Plug N’ Chug, Part Two
September 26th, 2008(This is the second post in a two-part series on earplugs, sound, decibels, and noise reduction ratings, the last one being the number used to “grade” earplugs. In part one, I laid the groundwork and discussed a couple of simpler problems with noise reduction ratings. In part two, I’ll discuss more complicated issues with the rating, loudness levels, and the different earplugs I tried, as well as my conclusions on how good various earplugs are and why. Reading the earlier post before this one is a must; Otherwise you’ll probably have no idea what the Noise Reduction Rating is, among other things.)
When we last left off, we found out that the Noise Reduction Ratings (NRR) for earplugs are flawed for two reasons. First, people often wear earplugs correctly, or not at all correctly - there’s no middle ground. The NRR takes this into account by averaging the effect, which is not ideal for either the right or the wrong wearers. Second, the NRR relies on a decibel (dB) rating, which is confusing and unintuitive, not least because the dB is based on a log scale (two of them, actually).
However, these are largely structural problems. There are more subtle and damning criticisms you can make, which I’ll be getting into in this post, along with a couple of practical issues with earplugs that I’ll go over later.
One big problem lies in how humans hear things: not all frequencies are created equal. This concept is best explained by what are called “equal loudness contour curves” (ELC curves). If you check out that link, you’ll see what I mean. The graph there plots sound frequency (Hz) vs. sound pressure (dB), and each curve is based on average subjective loudness as reported by many people in hearing studies.
As you can see, the shape of each curve is a bit like a ladle, with the lowest “dip” at 1k-5k Hz frequences, which means that we are best at hearing things in that range. A sound in this frequency range will sound louder to us than in other frequency ranges. An excellent example of this is when you hear a baby cry or a kid whine. Since they are often in this sensitive frequency range, they can seem absolutely ear-piercing at times.
By contrast, higher or lower frequencies than this seem much quieter to us for a given dB level. For example, a 60 dB (”regular conversation”) sound at 20 Hz (the lowest bass noise most humans can hear) sounds about as loud to us as quiet room “noise” at 1,000 Hz. Bass can even break the 85 dB threshold, where you can have your hearing damaged over time, and seem as loud as a high pitched whisper, if the bass is low-frequency enough. This is actually a frequent cause of hearing loss, since people don’t notice the damage the bass is doing to their hearing. It can also happen, to some extent, at the very high end of frequencies; however, those frequencies typically are only 10-15 dB higher on the ELC curve than the lowest point, so it’s really not as much of a problem. I guess if you love really loud dog whistles, it could be an issue.
Each equal loudness curve is measured in an unusual unit called the “phon”, which varies based on how high a given curve is on the graph. The higher the curve is on the ELC graph, the higher the phon. Thus, the phon is a unit of “equal loudness”, as humans perceive it. If you measure something in phon instead of dB, you automatically know how loud humans will think it is, without having to know the pressure (dB) or the frequency (Hz). That’s pretty convenient, until you try to tell someone what “phon” something is and they think you have a speech impediment.
Linguistic issues aside, you’ll notice that ELC curves get much flatter the higher you go. That’s easy enough to explain - if sounds increase in loudness (dB), you eventually stop caring what frequency they are. They’re just loud. And at dB levels high enough to hurt your hearing, your ears practically don’t care anymore. (Reminds me of the “big speaker” scene from “Back to the Future”.)
Where does the Noise Reduction Rating fit into this? Well, the base for the NRR is around 89-92 dB. You’re supposed to subtract the NRR from that base to see how well the earplugs work and what the average noise level would be with them on. (Remember, the typical NRR is between 12 and 33.)
But for what frequency? As we already know, the frequency of the sound involved goes a long way in determining how loud sounds are.
It turns out the NRR has a response for this too. They use something called “A-Weighting”, which attempts to incorporate the lessons of ELC curves and make a counter curve to cancel out their effect. If you examine the A-Weighting curve, you’ll notice that’s it’s basically the same as the 40 phon ELC curve flipped upside down. That’s no accident. A-Weighting was designed with the 40-phon ELC curve in mind.
So, if you take the performance of earplugs, in dB, and discount their effects via A-Weighting, you should get a straight line again. That is, the performance of earplugs, when A-Weighted, should be at a constant noise-reducing level across all frequencies. Then you can get away with giving people a single number as the NRR, since it stays constant.
But not so fast. That only works for ELC curves that are near (i.e. similar to) the 40 phon one. However, the ELC curves flatten as they go up. Therefore, if you’re in high-noise environments (the most dangerous ones), the NRR overestimates how well the earplugs will do against lower and higher frequency noises. (Especially the lower ones.) Watch out if you work on a construction site, I guess.
Even worse, the ELC curves that A-Weighting is based on were developed using pure tones. Unfortunately, unless you’re trying to block out a loud neighbor that only listens to experimental electronic music, you’re not always trying to block out pure tones. Just as often you’re trying to block out noise, which comes in many different varieties.
(As a side note, that link talks about something called “purple noise”. White noise I’ve heard of, but purple noise? Sounds like a Prince album gone bad or something. And it says here that the NRR actually uses something called “pink noise”, some C-Weighting, and has 100 dB as the baseline. Even with the wacky terms, though, the idea here is pretty much the same.)
And you’re not just trying to block out noise, usually - you’re also trying to block out just about any other annoying sound, like a very loud nightingale warbling (true story for me), a woodpecker banging on metal (also a true story), lawn equipment (don’t even get me started), and lots of other fun stuff that isn’t a pure tone. All these different noises can sound more or less loud to us, often as much as 8-11 dB louder. Even the duration of the noise matters.
Unless you want to become a noise-fighting vigilante, you’re probably going to need to know how well your earplugs will do against such scourges. (The funny part is that movie I linked shows the main character wearing earmuffs, which are among the least effective of all hearing protectors.) British people have run tests that simulate these and many other obnoxious noises, and they came up with a better weighting, called ITU-R 468 Weighting. (In addition to being an awful name that is impossible to say or remember, it also sounds disturbingly like the abortion pill RU-486. Who comes up with this stuff?)
Americans, not being fond of newer standards or names you can’t remember, never ended up adopting the new weighting. So the NRR will remain flawed for louder bass - or high-frequency noises and anything that isn’t a pure tone. (In true American form, we came up with our own new standard that we refuse to use, the NRR-SF.)
As if all that weren’t bad enough, there are practical problems to consider as well. Earplugs work best when the outside tip of them is barely extending past your inner ear. However, many people have different-sized inner (and outer) ears, which means the earplugs (be they foam, wax, rubber, or silicone) will often either stick out too far or go too far in. Plugs that go too far in provide less sound protection (oddly) and can be hard to get out. And if the plugs extend too far out, they may look ugly, come out, and be hard to lie down/sleep with. For example, I have long, thin inner ears, so the average earplug is too fat and short for me, making them very uncomfortable and much less effective.
If you want good earplugs, then, what are you to do? The NRR is only a good guide if you understand decibels and have only pure tones around 40 phon or so to deal with. (You also need to understand that if you use earplugs well, you need to add around 8-11 dB to its rating to get a real estimate.) Even worse, the NRR is also calculated using testing performed by the companies themselves - or by a relatively unknown third-party laboratory - which calls into doubt even the measurements used. That makes the NRR pretty well near useless, even as a relative measure.
Thankfully, some manufacturers understand this, and print a small frequency(Hz)/attenuation(dB) table on the package, complete with standard deviations to approximate how the differences in hearing, perceived loudness, and how people use the earplugs alter their effectiveness. (Yes, you have to be a pretty big audio geek like myself to actually use them, though.) However, that still doesn’t help the fact that louder noises will still be a problem, and that companies continue to do the testing themselves.
Since it’s no big deal to err on the side of caution, the best thing you can do (besides looking at those intimidating nerd tables) is to get the best earplugs you can stand (i.e. ones that include a frequency/attenuation table with high average dB ratings) and are comfortable. Foam is often best for pure attenuation, but again sticking things in your inner ear may not be for you. Also, you can maybe pair earplugs with some other form of noise attentuation, like earmuffs, white noise machines, or “canalphones” (headphones that go into your inner ear). By adding another source of attentuation, studies show you can often muffle another 5-7 dB on top of just the earplugs alone.
Canalphones are actually even more effective, since they’re basically earplugs that play music. In fact, they’re probably the best “earplug” solution out there. Not only are you getting something like 30-33 dB of average attenuation right off the bat (especially with foam tips), you are adding at least 10 dB more in “effective attenuation” by playing music. (Louder sounds can block out other sounds, but I read they’re about 1/3 as effective as earplugs.) Even those troublesome bass frequencies can be helped by canalphones, since properly-fitted canalphones have fairly decent bass and can mask outside noises better. This is normally a big problem, actually, because bass frequencies go through earplugs easier (it’s harder to stop lower-frequency noise) and can even vibrate your body instead of just your ears. No earplugs will help attenuate that noise. (Or your anger level.)
Before I finish up here, I’d like to talk about one last thing that *doesn’t* work: “active” noise cancellation. As they mention at that Head-Fi link, active noise-canceling headphones typically have poor-quality sound and do almost nothing to attenuate noise. They work off the principle that equal and opposite sound waves cancel each other out, creating a flat wave, or silence. The idea that consumer-level headphones can do this on-the-fly is patently ludicrous. I have read, seen studies on, and personally experienced the fact that active noise-canceling headphones just don’t work at all. Not even a little bit. They are often sold for a hefty markup by snake oil salesmen such as Bose. You’d be better off humming the music to yourself.
Then again, if nothing else works, you could always buy the “World’s finest earplugs”, which are “blended from the purest beeswax, cotton and lanolin”. And look, they even have a NRR of 34 dB (a full 4 dB outside of the 30 dB range cited by the government). Foam is so passe.
| | | del.icio.us |
October 4th, 2008 at 10:27 pm
Who knew that the differences in earplugs and how they actually work could be so complex?
I checked out the Wikipedia link on the different “colors” of noise. Of interest, I found both the “pink” and “red (or brown)” noise to sound the most like rushing water.
On top of that, “red (or brown)” noise was, to me, the most soothing of all the samples to listen to.
Conversely, the “purple noise” damn near threw me into a homicidal rage
I like to hear these different descriptions, so thanks. That’s too funny about the purple noise. )
- Dave