Carpet underlay as 703 substitute, bass absorbtion

[took-the-red-pill]

Hi all,

I searched but couldn't find this specific question asked.

To help control lows, I was hoping to make a significant square footage of 703 panels, at 4"-6" thick, and place them over all the corners I can. (I have QRD and absorber panels in middle of walls and ceiling) However, either the supplier hasn't heard of 703/705, or wants me to bring in a huge amount of it.

I've had an idea: I can get carpet underlay by the giant roll, but it comes in at 5-8 lb/ft3.. I would spray glue or contact cement enough layers together to get a 3-4" thick panel.

Would the composition of the underlay(foam chunks or fibre) have an effect on its ability to turn sound energy to heat, or is it just basically density?

Would a panel made in this way, in your opinion, do just fine a job of absorbing low end? Or do I need to hunt down some 703?

Thank you
Keith

[Gregwor]

What really matters is gas flow resistivity. I'm not sure what characteristics underlay has acoustically, but I know it is very very heavy and I will guarantee that it is horrible for dampening bass frequencies.

Greg

[JasonFoi]

Heh, exact same advice he recieved at gearsluts

[Soundman2020]

Yup. Carpet and carpet underlay are not much use for low frequencies. The only data I have on underlay shows that it is reasonable for absorbing highs, lousy for lows. As Greg said, unless you have the actual GFR numbers for it, or can get an acoustic lab to run the necessary tests for you, then I would not consider it.

- Stuart -

[took-the-red-pill]

Hi,

I erroneously thought it was the density that mattered. In my reading I'd never come across GFR

Incidentally, Jason above included a link to a GFR calculator, so if it's helpful for anyone:

http://www.acousticmodelling.com/porous.php

Thanks
Keith

[Soundman2020]

I erroneously thought it was the density that mattered. In my reading I'd never come across GFR

Right. In fact, GFR is the measurement of acoustic impedance, which is what we are most interested in. Here's an actual definition that I have from some place that I don't recall:

"It is defined as the ratio of the pressure difference across a sample to velocity of flow of air through that sample."

That same source goes on to explain why density isn't really what you need to know:

"The flow resistivity depends on the porosity of a material as well as its tortuosity, but for high porosity, low tortuosity fibrous materials, the flow resistivity is approximately inversely proportional to fiber radius squared at a constant bulk density i.e., a large number of small fiber diameters results in a higher flow resistance than does a small number of larger fibers."

And here's something I wrote a while back, to help explain some of the intricacies:

"Two things happen when sound strikes a porous absorber. First, there's a sudden change in impedance as it hits the surface, so part of the wave is reflected back, and part carries on (plus part could be refracted too, and even diffracted, but let's leave those aside for now!). That initial impedance change varies according to the porosity of the material: less porous (lower density) means less impedance, means that the wave is less reflected. Lower porosity (higher density) means that more of the wave is reflected. And since this is impedance, the effect varies with frequency. So some portion of the sound spectrum is reflected right at the surface, and some portion carries on into the material. (To confuse matters even more, this also depends on the angle that the sound strikes the surface...).

GFR doesn't tell you much about the part that bounces back, but it does tell you a lot about the part that didn't.

The part of the sound that is carries on into the absorber, runs into resistance as it moves forward, because there are numerous "barriers" (fibers) in its path. GFR measures this, as the relationship between pressure and flow. In other words: what amount of gas pressure causes what amount of gas flow. Since sound waves are basically made of of air pressure "vibrations", and air is made up of a bunch of gasses, this makes perfect sense. (After all, the way we measure sound intensity is by measuring how much the air pressure changes as sound waves go by, and we do this with an SPL meter: Sound Pressure Meter. The reading on the meter shows dB SPL)

In essence, a sound wave is just gas moving backwards and forwards through the maze of fibers inside the insulation material. Each wave is a pressure change that causes gas to flow, and the GFR number just tells you about how that works inside the material: what amount of sound pressure causes what amount of air to flow, and how that differs from the way it would be if there were no insulation, only air. Bigger numbers mean that the materials resits the flow of gas (= passage of sound) more than smaller numbers.

Unfortunately, most manufacturers don't bother measuring or publishing the GFR numbers for their insulation products, for a simple reason: GFR is not a useful number for talking about thermal isolation, and those guys sell the vast majority of their products for thermal applications. The fact that the same products also work for acoustic absorption is sort of a luck coincidence, but its just a very small fraction of the total market, so it's not worth it for them to measure and publish.

Fortunately, there is a rough (very rough, and not linear) relationship between the density of each type of material, and its gas flow resistivity."

included a link to a GFR calculator, so if it's helpful for anyone:

It's helpful to a certain extent, and I use that quite a lot when designing acoustic treatment, but you do have to be careful when using a simple calculator to predict complex phenomena. That calculator makes a lot of assumptions about the absorber, that might or might not be true. They way you linked to it, for example, is for normally incident sound, which is not typically what you are interested in a studio: you have to click on the box that says "random incidence" to get a slightly more useful prediction, but even then it falls short. For example, it totally fails to predict the real characteristics of plain old OC-703, or its siblings, OC-701 and OC-705. Here are the actual measured coefficients for the 70x products:

OC-703-specs.jpg

Try plugging in any GFR you feel like to that calculator for each of those thicknesses, and see if you can get it to predict the actual results... For example, the highlighted line in that table is for 4" of 703 up against the wall, and when tested in an acoustic lab it showed a coefficient of 0.84 at 125 Hz.... There's no way you can get that calculator to produce this result. Because the assumptions made by the calculator are not correct for OC-70x.

That does not mean that the calculator is no use! Far from it! It is still a very useful tool, provided that you understand its limitations and do not assume that will always predict correctly. That's why it is so important to check the actual specifications for any product you plan to use in your studio, to make sure that reality matches theory... and if not, then you need to modify the theory until it does match, not try to change reality! (as some folks love to do...)

The acoustic rabbit-hole goes deeper than you's ever think....

- Stuart -

[JasonFoi]

Hmmmmm... this intrigues me, frustrates me, excites me and makes me happy at the same time. If you go over to GS land and start saying to model anything in the PAC at random incidence youre gonna get a huge backlash of bull poop thrown at you stating that random incidence doesnt apply to small room acoustics and that we're only interested in normal incidence to about 60 degrees. Well, maybe not you, but certainly me!! Ive had it shoved down my throat that random incidence is from reverberation chambers and only useful for larger rooms. BUT the biggest source of frustration with that for me stems back to a thread i read a long time ago where Eric Desart said random incidence WAS useful in small room acoustics because you get a bazillion reflections from every which way in a small rooms. I really wish everyone would get on the same damn page about everything. It would make learning a lot easier.

[Gregwor]

Jason,

Whoever says you shouldn't consider random incidence is crazy. Of course sound bounces around the room (hence modes) at which point it is inevitable that it's going to hit the insulation at every angle possible. Maybe they're flat earth believers too.

Greg

[JasonFoi]

The person who told me that random incidence doesnt apply to small rooms is Jens Elkund, and i hold his advice in extremely high regard. Hence my frustration when Stuart, whose input i greatly value as well, doesnt agree.

[Soundman2020]

Think of it this way: the sound arriving at any given surface in a room certainly is not normally incident: that would only be true for a few small areas that happen to be lined up with the speakers, face-on. And for low frequencies, it isn't randomly incident either! Because small rooms do not have statistical reverberant fields... at low frequencies. (But they do at high frequencies...) On the other hand (or maybe "on the other, other, OTHER hand....) it is impossible to get beyond the critical distance for most home studios: the critical distance is further away than the walls. So from that point of view, there's no random incidence either... but there's also no normal incidence! Thus: There's no diffuse reverberant field (and thus no random incidence), but the sound arriving at most surfaces is not normally incident either. To be strictly technically correct, there's no true random diffuse field in ANY room, even very large ones! But that's a different story...)

Confusing.

So what do you do? Both random incidence and normal incidence are not valid.... But which one is closer? Think about: Normal incidence is a specific special case of sound waves arriving perfectly perpendicular to a surface. Random incidence is sound arriving from "all other angles at once, with equal probability". So which makes more sense to use in a home studio?

For the majority of surfaces in a room where porous absorption might be applied, and for the majority of the spectrum, the incidence is not normal. It isn't truly random either, as I pointed out above, but it's closer to being random than it is to being normal, because normal is one specific special case, and most of the sound arriving does not comply with that specific special case. Hence, real results from real rooms do not show what you'd expect from normal incidence, but they do show something approaching what you'd expect from random incidence. So, you are unlikely to see a real 0.84 coefficient for 703 in a real room, as shown on that chart, since that really was measured in a room with pseudo-random incidence, but the coefficient certainly will be higher than if you tested the same sample in an impedance tube, and thus really did attain perfect normal incidence, because real rooms are nothing like impedance tubes.

As Andre has pointed out repeatedyl: porous absorbers provide useful absorption where the thickness of the absorber is around 7% of the wavelength for normal incident sound, or around 3.5% for randomly incident sound. Real data from real rooms shows this to be correct, and you do, indeed, find plenty of cases where thicknesses of MUCH less than 7% of the wavelength shows considerable absorption. If the sound in the room were only ever normally incident, then there should NOT be any effect below 7%, .... but in reality there is some effect. Thus, the actual absorption is closer to that of randomly incident sound, than it is to normally incident sound.

I'm not arguing with Jens: what he says is technically correct: the sound field in a typical small studio is not randomly incident, and there is no true statistical reverberant field nor diffuse field in a small room: true! But reality also shows that the REAL sound field is very, very different from what it would be if it were only ever normally incident.

Conclusion; if you want to get realistic predictions, then use randomly incident settings, and assume it will be in the general ball-park but it won't be totally accurate.... Or use normally incident settings, and assume that it will BETTER than that. Neither is correct, but random incidence is more correct than purely normal incidence.


- Stuart -

[JasonFoi]

Now THAT is a great explination. Thank you Stuart. I was reading through Jen's posts on the subject, most of which are very short, and i found one where he went into a little more detail. He's refering to treating the modal region and how modes will line up into their neat patterns. For this we should look at treatment placement and selection of materials based upon their behavior at normal incidence. The increased absorbtion across the spectrum from random incidence us an added bonus, but what matters is their behavior at normal incidence for targeting modes.