A Study of DMLs as a Full Range Speaker

https://www.freecad.org/

Unlimited. Amazing features. Open Source. Free!

From the Features page:
"FreeCAD is made for everybody, by everybody. It is developed and maintained by a community of developers, users, moderators, translators, all united by their wish make FreeCAD a free and powerful tool. There is no commercial aim behind decisions being taken, no urge to make you upgrade your version of FreeCAD or to corner you into a specific workflow or ecosystem. FreeCAD and the files and data you produce with FreeCAD are truly yours, forever.
 
The panels sound fantastic with most genres of music but some hiphop sounds a little strange, not sure if it's male vocals getting a little echo/reverb from inside the panel itself. I have them crossed to a front loaded horn subwoofer at 130hz. Anyone else noticed spoken things sounding strange?
Yes
I had this in my first Cedar panels. I used news reports to judge spoken audio quality... Started off sounding hollow/honky..I used thin felt dampening to overcome it. Trial and error process for me at the time
Eucy
 
Last edited:
  • Like
Reactions: 1 users
I have also noticed increased HF off axis. For a HiFi application it is of course not ideal with extra HF content to the sides adding reflections. You basically have one listening position and want balanced sound emitted in that direction.
My untreated sandwich paper panels also have this very pronounced high peak 90 degrees to the side (almost all the other sound disappear like a true figure of 8). Some of the panels I have filled the edge of the panel with flexible pu glue, and they don't exhibit this peak. I think the outer edges may be resonating wildly at some high frequencies.
 
Hi Christian,

My measurements were pretty quick and I didn't pay a lot of attention to the set up, I was pleasantly surprised with the FR so posted. I think the reflection could be because the mic was on the edge of a set of drawers about 1m high, the panel was 1m away, 1m from the floor the wall behind it was probably around 1.2m behind the panel so maybe some of these reflections caused the spike.

The panels sound fantastic with most genres of music but some hiphop sounds a little strange, not sure if it's male vocals getting a little echo/reverb from inside the panel itself. I have them crossed to a front loaded horn subwoofer at 130hz. Anyone else noticed spoken things sounding strange?
Hello Joska
I am probably also too enthusiastic and prone to post measurements... most of the time with not so much time for doing them... don't worry. For 1m height and distance, the delay of the reflections is around 4ms (a bit less). For a back wall at 1.2m it is a bit more than 7.2ms. I think both of them are visible in your measurement. Those reflections happen in I think all the set up except specific arrangement (like outdoor?) to reduce them.
Anyway, a HF spectral content at some angle seems here.
Christian
PS : about the voices a bit strange, you might try to see if you can reduce the distortion in the mid range. Before having a look to the distortion in the panel I have in test, it was also male voice that was strange.
 
  • Like
Reactions: 1 user
I experimented a bit with angled baffle wings today. My frame has 6" wings parallel to the panel all around. I placed some extra boards left and right of my high aspect ratio panel and I tried different angles, sorta like a horn, facing the listener. I noticed that at any angle, and even when flat because of the extra thickness, the baffles focused the sound some. With no angled baffles everything sounds very open and diffuse and that does sometimes sound a little odd with vocals to me. Adding these baffle wings and angling them has a significant impact. The higher the angle the more focused the sound stage seems to be. I did a few FR measurements but there weren't any huge changes there that I could identify by graph alone because a few dB difference here and there is not very significant on these charts. The difference in soundstage and vocals was audible though.

I don't know how I feel about using angled baffles but I did take them off for a listening session after I was done playing around. This might help you if you think vocals sound weird - try putting some angled baffle wings on and see how it sounds.
I have in mind a post from DMLBES (?tbc) who said the sound of a DML can be changed depending how far it is in a H frame. This makes sense with some HF at high angle which is then reflected/diffracted by the frame or the wings.
If you have a look to the FR with for example a 500ms, it might not be so visible.
A spectrogram showing the SPL versus frequency and time should be more informative. The HF content reflected by the frame should arrive sooner.
Christian
 
  • Like
Reactions: 1 user
Yes
I had this in my first Cedar panels. I used news reports to judge spoken audio quality... Started off sounding hollow/honky..I used thin felt dampening to overcome it. Trial and error process for me at the time
Eucy
Hello Eucy
See the possible link with distortion and some unexpected resonances in previous posts. When the distortion frequencies are pointed, the 1/3 octave noise can help.
Christian
 
  • Like
Reactions: 1 user
Hello
I haven't found the weight of the EX32EP2-4 to extract your panel areal mass. Do you have it? The spec I have found as the one of the tweeter is weak.
Other question :
  • what leads you to add a tweeter (sorry if it is in one of your previous post)?
  • its cut off frequency is 6.2kHz?
Christian
Hi there,
I've just been through my quite extensive Exciter spec. sheet and strangely the weight is not specified.
I wish I had weighed my panels before attaching my TWO exciters per panel. Sorry I can't help more.

I don't really understand your first mention of "tweeter is weak" ?
https://www.pioneerelectronics.com/PUSA/Car/Speakers/PRO+Series/TS-B350PRO

I decided to add a tweeter (actually Super Tweeter) because the DML panels sounded & measured > DULL.
The Pioneer spec. says 3.5Khz to 27Khz. With my crossover, I hear it as roughly 5Khz upwards.
Even although I had to substantially drop its output (volume), I had a curiosity to use a 'bullet tweeter'.
It's not too bad, but I can hear it 'SING' a little bit around 18Khz. I would probably recommend the great value Vifa D25AG.

Cheers :)
 
I don't really understand your first mention of "tweeter is weak" ?
https://www.pioneerelectronics.com/PUSA/Car/Speakers/PRO+Series/TS-B350PRO
Hello
Sorry for the too short or not adapted wording. I found the same page which doesn't answer to what I expect from the specification pages of a loudspeaker : no curves showing the frequency response at different angles (the most important), the impedance.
Some Dayton audio specifications about exciters suffer also from a lack of data.
When I first looked at this page, I didn't remark the capacitor delivered with the tweeter.
Christian
 
@Veleric

Hello Eric, it's been a long time!
Just to share observations from last week.
When I commented the measurements from Joskaz, I was surprised by the level of the 1st reflection :
View attachment 1292638
I was even more surprised when I made just after some measurements.
Here is my panel with a polypropylene membrane in association with a tweeter (a FT17H, so a bit directive)
View attachment 1292644
Almost no reflection... and now an EPS panel at exactly the same position with the same mic position
View attachment 1292646
There is a stronger reflection...
Which in my opinion shows the DML throws some high frequency energy at a large angle... which is expected by the scientific papers about DML (important lobbying at high angle at the coincidence frequency)

After that, I updated my REW to last 5.31 version. From 5.30, REW offers more variant of wavelet spectrogram. So I come again to the idea to check the content of the response around 0s but now also at the time of the first reflection and it appears that just after 0s, there is a hole in the HF with a peak (rather flat) at the same frequency at the time of the reflection. Here is the example of my EPS panel.
View attachment 1292649
View attachment 1292651
To be verified if it is a plausible value...
Even if it is not the coincidence frequency, it is interesting to have a better look to the content of this reflection... I read somewhere that DML are not fully suitable (even if it work not so badly in my opinion) for an accurate stereo imaging because of this particular increase of the energy in HF at high angle.
Christian
Hello Christian.
So , how long does it take for the reverberation to stop altogether (flat line).

Panels do have an internal sound, this is part of the panel self noise I go on about.
I showed how to change this sound and significantly remove it in my finger positions recording.
Steve.
 
  • Like
Reactions: 1 user
XPS with PVA

I really want PS foam to work well for me in a DML. In my former job, I made PS foam! The output of PS foam (XPS or EPS) is impressive, and far higher than any other panel material I have tried so far. But, so far, I have not been able to make it sound good. It's LOUD, for sure, but not good.
Like many of us, inspired by the Parts Express post, I tried PS foam in my first DML attempt. I was impressed by how loud it was. But I thought it sounded "muddy". When I played it for my wife, she called it "muffled." We agreed, it sounded "unfocused". Soon thereafter, I tried plywood. I was disappointed by how much less sound volume plywood produced, but loved how much crisper and clearer the music, and especially vocals, sounded.
I've spent a good part of the last six months trying to find something as loud as PS and as clear as plywood. Sorry to say, I haven't found it yet.
I have to admit, it took me a while before I tried the PVA coating technique for PS foam. I was skeptical. But I finally did. Spoiler alert: for me it doesn't hurt, but doesn't help much either, as far as I can tell.
The results below are for 1/2 inch thick pink XPS foam. The foam I bought came with thin plastic films applied to both front and back. I tried that in my usual 16"x23" size, attached to a frame by a soft, thick PVC insulation foam around the full perimeter.


As I had observed previously for PS foams, it was loud, but it faded badly above 2 khz. It also sound muddy. Sadly, I can't measure muddy, but I can hear it.
I peeled of the film and sanded it, then remeasured:



Removing the plastic skin and sanding resulted in a dip output between 100 and 200 Hz. Still faded above 2 kHz. Still sounded muddy.
I applied 2 coats of PVA and measured this:

The dip between 100 and 200 Hz improved. Still faded above 2 kHz. Still sounded muddy.
I applied 2 more coats of PVA and measured this:

Output now peaked at 200 Hz. Still faded above 2 kHz. Still sounded muddy.
My conclusion is that applying the PVA coating made a slight improvement, basically restoring the performance provided by the pre applied film. But it did nothing to improve the major deficiencies of PS foam, that is, the muddy sound and the fade above 2 kHz. YMMV.
Eric
Hello Eric were the PVA coats 50/50 PVA/water or were they full strength? did you mount the exciter directoy onto the foam board or did you have any sort of interface material between the panel and the driver? thanks.
 
Veleric.
I agree with you about foam core ,it is heavily damped by the paper .
Eps on on the other hand has no problem reaching 10k (1inch 70 grade) or 20k and beyond with the vh grade.
Steve do you recall the density of VH grade XPS/ I had not realized that XPS was available in thin flooring under lay sizes, I will see if I can find some. Do you still consider this to be a superior panel material? I know that you have also used thin light grade white XPS with very good results. I this a situation where either can be made to perform well with equal high end extension? thanks.
 
do you think that making the wings from an absorbent material such as high density fibreglass would help? I am thinking about the fibreglass panels used for drop ceilings which is can be found in thickness of 1/2" and thicker. thanks.
I'm still playing with baffles here and there so I can't say with certainty how they affect the sound. There are a lot of variables across my tests and results have been 'mixed' because I've tested so many things without being methodical enough to draw absolutely correction conclusions.

Right now I'm pretty sure that adding a 6" wood baffle all around my high aspect ratio panels boosts LF and evens it out some. A gap of 1" between baffle wings and panel is too big to show much impact, it needs to be as tight as possible. Under 1/4" for sure. This rule may not be true for lower aspect ratio panels. I use higher aspect ratio because I believe that a narrower panel has reduced modal intensities for a more even FR. This might be more dependent on panel rigidity and damping though, and my setups may have changed those variables at the same time that I changed panel dimensions.

The baffles need to resist the pressure differential between front and back so as long as they are not floppy they should work. Even normal cardboard box shows results in test, but the stiffer the better, I think. The surface will affect the higher frequency reflections which might be important for DML, idk, I haven't tested.
 
  • Thank You
  • Like
Reactions: 1 users
Guys, I don't want to lecture too often. But...
I sometimes see comments regarding the "sound" of higher frequencies in the plus-10khz-or-so region.
I've even seen a comment somewhere about this range, an excess of which, is apparently the cause of sibilance in speech, or where cymbals and hi-hats are most active.
Two things pertain:
1. Any male over the age of 40 or so is unlikely to hear much information in that range. It's reserved for teenagers and maybe females up to the age of about 30 ot so.
Hi-fi pundits, being older males in general, will have a neg-3dB point between 10 and 15khz.
2. Sibilance generally occurs at about 5-7kz for males and about 6-8khz for females. That's where recording studios set their de-essers. Hi-hats about 4-7khz; overall brightness about 8khz, and "air" and "breath" in the sound around 10-12khz. There's no real, useful sonic information above that.

The above information might help you to target deficiencies in your designs purely by ear.
If all else fails, then fire up your REW signal generator with variable sine waves or narrow-band white noise, activate the RTA, and test for yourself what those HF frequencies actually sound like.
 

Attachments

  • full_1022x1022.jpg
    full_1022x1022.jpg
    226.3 KB · Views: 38
Last edited:
  • Like
Reactions: 2 users
Hello Christian.
So , how long does it take for the reverberation to stop altogether (flat line).

Panels do have an internal sound, this is part of the panel self noise I go on about.
I showed how to change this sound and significantly remove it in my finger positions recording.
Steve.
I learned something about reverberation yesterday through experience. I always knew that panel damping is important, that less 'ringing' in panels is a goal, but I didn't understand how it sounds.

I was testing a treated 1/4" XPS panel and it sounded like my tests sounds were playing in a long tube. This effect was quite dramatic; in a way it sounded like this thin panel was a huge room. Then I realized that a lot of panels, EPS and XPS in particular, sound this way through a section of FR, especially around 300-600 hz. Panel material, thickness, and surface treatment have a big effect on this and it can be easy to mistake this reverb for modal response if you are only paying attention to volume and not timing. Panel mounting methods also have a big effect and changing panel dimensions also changes the damping effectiveness of the mounting.

Thanks to Christian's hint I was looking at 3rd order distortion yesterday and it does seem like there is a good correlation with high 3rd order distortion and reverb in the panel. Idk if high 3rd ONLY comes from reverb inside panel, or if ALL reverb in panel reflects as 3rd order distortion, but there does seem to be a correlation.

So, if vocals sounds strange and if it sounds like a reverb effect, then panel 'ringing' might be the cause.

Oh, and here's a cool video I watched about reverb last year. It's a very different use of an exciter and a large panel:
 
  • Like
Reactions: 2 users
I was looking at 3rd order distortion yesterday
The spectrogram view using wavelet or CWT is also an excellent tool to get information about the level in frequency AND time.
XPS was the first material I tested before coming to plywood. I remember from it an important energy stored in the range you mention. For the plywood I don't know The current small EPS panel for testing at the moment shows level in this frequency range at about 10ms.
All of that doesn't give the recipe to built a good panel for sure!