
from your pictures it looks like you've tried many EPS samples from 1/2 and 3/4 in or 12.7 mm and 19.mm . Can you please tell me how they compare?
Thank you for your time:
Pete
Hi SteveEric.
You have said it yourself marketing BS !!
It is only DML if we say it is.
Before NXT re invented DML it was called a soundboard.
As I have said before, get a life !!!
Steve.
Please see post 6168 and please explain how you do the 50/50 pva coating
Thank you
Pete
Alaska129.
The pva when mixed should have the consistency and look of 0% milk, very thin.
If the EPS has a hard skin (which is the main cause of the EPS bad sound).
Lower grades do not seem to have the hard skin, so sanding is not necessary.
I gently sand this skin off, and brush on the thin mix of pva, which will soak into the soft surface and dry to a hard finnish.
I usually add a thin neat coat of pva when gluing on the exciter,about 1inch around the exciter area to help the surface to support the exciter weight,
my exciters only weigh 85g so do not need rear support, but heavier exciters will probably need support ?
EPS usually has dropouts between 100hz and 300hz ,so I would recommend using a low frequency speaker to help fill in this response for best sound, or it could sound thin in the lower midrange ?
My transmission line speakers work from below 20hz to 300hz hz in room usually, depending on the panel type .
I would usually only run the EPS panels down to about 100hz to save them from large LF movement and also low frequency resonance (RINGING).
I intend to purchase some 2ftx4ft x 20mm graphite EPS 70grade, for comparisons, and testing, and show the frequency response and do some recordings of the panel sound.
I also intend to do a 2ftx2ft EPS comparison to the proplex panel, just for fun😁
Steve.
The pva when mixed should have the consistency and look of 0% milk, very thin.
If the EPS has a hard skin (which is the main cause of the EPS bad sound).
Lower grades do not seem to have the hard skin, so sanding is not necessary.
I gently sand this skin off, and brush on the thin mix of pva, which will soak into the soft surface and dry to a hard finnish.
I usually add a thin neat coat of pva when gluing on the exciter,about 1inch around the exciter area to help the surface to support the exciter weight,
my exciters only weigh 85g so do not need rear support, but heavier exciters will probably need support ?
EPS usually has dropouts between 100hz and 300hz ,so I would recommend using a low frequency speaker to help fill in this response for best sound, or it could sound thin in the lower midrange ?
My transmission line speakers work from below 20hz to 300hz hz in room usually, depending on the panel type .
I would usually only run the EPS panels down to about 100hz to save them from large LF movement and also low frequency resonance (RINGING).
I intend to purchase some 2ftx4ft x 20mm graphite EPS 70grade, for comparisons, and testing, and show the frequency response and do some recordings of the panel sound.
I also intend to do a 2ftx2ft EPS comparison to the proplex panel, just for fun😁
Steve.
I noticed in the tectonic frequency plots, that they have a raised response below 1k or 500hz depending on which plots you are looking at, which I prefer for my panels.
They also ignore the spike above 10k ? Which is easy to correct.
There are also a couple of suck outs at 1m in the response too ? At about 600hz and 5.5k.
Which sorts itself out in the hemispherical power response.
Steve.
They also ignore the spike above 10k ? Which is easy to correct.
There are also a couple of suck outs at 1m in the response too ? At about 600hz and 5.5k.
Which sorts itself out in the hemispherical power response.
Steve.
To have the axis FR or at at least in the listening window might be interesting... The power response seems "tolerant". A change in color in the polar response is a 3dB step, the FR axis seems not so smooth. Nothing surprising with a DMLI noticed in the tectonic frequency plots, that they have a raised response below 1k or 500hz depending on which plots you are looking at, which I prefer for my panels.
They also ignore the spike above 10k ? Which is easy to correct.
There are also a couple of suck outs at 1m in the response too ? At about 600hz and 5.5k.
Which sorts itself out in the hemispherical power response.
Steve.
I did too much test on ratio size thickness and weight that I really can't remember all of them.hkguy6;
from your pictures it looks like you've tried many EPS samples from 1/2 and 3/4 in or 12.7 mm and 19.mm . Can you please tell me how they compare?
Thank you for your time:
Pete
Basically when I get a better sounding board I will throw away the inferior.
However some common result I studied was, 2cm is better then 1cm especially on large size. Any thicker then 2cm is not my sound. But it's not a constant, my best sounding eps is a 1cm borad with 41kg/m3 so density also a matter.
I suggest you try them all by yourself if possible. Every ears diferent.
Hope this helps.
on page 250 and 251 I talk about my 1cm low grade packing eps and the 1cm 70 grade eps , the rigidity caused by the pva skin seems more important for HF performance than density as the they performed similarly when sanded and coated.I did too much test on ratio size thickness and weight that I really can't remember all of them.
Basically when I get a better sounding board I will throw away the inferior.
However some common result I studied was, 2cm is better then 1cm especially on large size. Any thicker then 2cm is not my sound. But it's not a constant, my best sounding eps is a 1cm borad with 41kg/m3 so density also a matter.
I suggest you try them all by yourself if possible. Every ears diferent.
Hope this helps.
Later on I also made a recording of the two panels.
Steve.
Thank you SteveAlaska129.
The pva when mixed should have the consistency and look of 0% milk, very thin.
If the EPS has a hard skin (which is the main cause of the EPS bad sound).
Lower grades do not seem to have the hard skin, so sanding is not necessary.
I gently sand this skin off, and brush on the thin mix of pva, which will soak into the soft surface and dry to a hard finnish.
I usually add a thin neat coat of pva when gluing on the exciter,about 1inch around the exciter area to help the surface to support the exciter weight,
my exciters only weigh 85g so do not need rear support, but heavier exciters will probably need support ?
EPS usually has dropouts between 100hz and 300hz ,so I would recommend using a low frequency speaker to help fill in this response for best sound, or it could sound thin in the lower midrange ?
My transmission line speakers work from below 20hz to 300hz hz in room usually, depending on the panel type .
I would usually only run the EPS panels down to about 100hz to save them from large LF movement and also low frequency resonance (RINGING).
I intend to purchase some 2ftx4ft x 20mm graphite EPS 70grade, for comparisons, and testing, and show the frequency response and do some recordings of the panel sound.
I also intend to do a 2ftx2ft EPS comparison to the proplex panel, just for fun😁
Steve.
That is a great help
Pete
Yes the skin make the most different. You can clearly listen the big different when just simply knock the boards on ground. Without sanded the sound always honky and glassy. The sanded is more dull and less decay.on page 250 and 251 I talk about my 1cm low grade packing eps and the 1cm 70 grade eps , the rigidity caused by the pva skin seems more important for HF performance than density as the they performed similarly when sanded and coated.
Later on I also made a recording of the two panels.
Steve.
Sorry I don't use measurement. My test on density different is, same size thick but different density with same sanded and coated. Listen them one Left and one Right for different music in stereo and mono. Their difference show as imblance. Sometimes is the FR sometimes is the diamension.
This test is good for me but definitely not for communciate with others.
Hi Steveon page 250 and 251 I talk about my 1cm low grade packing eps and the 1cm 70 grade eps , the rigidity caused by the pva skin seems more important for HF performance than density as the they performed similarly when sanded and coated.
Later on I also made a recording of the two panels.
Steve.
I went back and reread the above post... then ended up just rereading from there 😉
I'm wondering when you need to cut or round corners of EPS what technique you use?
Hot knife and sanding perhaps ? My EPS is hot wire cut so I'm guessing start sanding with say 220grit?
Thanks
Pete
I'm sceptical if rounding the corners is useful at all. If will affect the modes a bit like any modification of the plate, but since you anyway need to adjust dimensions and clamping to get a smooth FR, I don't see the point of bringing in an extra variable.Hi Steve
I went back and reread the above post... then ended up just rereading from there 😉
I'm wondering when you need to cut or round corners of EPS what technique you use?
Hot knife and sanding perhaps ? My EPS is hot wire cut so I'm guessing start sanding with say 220grit?
Thanks
Pete
So far I have only seen tech ingredients video suggesting it is useful, but without any real data to prove it.
Alaska129.
Sanding large eps panels is a pain in the neck , and it is easy to accidentally score a groove into the material, if rushed.
You could sand the panel and only coat the panel an inch or two around the exciter, to hear how it sounds without the skin, similar to tech ingredients, I have an uncoated eps panel still lying around from years ago for comparisons.
The sanded surface will have a soft and fluffy feel ,and sounds a little soft too, tech ingredients seem to like this and say it acts as a sound absorber?
Just thought you might like to experience this first ?
Are you thinking of using the panel full range, or using a low frequency driver for support ?
Which exciter model are you using?
A fairly fine sand paper should be OK, you don't want to go too deep, it's more like scuffing the surface to get that fluffy feel.
The very thinned pva will just soak into the fluffy surface ,producing a better sounding hard skin, pva will take a couple of days to fully harden for full HF performance.
I use the nxt recomended 4/9 x 3/7 exciter positions myself ,for a more central position .
Steve.
Sanding large eps panels is a pain in the neck , and it is easy to accidentally score a groove into the material, if rushed.
You could sand the panel and only coat the panel an inch or two around the exciter, to hear how it sounds without the skin, similar to tech ingredients, I have an uncoated eps panel still lying around from years ago for comparisons.
The sanded surface will have a soft and fluffy feel ,and sounds a little soft too, tech ingredients seem to like this and say it acts as a sound absorber?
Just thought you might like to experience this first ?
Are you thinking of using the panel full range, or using a low frequency driver for support ?
Which exciter model are you using?
A fairly fine sand paper should be OK, you don't want to go too deep, it's more like scuffing the surface to get that fluffy feel.
The very thinned pva will just soak into the fluffy surface ,producing a better sounding hard skin, pva will take a couple of days to fully harden for full HF performance.
I use the nxt recomended 4/9 x 3/7 exciter positions myself ,for a more central position .
Steve.
- Home
- Loudspeakers
- Full Range
- A Study of DMLs as a Full Range Speaker