Understanding Decibels

So it is really really important to understand something: that meter isn't reading the peak sound. 

So I always want to give props to Steve. I mentioned this to him a few days back and he took the comments gracefully and (seemingly) to heart. Obviously not everyone would be thrilled to hear their expensive piece of test equipment isn't really suitable for doing what they want it to do. I suspect he'll continue to use the meter, however it is very important to understand that the decibel number it is showing you is only representative of the sound output in the broadest possible sense. 

So Steve/AEAC uses an Extech 407730. If you look at the manual, you can see that unlike cheaper meters it actually lists its response time, and it has both a fast and a slow response time setting. (you can find the manual here: http://www.extech.com/resources/407730_UM-en.pdf) The fast response time setting is 125 milliseconds, and the maximum frequency is 8khz. Both of these make it unsuitable for measuring impact sounds such as what comes out the end of our moderators. Let me illustrate why: 

This is what the output trace from the sound of an airgun meter ACTUALLY looks like:


The above is a DonnyFL Sumo on a .22 caliber FX Crown running about 32 foot pounds. The X-axis delineations are each ONE millisecond, that is to say moving from one grid line to the next represents the passage of a single millisecond. So the entire measured trace represents just 12 miliseconds, one tenth the response time of that Extech meter. So, in short, it just can't possibly measure that peak, which occurs in well under one tenth of one millisecond.

Here we can see another trace. What you're trying to measure is highlighted in purple. What meters like this are actually measuring is highlighted in teal. 



Basically the Extech meter's response time is over two orders of magnitude too slow to measure airgun moderator sound. Given that Extech usually makes pretty decent entry level test equipment, and that the meter has an optimization for fast response time, I'm guessing it is considerably better than most alternatives out there. 

Now it is important to note that human perception is highly frequency weighted, and this varies from individual to individual, and is also highly time weighted. (a hammer blow on a nail is tolerable, however that sound played continuously would be physically debilitating) As a result, the Extech meter used here, and other unsuitable meters like it, may bear a casual resemblance to perceived loudness. It is not however a suitable way to differentiate between airgun moderators' performance. You'll note this probably also is why the differences between stock shroud and the three moderators all were almost negligible, well within what you'd expect the margin of error to be. 


I hope that is a clear and helpful explanation of the issues surrounding the measurement of sounds like this with inexpensive (relatively) sound meters. Systems which are actually capable of taking measurements like this are fairly specialized and expensive for a reason unfortunately. Anyway I hope that helps.  

I think the least expensive way to get a rough idea of relative decibels his to use a PC and Audacity. I know it has shortcoming, many shortcomings, the response curve of the microphone isn't linear the PC A/D converter isn't accurate etc etc. However in practice these errors are minimal compared to the sound environment where the measurements are made. They're are reflections absorptions standing waves frequency null points etc. But this is free and way better than most sound meters the cost a few hundred dollars. You can get soundwave amplitude versus time plot as sto shows and a frequency spectrum vs. Decibel.

To use it you have to ensure that you never clip the input by setting the gain correctly for the peak sound pressure. You can sample at the standard 44k or you can even go higher 96k or more depending on the PC. This is more than enough the capture and he sound peaks humans would hear. Humans can't hear anything over 20,000 to 22000 Hertz when they are young animals could hear much higher.

Since we're discussing sound I find it interesting that the audio marketing experts keep pushing amplifiers with low total harmonic Distortion super good signal-to-noise ratios in great flat linear responses. Where in real life all speakers are way worse than any of the electronics upstream and the room Acoustics are dismal as well.

Measuring the loudness of a gun in decibels is very inconsistent and in no way let’s a person know how loud it is. 10 people can measure the same gun and get very different results. Too many variables. This is the biggest waste of any review.

i would rather the reviewer tell me “ its pretty quiet”. Rather than say 77 dB.

When I don’t hear my wife asking what I’m doing out back I know the gun is quiet.

Click to expand...

AGREE ... that's enough scientific evidence for me

When I don’t hear my wife asking what I’m doing out back I know the gun is quiet.

Click to expand...

AGREE ... that's enough scientific evidence for me

Click to expand...

So true, but I would add equally as important is when the neighbors don't say "what was that."

i just lucked out on my old .22 disco , i guess. tko stage5 from ten yrs ago and i can really just hear the hammer and spring and maybe some flatulence when pressures get around 1400 psi . i fill it to 2000 psi . but why cant we compare the freaking loudness of these things to perhaps , other well known airguns ? the gun would not be sneaky enough for what we usually do if it were as loud or louder than a sheridan on 6 pumps - most break barrels would be too loud unless soft tuned like my old gamo 1000 - thunk ! daisy red rider could be a level two .... wonder what decibel those a.z. theoben rapid mk2 / or mfr bottle guns register . those are so quiet you really dont even hear the trigger and firing process in the receiver. ........ just the .25 cal pellets hitting backstop would get us busted , anyway - especially from a gauntlet . we should ask those n.western pcp tuning guys - i saw a video of them tuning a .25 gauntlet . seems they quietened it down , too . . http://www.jsairrifles.com/Umarex-Gauntlet-Air-Stripper-25_p_147.html i'm just blabbing , but i always wished that those noise ratings at pyramyd would actually have clear meanings.

I have been using centerfire and rimfire moderators for over 20 years and can tell you the sound produced by one mod VS another is very subjective indeed as said above.

But with respect to airgun moderators which is a different thing altogether, what surprised me the most is that the FX moderators which have no internals, effectively just a single dump chamber can be as or even more effective than some airgun moderators of the same size or volume but with complex baffle arrangements.

Some companies sell true db meters that can give you very accurate db/SPL readouts regarding sound/pressure. I used to own one almost 20 years ago when I was trying to win SPL competitions with the bass coming from the sound system in my truck. I peaked at 149 db... I lost by 3 db.

Sound "loudness," or sensed in terms of db/SPL is not linear, it is exponential. Jumping from 80 db to 90 db is not a 12% increase, it is more like 40+. Going from 140 to 150, there is literally hundreds of percent in power needed to achieve that.

So thinking backwards, when someone says that a gun with "the best" mod on it only quiets the report by 30 db... Again, it is not linear. Dropping 30 db off the top shelf of the report (let's say, going from 90 db to 60 db) is actually the same as saying that you are reducing SPL pressure to your ears by over 70%. Fuzzy math on my end, but you get the point I am sure.

I love talking sound, so stop reading when I've gone too far.

A bel is simply the log (base 10) of a measurement compared to a reference. Thank Alexander Graham Bell for that. A decibel is just 10 times a bel.

We measure sound pressure levels as decibels. Problem is, sound pressure varies with distance from the source so most of those charts you referenced are BS since they don't tell you where the pressure was measured. Another problem with comparisons is sound pressure also varies with the environment, absorptive versus reflective, indoors versus outdoors, etc. The only way to really compare is to take measurements at the same distance from the source in the same environment. If you double the distance from a source, say 10 feet versus 5 feet, the sound pressure decreases by 6 dB. This is known as the acoustic inverse square law. You can also look at it by saying it takes a 100% increase in pressure (doubling) to raise dB by 6.

There is also another dB associated with sound called the sound power level. But that's another topic.

Here's another problem, as mentioned in another post. Sound level meters have different response times. Basically, it's the time they "listen" to a sound and average the peaks/valleys of the wave form to give you a single dB reading. You can have slow, fast, impact, peak.

Oh, and here's another confounder. Sound level meters also have different weighting networks (sort of like filters). A, C and Flat are the most common.

When looking at damage risk from noise, we use A-weighted, slow response measurements since that approximates the way your ear responds to pressure waves. 

What I'm getting at is that to compare noise levels, you must take measurements exactly the same way, in the same environment. Peak measurements may be relevant for comparing moderators but they may not be relevant to how we hear the sound. 

Finally, there is the issue of the "quality" of sound. One moderator might be more quiet than another relative to dB, but it might seem more harsh to our ear.

I'll stop now.

I don’t understand the obsession of spending $200 to quiet an air gun! I think mostly some people like the looks of a moderator more than worry about noise.

Really good post.  

I don’t understand the obsession of spending $200 to quiet an air gun! I think mostly some people like the looks of a moderator more than worry about noise.

Click to expand...

Plenty of people have said it. Nosy neighbors, jumpy prey, or just the joy of precision shooting in peace and quiet. I'm sure some people like the looks as well, I know Tominco said something like it was half about style and half sound for him. Depending on the moderator, you can get CONSIDERABLE sound attenuation. Let me just give you an easy to visualize example:



This trace is with the stock shroud on my FX Crown fully extended, as quiet as the gun gets "factory." It isn't painfully loud, but it does have quite a bit of bark to it, the neighbors will certainly wonder if not know what I'm doing. If you have a sound reflective surface in front of you, you'll definitely get a little wake-up call with every shot as that muzzle blast comes back at you. I didn't like that and I wanted to set about quieting it. 







So this is a moderator I built for it. It is significantly quieter, both from a measurement perspective and from the shooter's perspective. The hammer slap is now the loudest part of the gun's action, and at 150 yards the sound of the pellet hitting a steel target is louder than the gun going off. It makes shooting more pleasant, and somehow making an airgun very quiet seems vaguely nefarious and slightly giggle-worthy. 

I should note the Y axis scale was changed between those two traces, if memory serves the shroud trace is twice the vertical scale of the mod trace. This has to be done, otherwise the factory shroud would be off the chart or the moderator would be too low Y-res to nicely visualize. The actual raw values (averaged across multiple pickups and shots) were 288 for the factory shroud and 52 for my moderator design. I'm not particularly confident in my meter's calibration, which is why I don't provide the dB conversions for those, just the raw values, however that probably corresponds to very roughly 15 decibels of sound attenuation. Rough rule of thumb is that every time you double the distance between yourself and a sound source, you cut the dB by 6. So this would be equivalent attenuation of being over four times further away from the sound source. So lets say your neighbors are 150 feet away from you, and are getting a little curious about what this air-burst sound is coming from. By screwing this moderator on, you'd change their perception of the sound to be more like they were 200 yards (600 feet) away. Again this is all rough napkin math here, but you see the point I'm trying to illustrate I hope.  





I hope that explains it a little better, even though it is just the reasoning/perspective of one crazy guy out there on the internet.  

Those are time-domain captures - give you some idea of amplitude but does little to describe the "sound."

You need to do an FFT to convert to the frequency domain so you can see how the sound varies with frequency.

Then you should be able to see how various moderators do their thing.

Those are time-domain captures - give you some idea of amplitude but does little to describe the "sound."

You need to do an FFT to convert to the frequency domain so you can see how the sound varies with frequency.

Then you should be able to see how various moderators do their thing.

Click to expand...

So I would like to try doing FFTs at some point but I've been mainly focusing on total amplitude regardless of frequency. Taking a closer look at the captures, across a bunch of different moderator designs, and just by looking you can see there seems to be a dominant frequency that doesn't change much. It is in the 4-8khz range and that is well within human hearing range, and seems rather consistent throughout my testing. In essence then, the double-digest of the data might not actually provide you as much additional information as you'd hope. I'd still like to do it at some point, it just hasn't been the priority. 

I should point out that our peak is exactly that, a single peak, not really a fully formed wave at all. This is how these things are measured in the silencer industry, frequency blind, and I just carried this over to our analogous airgun application. In all honesty, I'm still in the stage of pursuing big gains such that frequency tuning just isn't on the menu quite yet. The difference in gross amplitude is so large...... 

Knowing the frequency of peak would steer you toward using a design with a nulling response AT that frequency.

Ultimately you would want both, a measure of the specific sonic signatures of each "producer" as well as the "suppression-vs-frequency" curve of each moderator.

Then mate the moderator that has the biggest null matching the frequency of peak output of your particular rifle.

I found this post to be very interesting as well. I have used cell phone app, computer app, and off the shelf db meter. None of them gave me an idea of how loud it is or be able to compare to someone else to give solid advice.

Just a thought here,

Maybe it's time to make some kind of "cheap" sound standard gadget. Like an adjustable noise maker that runs around a similar frequency and time that our airgun run. Something like a " calibrated" 5db increment adjustable noise maker.

Using a standard like this should really give the buyer of an airgun or LDC a great idea "with their own ears" what to expect.

Good stuff. You can learn something every day. Tx gentlemen, very interesting. Also, nothing wrong with moderating sound for a given environment and objective. You never hear the one that hits you...

Measuring the loudness of a gun in decibels is very inconsistent and in no way let’s a person know how loud it is. 10 people can measure the same gun and get very different results. Too many variables. This is the biggest waste of any review.

i would rather the reviewer tell me “ its pretty quiet”. Rather than say 77 dB.

Click to expand...

Yes, I agree, for what it’s worth I make all my own silencers and shrouds for my guns ( legal in the uk despite our restrictive gun laws) up until recently I have done all my testing by ear. To me if one sounds quieter than another, then it is but I have been given a calibrated dB meter, and the results are surprising to say the least.

with our 12fpe limit they are never going to be loud but unmoderated they do make a hefty crack, with just a simple tube filled with felt the crack becomes a pop, a custom made calibre specific mod then the crack becomes a muted fart, but how loud is that?

While testing one of my lasted shrouds fitted to a custom theoben, I was very pleased with its sound level or lack of it. The strange thing is, on the dB meter, cycling the bolt peeked 2dB more than the muzzle report yet to my ears the muzzle report was much louder than the bolt and compared to some of my less quiet mods the results were almost the same.

I have given up on the dB meter as it just doesn’t equate to anything that I hear, so my mods are either “quiet” “very quiet” “extremely quiet” or “insanely quiet” so Insanely quiet a .177 pellet sounds louder as it tears through a paper target at 30 yards than the report of the gun!

An example of insanely quiet on a heavily modified theoben mfr



Bb

I found this post to be very interesting as well. I have used cell phone app, computer app, and off the shelf db meter. None of them gave me an idea of how loud it is or be able to compare to someone else to give solid advice.

Just a thought here,

Maybe it's time to make some kind of "cheap" sound standard gadget. Like an adjustable noise maker that runs around a similar frequency and time that our airgun run. Something like a " calibrated" 5db increment adjustable noise maker.

Using a standard like this should really give the buyer of an airgun or LDC a great idea "with their own ears" what to expect.

Click to expand...

About 20 years ago, Radio Shack used to sell a few analog meters, and they were very responsive. I wish I still had the one I bought, it was the most expensive one (can't recall the brand now) but it had a needle that would hold and was resettable at the peak measurement.