Preserve high frequencies with plastic plates

[andymel]

Hi,

I'm currently helping someone improving his room acoustics. He is deep into Hifi and does not want to use an eq in his signal path. OK so I try to give tips that preserve as much high frequency energy as possible. My question: do you have any experience using thin plastic plates to completely cover absorbers? A foil can't be used for areas that size because of mechanical and optical demands. The idea is to have the least possible weight that reflects high frequencies and needs no other cover to protect it (that would absorb high frequencies again).

What we already did:
I made a measurement and told him to build edge absorbers (mineral wool) for a start. We measured again and the result is as expected. But now we need to cover them with something. In addition the whole rest of the backwall (~12m^2) will be covered with 20cm mineral wool (plus 20cm air gap) and also needs coverage. I try to get him to a RT60 of 0,4sec. According to my calculations we will need even more than that as well.

Actual RT60:

20160304_before and after edge absorbers.png

As you can see on the pic above we already reach the 0,4sec at 10k. I think it would be best to have a cover that reflects above ~6k Hz. There is this "Bergschersches Massengesetz" (I don't know the english word, should be something like "Mass-law of Berger") that is used to calculate the soundinsulation of walls. I don't know if that is accurate for that light-weight plates but if yes I would need 0,05 kg/m^2...that sound more like a foil than a plate.
Long story short, I don't have any real experience on that. That's all just numbers out of books.

So my main question I guess is...what would you try to do? Any thoughts on that?

Thanks
Andy

It's maybe not necessary to know for a general answer on how to preserve high frequencies but maybe it's good to know the context. This pic shows what I want to achieve. It's the equivalent absorption area. The gray lines mark my target area, the blue one is the actual state (with those edge absorbers) and the red lines mark the difference.
An "Equivalent absorption area" of 10m2 means I need 10m2 of an absorber with absorption factor 1. But I can also use 20m2 of absorbers with absorption factor 0.5. It's simply the product of the area and the alpha of the absorber.

Äquivalente Abs_fläche mit Kantenabsorbern.PNG

[Soundman2020]

Hi Andy! Good to see you around! You still have the best acoustic calculators out there!!! I'm your number one fan for those...

Please post t actual MDAT file somewhere that I can download it, to analyze it in more detail. Also, what are the dimensions of the room?

It seems to me that you have other issues in there as well, and will likely need some absorption around 2k, running to an octave either side of that.

that is used to calculate the soundinsulation of walls.

Are you talking about the Mass Law equation:

TL = 20 log (F * M) - 47 dB

Where:
F is the frequency (Hz),
M is the mass per unit area (kg/m²)
TL is Transmission Loss

That's not the equation you need here.

There's three ways to look at this that I can think of off hand (at least three ways). One is with a "foil", as you mentioned, which can indeed also be a thin membrane. I don't recall where I got it, but I do have this equation on file, and it seems to work reasonably well:

M = 90/F

F = Frequency that passes through the foil at 80% transmission (20% loss = 6 dB loss. "loss" = reflection here). (Higher frequencies are reflected progressively more)
M = Mass (surface density) of the foil in kg/m2

So for example to reflect above 6 KHz. you would need a foil foil with 15g/m2 (0.015kg/m2) is needed.

But since you want good reflection above 6K, and this is only the 20% point, you would probably want to shoot for a higher density (lower frequency). Maybe set 1.5k as your 20% reflection point (2 octaves down from your 6k need), so you'd want a foil (membrane) with a surface density of 60g/m2 (0.06kg/m2).

I'm not sure what the roll-off curves are above and below that "80%" point, but I'm assuming it is related to mass law, at 6 dB/octave. If I'm correct, then you'd have 18 dB loss above 6 kHz in transmission through the foil, and most of that would be reflected. (some would be lost in absorption by the membrane itself, and in other ways, but not much).

That's the first issue. The second issue is that you could use reflective surfaces that are sized according to wavelength. Since sound pretty much ignores objects in its path that are smaller than the wavelength, you could use square patches of wood, plastic, glass or whatever that are about 2-1/4" across, or 57mm, which is the wavelength of a 6Khz tone. Larger Wavelengths (lower frequency) will pass by to be absorbed by whatever is behind, while smaller waves (high frequencies) will be reflected back. You could do widely spaced slats that are 57mm wide, for example.

The third issue is to use some type of very high density porous absorber, cut very thin, with an air gap behind it, such that it is tuned to the frequencies you need to absorb, but rolls off above and below. I played around for a couple of minutes with a porous absorber calculator, and found this combination:

tuned-porous-absorption.jpg

That's tuned to 2 kHz, where you do need it, but rolls off above that. Put the foil or slats in front of that, and you have a combination trap that absorbs what you need in the mid range but rejects higher frequencies. You'd need to find insulation that has a GFR of 60,000 MKS rayls for that, and you'd need to figure out how to cut it very thin (only 5mm) and space it 30mm away from a wall, but in theory it should work.

Or do all three...

.that sound more like a foil than a plate.

Right! Foil is what you need. For even a rather light-weight plate at say only 1kg/m2, you'd be reflecting from 90 Hz upwards.

But don't take that "foil" equation as written in stone! I can't vouch for the original source, and the transmission / reflection isn't going to be perfect in any case. However, it does seem to be a good starting point to play around.



- Stuart -

[andymel]

Hi Stuart,

thanks a lot for your fast and laborious answer. I guess now we are each others fan It's so sad I don't have more time, my head explodes because of all the calculators and other ideas I want to build...but I am working a lot, so it will happen! I guess the next published thing will be an app. I count on you as my tester

Please post t actual MDAT file somewhere that I can download it, to analyze it in more detail. Also, what are the dimensions of the room?

The room size is 730x630x280 (cm) and the mdat file is here https://drive.google.com/file/d/0B ... sp=sharing

Are you talking about the Mass Law equation: TL = 20 log (F * M) - 47 dB

Yes that's exactly it. Seems like the resulting weight is similar to your easier 90/f formula. Thanks a lot for that easier way to approximate it!

You could do widely spaced slats that are 57mm wide, for example.

That was my first suggestion as well but he does not like the whole stripes-idea. Maybe I should try to force him a little bit more in that direction. What do you think should the space between the slats be at least?

The third idea is the craziest of your ideas I think because of all the work to do that would not be the first choice. This strong frequency dependent absorption coefficient could work for the first wave front as most of the energy should come from 0°. But after that I would guess that we would get a much more even alpha over frequency because of the random incident angles.

Backwall_20160304.png

sketchup file

The left is inspired by you ...57mm
The others would be easier to build though.

Waiting for your thoughts

[Soundman2020]

I count on you as my tester

For sure! I love your apps. I'm just waiting for the one that has a single button saying "Design The Complete Studio"....

The room size is 730x630x280 (cm) and the mdat file is

Great! I'll take a look at it, then get back to you.

That was my first suggestion as well but he does not like the whole stripes-idea.

The could be circles, squares, hexagons, octagons... pretty much any geometric shape that is the right size.. You could have a whole "surrealist modern geometric treatment montage unit" back there, if you wanted... They could even be made from thick transparent material, such as glass or acrylic panels.

What do you think should the space between the slats be at least?

A couple of cm would be fine. And don't seal the cavity behind! If it is sealed, then the slats would act as tuned Helmholtz resonators, but if the cavity is not sealed, then you should be fine.

The third idea is the craziest of your ideas I think because of all the work to do that would not be the first choice.

I love coming up with complicated solutions to simple problems! Rube Goldberg is my hero...

This strong frequency dependent absorption coefficient could work for the first wave front as most of the energy should come from 0°. But after that I would guess that we would get a much more even alpha over frequency because of the random incident angles.

At higher angles it doesn't change too much: the frequency curve moves over to the right a bit (higher frequency), and the coefficients of absorption drop down a bit, but the overall effect is still usable. Even for random incidence the curve is not all that much different: lower Q of course, and lower overall coefficient (maybe 0.7 or so) but it would still do something useful.... I think!

The left is inspired by you ...57mm

You could space them randomly too, if you want! You could even have several different sizes, to reflect different frequency ranges, if needed. Lots of flexibility.

The others would be easier to build though.

Yup. Just call me "Rube"...


- Stuart -

[Soundman2020]

OK, I took a quick look at that, and ... You got some serious issues going on there!

He is deep into Hifi and does not want to use an eq in his signal path.

Fair enough... but he also doesn't want this, and doesn't want that, and doesn't want something else... Maybe you should have a chat to him, and tell him that he is imposing unrealistic and arbitrary limitations on the results. There are limits to what the laws of physics can accomplish for a small room like that anyway, so if he places even more limitations out of pure subjective "likes" and "dislikes", then you won't be able to actually do much at all, except for interior decorating!

That room needs major bass trapping. There's a lot of stuff going in in the low end, and the high end isn't looking too bad. The mid-range "bulge" could be due to many things... hard to say.

In addition the whole rest of the backwall (~12m^2) will be covered with 20cm mineral wool (plus 20cm air gap)

Absolutely! The x.0.0 axial modes are very prominent, and I suspect they are masking other stuff that is going on in the other directions, that we can't even see yet.

I try to get him to a RT60 of 0,4sec.

For that size room, I would go for something dryer than that. It's not a big room: I'd shoot for something close to ITU / EBU specs, at around 270 to 300 ms. Is there any reason you wanted to go for something so live? If this is an audiophile room, then neutral would be better than live. With a decay times that are tool long, remind him that this will hide some f the subtle details in his music: he won't be able to hear the actual reverb tails on his music, as they will be masked by the reverb tails of the room itself.

As you can see on the pic above we already reach the 0,4sec at 10k.

Go to the "Filtered IR" tab in REW. The small overlay window on the left shows the estimated RT using different methods: Forget EDT, and look at T20 and T30. They are showing that you are much closer to 600ms than 400ms, overall. Look at where you are for 1kHz: That's the reference point for RT60. 10k isn't going to be a problem. You are not going to lose much more in the high end if you use the above strategies, but you do need to get a lot more absorption in the low end.

An "Equivalent absorption area" of 10m2 means I need 10m2 of an absorber with absorption factor 1. But I can also use 20m2 of absorbers with absorption factor 0.5.

Wwweeeellllll.... Hmmmmm... That sort of assumes that all absorbers are prefect absorbers, with flat absorption curves across the entire spectrum, from 20 Hz, to 20,000 Hz. Real-world absorbers don't behave like that. Alpha changes even according to the shape and installation of the absorber! And it can be above 1 for some installations, while also being below 1 for others.... You'll have to "tune" your absorption strategy... and you might even need a little diffusion in the high end, if this is an audiophile room.

Photos of the room as it is right now would help!


- Stuart -