absorption and doors

[Brian Dayton]

Recently, more or less for the heck of it, while running some tests at Orfields lab this experiment was conducted:

A frame of 25 gauge steel studs was in the opening, with 2 sheets of 5/8" drywall (damped) installed on one side only - a simple panel that should approximately follow mass law.

Then we had R13 insulation installed in first 1/3, then 2/3, then all of the spaces between the studs.  the results were very interesting and are presented below in the first diagram.  What was observed was a marked improvement in middle and high frequency TL when the absorption covered the entire specimen.  

The results appear to follow the absorption of the insulation itself (R13), and i would guess that if you took an impedance tube measurement of absorption of R13 fiberglass, you would find a strong correlation between this improvement in TL and the results of the impedance tube absorption test.

I then dug through the NRC's published tests to find a historical precedent, and located one.  The TL of a floor/ceiling with and without carpet is shown.  The same phenomenon is observable - marked improvement in TL at those frequencies where the carpet is a good absorber.


So it is established that absorption can improve TL at mid and high frequencies (or possibly lower freq's if the absorption in question was very effective at those freqs).  But while it isn't cost-effective (especially at low-frequencies), or probably practical from a sound-quality standpoitn to cover your entire wall in absorbing material... it isn't that expensive or that impractical to cover your entire door.

And with a door you have this:

-generally a solid core single-door is quite heavy and
-generally doors found on studios or home theaters are quite small

and both of those will work to have the low-frequency performance of the door be generally pretty good.  But a strong coincidence dip is generally found in doors such as solid wood and other heavy doors, and the lack of an air cavity/absorption/decoupling, etc. in doors will generally keep their mid and high frequency TL well below that of the walls they serve.

Therefore i hypothesize (and hope to test oneday) the following:

-when planning acoustic treatment for your studio you should plan to cover the entire door and frame with absorption.  This will accomplish the following for you:

1.  improve the TL of the door at the frequencies where it needs help the most
2.  help (especially if absorption is present on both the frame and the door, making a small "tunnel" for sound passing through the seals of the door to pass through) mitigate less than perfect seal conditions.


Hope that's helpful, it really should be.

[Brian Dayton]

following up on this basic concept, i take some data from http://irc.nrc-cnrc.gc.ca/pubs/bsi/85-3_e.html

for a solid core door with seals.

Taking a leap of faith (it will ALWAYS be critical to ensure that your doors are well sealed) that the seals aren't the limiting factor, and an educated guess as to the effect of covering this door in different tyeps of insulation, we'd get these estimates.

raw door:  STC=26, tennekes=24, flat nosie = 26
with R13 = STC=34, tennekes=27, flat nosie=34

presumably superior absorbing materials could help more.  USG's thermafiber comes to mind.  or 703 or whatever.



So, for a single panel door where middle and high freq's are likely to be weaker, relative to the surrounding walls than low freqs, I think this should prove a valuable tool.

It is thus recommended that your entire doors be covered in absorption.

[Scott R. Foster]

Brian:

That is very interesting work.  Thanks for sharing.

What do you reckon causes the large step up in diminution of transmission once the entire panel is covered... suppression of bending waves in the panel by attenuating the impact of the pressure wave all the way across the panel?  Somehow leaving a third of the panel exposed greatly reduces effectiveness, and it would be interesting to know you speculations on why this arises.

Also, the installation you propose for doors is reminiscent of a the "splitter silencer" effect Eric suggested many times be incorporated in TL doors - for example, in lieu of a "sweep" at the bottom edge of the door or other element of door seal that is likely to quickly wear out.  The concept as I understood his explanations was that a pair of absorptive surfaces with a "crack" between them can behave almost as though there is no crack.

[Brian Dayton]

Brian:

That is very interesting work.  Thanks for sharing.

What do you reckon causes the large step up in diminution of transmission once the entire panel is covered... suppression of bending waves in the panel by attenuating the impact of the pressure wave all the way across the panel?  Somehow leaving a third of the panel exposed greatly reduces effectiveness, and it would be interesting to know you speculations on why this arises.

Also, the installation you propose for doors is reminiscent of a the "splitter silencer" effect Eric suggested many times be incorporated in TL doors - for example, in lieu of a "sweep" at the bottom edge of the door or other element of door seal that is likely to quickly wear out.  The concept as I understood his explanations was that a pair of absorptive surfaces with a "crack" between them can behave almost as though there is no crack.

Scott,

   That's a good question that i've pondered quite a bit dating back to the panel-trap triple leaf experiment, which showed a similar phenomenon (i posted that a while ago).

    I think that the answer can be understood in two parts:

If we imagine each of the 6 spaces between the studs in the experiment above as an individual loudspeaker making 70dB of noise (bear with me), and assume that adding insulation in front of any given stud cavity would reduce the noise radiated by that "loudspeaker" by 10dB, we'd get tihs:

all speakers radiating at full (70dB) level:  total nosie=78dB
4 speakers radiating full volume, 2 speakers radiating at lower volume:  total noise = 76dB
2 speakers full volume, 4 speakers lower volume = 74 dB
all speakers lower volume = 68 dB

   So that calculation of just adding each section of the wall as a random noise source implies that the biggest gain will be had when ALL of the loudspeakers are turned down, similar to when all of the partition is covered in absorption.

 
And, second, at and around the coincidence dip:  note in the graph i posted above taht in the vicinity of the coincidence dip (around 2000hz) the whole-panel-covered effect is considerably MORE pronounced.   This probably/possibly relates to the nature of coincidence and if you expose any of the wall to open sound, rather than covering it with absorption, the coincidence can and will spread and radiate from all of the wall on the other side.


We didn't run the test "reversed", and in the case of the tset the absorption was on the "noise" side.

[Scott R. Foster]

This probably/possibly relates to the nature of coincidence and if you expose any of the wall to open sound, rather than covering it with absorption, the coincidence can and will spread and radiate from all of the wall on the other side.

Very intersting indeed.

that was the idea I was groping for with this:

suppression of bending waves in the panel by attenuating the impact of the pressure wave all the way across the panel?

Thanks for the clarification... the "coincidence region" is sort of like hitting a railroad rail with a hammer... if you let it take a smack it will ring for miles up and down the line.  A small part of why railroad rails make such bad pillows.  

[J.F.Oros]

This is verry interesting data, Brian, thanks for bring it and sharing it. And we are glad to have you back here  :)

Now I know why they massively pad those big managerial office doors. I thought that it doesn't do much for soundproofing, but now I see there is a real (like 10 dB) bennefit on the voice frequency range, where it was afterall intended to be used.

[bert stoltenborg]

Hey Brian, cool!

Philip still didn't make that damped speaker cabinet (have to do it myself) but they did manufacture a speaker testing facility.
18mm mdf-GG-18mm mdf, 180 mm gap with wool, 18mm mdf-GG-18mm mdf, inner space floating on sylomer.
He says you can shout what you want inside, and even with the ear on the door you don't hear shit!
They kill speakers from 130 Hz up, at 140 dB.
They are a happy bunch of speaker developers.
 
 :D

[Bob]

Now I know why they massively pad those big managerial office doors.

The only ones I ever saw like that were in the old James Bond films.

[J.F.Oros]

I know. I probably still live in the Golden Age !  

(or should I say GoldFinger )

[bert stoltenborg]

I feel a urgent craving for watching Diamonds are forever.

[J.F.Oros]

The only Bond movie I have is "Golden Eye". The first Bond with Pierce Brosnan.
Eric Serra did a great soundtrack on this one, diffrent from all the others.
And there are also Issabela Scorupco and Famke Jensen ...  :mrgreen:

[bert stoltenborg]

Famke Jansen is Dutch.
Nevetheless I like her in X-men I.
In Golden Eye she sucked ass (hmm, not litterally).

[J.F.Oros]

Hey, this is a first :

A verry interesting and on-topic thread (where Bert and I were actually making some pertinent and on-topic observations) was hijacked by Bob !    8O

Cool !    :mrgreen:

[bert stoltenborg]

We'll spoil 'm all, Flav!

[Brian Dayton]

my new internet-secure computer won't let me post at john sayers forum as it claims that forum is trying to hack my computer, won't let me log into recording.org, and signs me out of web forums after 15 minutes or so and won't remember my password.  So i have to go next door to post at sayers, and i can't remember my recording.org password anyway.

I wonder if life wasn't better getting hacked, but having a convenient life?


my pleasure to share this info.  these same tests may  have some use with concrete slabs as well.  concrete generally doesn't need any low-freq help, but can at times need some high-freq help.  i wonder?

[Brian Dayton]

does anyone know of a door handle that sticks out a bit from the door, so as to be clear of the absorption like shown in the picture:

[Ido]

my new internet-secure computer won't let me post at john sayers forum as it claims that forum is trying to hack my computer, won't let me log into recording.org, and signs me out of web forums after 15 minutes or so and won't remember my password.  So i have to go next door to post at sayers, and i can't remember my recording.org password anyway.
....

good. we have you all to ourselves.
this is great stuff, thanks.
2 questions come to my mind:

how much absorbance (%) per frequency is needed for said behaviour, and how does it differ accordingly to lower %?
(how thick is this R 13 ?, it's 24 kg/m3, right?)

how about walls with higher TL, there should be some relationship versus said effect?
the better the TL per frequency, the less pronounced this absorbing effect should be?
I'm also thinking how these 2 points would relate to LF behaviour (where it's harder to reach 100% absorption).

[Bob]

R-13 is 3.5 inch thickness.
It's fiberglass batt insulation, aka fluffy fiberglass pink. It's low density.
I think it's somewhere between 0.5lb/ft^3 and 1 lb/ft^3, which is 10kg/m^3 to 14kg/m^3

Basically, Brian put 3.5" steel studs on the wall, and stuffed batt insulation between them that was made to fit between them.

Unfaced 3.5" R11 on wall 0.34 0.85 1.09 0.97 0.97 1.12 0.95 (125hz to 4000hz, NRC)

[Brian Dayton]

good. we have you all to ourselves.
this is great stuff, thanks.
2 questions come to my mind:

how much absorbance (%) per frequency is needed for said behaviour, and how does it differ accordingly to lower %?
(how thick is this R 13 ?, it's 24 kg/m3, right?)

OK, that's a great poitn and question and topic.  

If you are familiar with an impedance tube absorptiotn est, in that test you basically put some absorption material in a device and calculate absorption based on how much it attenuates sound ( i think).  These numbers from the impedance tube test cannot be greater than 1, and would always only approach 1 rather than reaching 1.

My anticipation is that absorption placed over a non-resonant panel (like a thin piece of metal or lead or something) would improve transmission loss by:

10*log(1-abs) where abs=impedance tube absorption (NOT normal reverb room absorption like on Bob's absorption page).

This implies that, like always, denser and thicker material will be effective to a lower frequency.

The sketch below compares the reference wall to R13 and R19 (3.5" and 5.5" insulation).  R19 wiorks to a lower frequency.  Presumably 4" of Thermaviber or Rockwool would work to a still lower frequency, and so forth.

how about walls with higher TL, there should be some relationship versus said effect?
the better the TL per frequency, the less pronounced this absorbing effect should be?
I'm also thinking how these 2 points would relate to LF behaviour (where it's harder to reach 100% absorption).

I think this has essentially no application for whole walls for these reasons:

1.  its not cost effective.  covering a whole wall in an appreciable amount of absorption would be awesomely expensive by the time you covered it and made it look good and all of that.  Cheaper ways would exist to get the results you wanted (but those ways aren't always applicable to doors, and doors are far smaller)

2.  Doors are usually weakest at middle/high freqs, and usually aren't weakest at low freqs.  Its alsmot a given that if your wall is a normal double leaf design that its weak poitn will be at lower freqs, where this technique becomes almost totally ineffective.  

3.  Eric Desart once warned about the sound-isolation troulbe that comes witha dding too much absorption toa  room.  His comment was on this forum, and related to dance clubs and discos and the like.  He said that you have a noise problem, which is probably at low frequencies, you add foam over all the walls and it cuts down on noise inside the club at middle/high freqs, which then causes people to turn the volume up louder, which results in more low-frequency nosiet han before (where the absorption isn't effective), which in turn results in even MORE problems.

So i think that the applicability of this is primarily to doors.  Windows you have to see through, walls are just a different type of animal...

but on a single-leaf door, this is just great and cheap, i think.

[Brian Dayton]

R-13 is 3.5 inch thickness.
It's fiberglass batt insulation, aka fluffy fiberglass pink. It's low density.
I think it's somewhere between 0.5lb/ft^3 and 1 lb/ft^3, which is 10kg/m^3 to 14kg/m^3

Basically, Brian put 3.5" steel studs on the wall, and stuffed batt insulation between them that was made to fit between them.

Unfaced 3.5" R11 on wall 0.34 0.85 1.09 0.97 0.97 1.12 0.95 (125hz to 4000hz, NRC)

Yep.

And i think that in general, when applying this technique, Bobs absorption page is useful, BUT, you have to note that with the equation i offered in the post right above this one you can't use Bobs absorption data,b ecause that is reverb room data.

But with some educated guesswork, you could get an idea of which is best.

thicker, denser, probably = better in this case.