The next step in noise isolation: bomb shelter!

Go find a 50's-early 60's home - there's got to be one with a bomb shelter somewhere

BillTodd -- Thanks for your comments.

The floating room-in-room cost was taken into account way back when the fund raising began. It's less costly than you'd think, materials-wise. The single most expensive aspect is the QuietRock drywall substitute I would use -- these are around $100 per 8x4' sheet, and a minimum of 13-14 sheets would be needed for a one-sided wall (with the unclad side lined with 4" thick blue fill and a 4" gap between that and the existing room wall. The additional blue fill would run around $500. Then there are the steel studs (again, not that pricey as not many are needed), another door, and the materials for the floating floor -- 6~9 sheets of ply or MDF, the neoprene dampers (which can't be that pricey), and perhaps a new carpet/underlay for the new floor. I think in total it might run another $4000. The biggest cost is in time / energy / wear & tear on this body -- even if I have help.

The low freq noise thing -- your question is one of the dilemmas we've been discussing all along. It's planes that create the most low freq interruptions, along with cars. They've intruded enough in the past to make measurement/recording frustrating if not impossible for entire days. It's a little less critical for the SPL measurements but much more odious for recording. We've been using a low filter (-12dB/oct starting at 80Hz) on the mic preamp all along for recordings, but even with this, any cars within a couple blocks or a plane that we can hear ends up in the recording -- which means we have to stop. This happens on a very regular basis.

The main question is do we have enough sound isolation now to be able to ignore most of these -- or at least enough of them? If we do, then there's no point chasing ultimate sound isolation -- it doesn't have to be absolutely impervious to outside noise, just isolated enough to make our work go more smoothly.

Re the 60Hz + harmonics. No constant voltage transformer. My last post mentions what was producing most of those & how I dealt with it last night. There's very little left.

Thanks too, for the reminder about the cable -- I knew that (used to dress tonearm cables on suspended Linn turntables - lol!) but hadn't dealt w/it yet. I have tried placing the entire stand on foam to decouple it from the floor further -- no effect on low freq.

Actually, I believe there is a closer (well, closer to Mike anyway) source for viscoelastic sound damping material. I'm pretty sure that Green Glue is based in Canada.

If anyone is interested in the basics of soundproofing (and likes pretty pictures ), their web site has a good introduction to the subject. Their 5 Principals of Sound Proofing is where I first read up on it when Mike proposed his chamber.

Checked out the sites -- Green Glue is sold only by one Toronto company in Canada, and they don't have any resellers in Vanc. $17/tube + shipping or $180 for a dozen. But I want it here now. Today. I've used silicone glue/caulking for similar apps in the past, but they're pretty smelly/toxic -- not good for use in a room w/o ventilation.

You're right about the principles of soundproofing page -- pretty good info, but nothing new for me at this point.

I haven't posted because I really don't have much to add from a technical standpoint. However, I did want to say how much I'm enjoying reading about the progress, the discussion, and the problems. Well, I'm not "enjoying" the problems, but you know what I mean! I really appreciate the effort here and I look forward to the reviews utilizing the new chamber.

Also, I just wanted to say that, personally, I'd suggest doing this the right way, the first time. The last thing you want to do is have to re-examine all of this somewhere down the road and have to start again. Always think of the end goal, not the short-term objective. In the end, you need to be happy with what you've done and satisfied that it--at least--meets your original intention. Currently, I think you're clearly stating that it does NOT meet part of it.

Trust me... we'll wait!

Yes, you're willing to wait... but there are many complex relationships SPCR has with sample suppliers, advertisers, and others you could call partners -- and they've been waiting for some reviews for months. Their expectation that their samples get reviewed -- whether positively or not -- are starting to push in on priorities here, too. I don't know if anyone has noticed -- we have not touched anything that requires extensive acoustic testing for quite some time. I refer to cases, fans, systems. Why? Because we knew the new room and equipment was imminent and wanted to wait until it was all ready to do the reviews.

Then there's the simple fact that I have only the month of July as a bachelor geek -- my better half returns from an extended business trip Aug 1, and we'd both be much happier if the mess that really exploded after she left is cleaned up before she gets back.

I know the floating room-in-room would give better Sound Transmission Class (STC) results -- but keep in mind STC only looks at 100Hz & up. There may be no simple solution to keeping out the rumble of jet planes and distant traffic much beyond what has already been achieved. Diminishing returns.

Anyway -- back to the last bit -- applying damping to the inner door. Then v1 of the anechoic chamber will be done!

According to principle 5, you might possibly be gaining more noise right now through the floor as its the only part that has not recieved any serious soundproofing, and as far as I know that might be where your main low frequency noise is coming from. I have no idea about frequency characteristics of different materials, but it seems reasonable to suggest that concrete may be a good conductor of low frequency noise.

http://www.greengluesoundproofing.com/p ... index.html

My point is essentially, that if the floor is the problem then it would be quite easy to add a floating floor to your existing room, or would there be a degree of extra work involved on top of the floating floor that I am unaware of.


Andy

OK, it's ugly, the door doesn't really want to open, but the walls and ceiling are completely lined at this point. A table, some tools, a lamp, two stools and the mic on its stand are in the room. I've been listening and monitoring for about half an hour.

It's 4:30pm, Tuesday. Lots of traffic on Main St, tho I don't hear it at all at my desk. On the Freq Spectrum scale, w/o weighting, everything from about 150Hz up is at -23 dB or better, almost all the time, regardless of distant traffic or most planes. It's the sub-150Hz region, especially below 100Hz, which is actively jumping up and down along with traffic/panes at a distance. The A-weighted SPL throughout is at 10.2~10.6 dBA.

Only cars passing within a block affect the A-weighted SPL, which has stayed under 14 dBA even in those peaks.

The low freq rumbling is clearly discernible at full gain / playback volume but not audible at all at the level we recommend people set the volume when listening to the recordings. Applying the 80Hz - 12 dB/oct low cut filter makes it mostly inaudible even at full gain.



Enuf. I'm throwing in the towel for the day. Time for a beer!

Awesome project, Mike! The results look amazing.

If it were my house, I'd be tempted to use the room to listen to music - especially with really good open headphones (which let in ambient noise). And children (visiting or resident) might be tempted to hide or play in there, too. So it might be wise to seriously consider the suggestion of a countdown timer in the room, childproof locks, and some automated indicator to people outside the room that someone is inside.

(I don't have any specialized knowledge, so this is just a surmise based on what I've read in this thread...) About floating floors - since the walls and ceiling aren't transmitting much noise up from the floor, it doesn't seem necessary to isolate the walls and ceiling from the rest of the house. I mean, the next time you upgrade the room, you probably could install a floating floor inside the existing (lined) walls, right?

Just some idle thoughts.

Is the roof of the house steel? If so it might be why your picking up aeroplanes as the roof will just act like the worlds worst (and largest) drum. Although I doubt these rooves are very common in Canada.

I wonder if sealing the room with mdf/particleboard/gib (sorry brand name, its paper plasterboard) would cut down on some of the noise. It should hopefully reflect most of the sound the makes it through, although might not help with low frequency so the point is somewhat moot.

That's a damn good point, which obviously I couldn't be aware of. Sounds like you need to do exactly what you're already doing and maybe come back to the room at some point in the future.

Again, loving the progress and truly appreciate all the hard work you're putting into this!

MikeC, go ahead and do a review or two with the current configuration. While I understand - and generally agree with - Skirge's point about doing things right, the new setup sounds like it's ready for a trial run. I see no reason to tear it all apart before trying it in actual use.

golemB --

Yeah, it's a cool project but it's also lot of grunt work. I've been constantly beat by the end of the day for weeks now.

Regarding the safety issues around the airtight room:

The US Dept of Homeland Security suggests that ten square feet of floor space per person provides sufficient air to prevent carbon dioxide build up for up to 5 hours. With all the blue fill in there, let's say the the dimensions have been reduced a bit -- to 11x9' -- which is 108'. I, alone, could stay breathing and alive in there for a couple days, according to DHS.

A more detailed scientific article entitled What would happen if you were locked in an airtight room? at How Stuff Works says in a 12x12x8' room , there's 1,152 cubic feet or 32,621 liters of air, which would take 3 days for a person to inhale/exhale. However, the reduction in oxygen and buildup of CO2 come into play. "When oxygen falls to 19% and the carbon dioxide level reaches 2%, you're in trouble. Not only are you getting less oxygen, but you're now also taking in carbon dioxide, which is actually a poison. Therefore, in reality, you will only last a day and a half or so. Then your body will begin having problems. So, you had better start working on finding a way out of that room!"

I think it's safe enough. It might be great for naps in the middle of a noisy day or when there's a too-noisy night-time party in the neighborhood.

Aard --

The roof is tar & gravel.

Additional layers of drywall or equivalent on the existing walls with resilient stuff like Green Glue in between would certainly improve the overall noise isolation, but the effect is limited -- or at least assured -- to only frequencies above 125 Hz. STC ratings, widely used to rate acoustic transmission qualities of building materials, are specific to 125~4000 Hz.

Here's what STC Ratings has to say about low frequency noise: "The rating provides no evaluation of the barrier's ability to block low frequency noise, such as the bass in music or the noise of some mechanical equipment."

Everything I've read/studied suggests that to keep out low frequency noise, I need mass. Lots of it, everywhere. The low freq noise of planes has to be airborne and comes right through the house like a hot knife thru butter; it's not just a floor issue.

Having said that, there's no question that the floating room-in-room with perhaps duble layers of QuietRock and green Glue would help, but there's no question too, that this will not eliminate the external low freq noise interference, simply reduce it. By how much we can only guess or estimate. It's really a costs vs benefits question.

The 80Hz-12db/oct filter on the mic preamp is a really useful tool here. Turning it on essentially reduces the low freq noise interruptions by some 75% -- in other words, it makes it possible for recordings to be made through all but closest low freq noises, like a car within a block or a plane at its closest distance to the house (still several miles I'm sure, except for the pesky little ones that sometimes fly close to overhead!).

So then the question becomes -- could I hope to achieve this kind of low noise reduction from any practical construction in that room? Well, what is the reduction of the low cut filter? it's 12 dB/oct and starts at 80Hz. That means it's -12 dB at 40Hz, -24 dB at 20Hz, and -36 dB at 10 Hz. Anyone with even a modicum of acoustics knowledge will answer: NO WAY! But a floating room-in-room might be achieve a general... say... 5-6 dB reduction in the sub-150Hz region, along with probably 10-20 dB further reduction in the frequencies above (not that this is so useful).

But take a look at the low freq curve in my last post and tell me whether that level of reduction would be meaningful. At 100Hz, the level is already >15 dB higher than at 1kHz. At 50Hz, it is >25 dB. At 20 Hz, it's >40 dB. To get these freq to midband levels, we'd have to go underground -- which takes us right back to the ridiculous acoustic bunker under the back yard and garage fantasy I've had for six years!

At what point does this project become an obsessed "How low can we go?" crusade?

You might be kept awake by your own heartbeat and breathing noise...

Those oxyxen consumption/CO2 poisoning calculations don't seem to take into account metabolic rate, which would be increased by the "ooooh shinnyyy computer bitssss" factor.

Here's a relatively simple solution:

Apply an inverse equalization filter in the digital domain based on the freq spectrum of the room ambient. The filter would only apply to frequencies below 200Hz. The idea is to change the response of the mic/room so that it's essentially flat across the whole freq. spectrum. Then any changes from flat in the response curve would be easily visible.

Different EQ filters could be created for rush hour, midday and quiet external conditions -- since they change the low freq ambient in the room.

This can be done for the freq. spectrum display and SPL readings in Spectra+ as well as to all recordings. It would essentially eliminate the low freq boost of the room.

Actually, this would probably be useful regardless of whether further soundproofing work is done in the room or not.

Example of low freq EQ'd Spectra+ screen capture of room ambient on a relatively noisy morning. Note vertical scale -- top is around -5 dB.



Of course, the trouble with this method is that any components which have significant output below 200Hz will be "helped". Unfortunately HDDs' fundamental freqs are at 90, 120 or 167Hz depending on rotation speed... so this method is not ideal.

FYI, a Nexus 120 fan, which we've previously measured as 22 dBA/1m using the old B&K SLM in our untreated room, now measures 18 dBA in the anechoic chamber with the new Spectra+ SLM. (This is without any further filtration other than the A-weighting.) It so happens that Nexus claims 18 dBA for this fan. They go to some lengths to explain how the measurement was done in an anechoic chamber... with an ambient SPL of 15 dBA.

That's quite interesting. Looks like scythes and most other quite fans will prove to be even quieter than tested previously, making them appear even better.

However, I think that more ambient noise gives a more realistic sound measurement for most users. As nobody (I assume) has such sound isolating rooms for regular computer use, having a little bit more surrounding sound will give more realistic fan performance.

I hope you could understand that.

Here's a response curve of the Nexus 120 overlaid against the ambient of the room. The green line is the room ambient; the darker blackish line is the Nexus. 10.58 dBA was the ambient; 17.94 dBA was with fan turned on, mic 1m away.



You can see there's very little noise from the fan below ~80Hz. Most of the noise is 100~2000 Hz. But there are tonal peaks at ~110Hz, ~140Hz, 270Hz -- there are a few others but the levels are probably too low to be audible.

Of course, the problem with both of these lines is that each represents the sound only at one moment, a moment in which there could have been different external influences on each line, which would affect the <150Hz region. I have to figure out how to obtain an averaged curve.

Keep in mind that they won't be any quieter, the fans won't change one iota -- they'll simply measure better.


This is the line of thinking that made me disdain anechoic chamber measurements for a while. However, the basic reasoning is flawed. Measuring the fan in a noisier ambient simply means you're getting more of the ambient and less of the fan in your measurement.

Let's put it this way. Let's say we find a fan that measures 12 dBA in our chamber. Now if your room ambient is actually 22 dBA you probably won't hear the difference between this fan and the Nexus 120. But you can add 3 dBA for the subjective sound of each fan -- due to the sonic reflections in your room. Now we have 21 and 15 dBA. You need 2 fans tho -- one on the heatsink and one for case exhaust. Now you have 24 dBA and 18 dBA.

Which fan will you use? The 12 dBA fan, of course.

However, if our ambient was 18 dBA, we'd get something like 18 and 19 dBA for these fans. The difference between them would be obscured by the higher ambient.

One point of this project is to improve the resolution of our SPL and other audio measurements. This will be a real benefit for silent computing.

Would it be possible to subtract the background noise spectra from the measurements in linear space instead of changing the response? I'm not sure if this would interfere with the A weighting, and it would be problematic if the noise levels during the measurements were lower than the assumed noise level. But to me it would be the most natural way to treat the measurement data.

edit: I just want to add that its a very impressive job you have done here, I'm really looking forward to new reviews with the new setup!

I should hope not!



That is closer than Sweden, right?



If you are serious about using the product I'm sure you could pick up a leftover tube from someone on-line (ebay?). With your current plan you would probably only benefit from using it on your two doors, so you would only need one (and some wallboard). Of course, you will need more if you plan to use it when building that floating floor. From what I have seen online, this stuff is good for reducing low frequency noise transmission when used on a floor (and walls too, but I don't think it would help your walls).

Hey Mike,

I just quickly browsed through the thread, so I'm not sure if it has been mentioned before, but maybe isolating a platform on which you can do your measurements with bungee cords might help. That's what a lot of research labs do to eliminate vibrations for atomic force microscope measurements.
An example is shown here:
http://comp.uark.edu/~jchakhal/the%20afm%20setup.jpg

I have seen constructions with bungee cords up to 2m long.
Well there are also optical floating tables, but they won't fit into your budget.

As "Webfire" just mentioned tables, and i vaugely remember seeing a mention to flat surfaces within the test chamber - could something as simple as the flat floor surface or table cause measurable anomalies with your high tech kit.

Have you tried covering the floor with a soft material (duvet, pillows etc that can cover a large space and are easy to test with as they are available), and removing the table leaving just the mic in the room.? At the very least a test with such things would tell you very quickly whether you are heading in the right direction, and should be doable with a short time gap between tests so the outside ambient wont have changed much. I remember reading (on SPCR) about flat surfaces that face each other, such as the insides of a computer case, maybe the noise is literally bouncing up and down between the floor and table.? Just a thought.


Andy

Webfire --

The low freq noise is in the air. No question about that. It might shake the walls, ceiling and floor, but that in turn vibrates the air. I tried placing the whole mic/stand on soft foam -- half compressed -- and there was no significant change in low freq levels.

andyb --

Not a bad suggestion, but remember that the walls and ceiling are covered with 16" of blue fill. There's very little reflection in there -- and those show up as narrow spikes, which aren't the problem. Removing the table won't change the low freq acoustics.

I dont know nearly enough about accoustics to help much with your spike and low frequency issues, but I do know that the floor is the only hard flat surface with no damping at all - could it possibly be the floor to blame in its current state for either the spikes or low Hz noise levels.? Could it be tested to any degree of accuracy with such crude methods as to throw a couple of duvets in the room covering the floor to see if the noise characteristics change by 1/2 a Db, or would youre testing have to run as far as a fully isolated (floating) floor to even find out if its the culprit.?

I cant claim to be thinking outside the box, but I am having a go at thinking up a quick and easy test so that you can identify your problem and fix it later, leaving you open to put your new room and testing equipment through its paces, or fix it once it is found delaying you testing even further, but knowing that its perfect (withing reasonable limits) and doing your testing Dont think I am impatient, I just want to see you achieve your goal and make SPCR well beyond the reach of the best amateurs (and of course other review sites) - your their right now (and 3-years ago, but their is no problem with making everyone else look like 2nd rate amateurs


Andy

I appreciate your thoughts, Andy. As it stands, it's safe to say...

1) the new room and test gear is eminently usable, and much better than what we had before. Our low noise resolution should be good to below 15 dBA. The ambient noise of 10 dBA is probably close to our measurement limits. I've been told that the ACO Pacific mic has a noise floor of 8 dBA, which is probably lower than the noise of the M-Audio Tampa digital mic preamp being used.

2) we still have to pay attention to external noise but it's not nearly as bad as before. On the noisiest days (like if/when there is major house construction less than 2 blocks away), we could probably still do measurements but not recordings. In the past we'd left both mics and SLMs alone on such days.

I'm very impressed by the scientific way you've managed to find and eliminate sources of noise contamination. I know you are a perfectionist, but I think you have probably done all you can to easily lower the ambient noise. Its time to start to use the chamber and prosper through your handiwork

Alex.

Mike, do what you think is right but I won't complain if you stop here. You've made leaps and bounds over the old testing conditions.

You know, I'd love to have the floating floor room-in-room all built and lined with batts already. I know it would have better isolation -- I don't know exactly how much better, but definitely better. It's what I'd prefer to have.

But...

The practicalities are difficult to sort through. It's not like the materials are all on hand and ready to be used; they have to be ordered, delivered, stored. I'd love to have expert help, but Rob's also a busy guy with lots of demands on his time, so I don't know if I can count on him. And finding a good construction guy in Vancouver is no easy task -- they're in extremely high demand, there's lots of work for them, so this would just be a piddling little thing.

Anyway.... I will decide in the next day or two whether to move on to v.2 now or postpone it to a later date. And yes, it would be a postponement, not a cancellation.

In the meanwhile, I have some writing work to be done!