Do Digital Sound Meters meet testing requirements?

[John Titsworth]

I have been looking into getting some new sound metering equipment lately. After reviewing MIL STD 1474D and reading the requirements for digital recording.

MIL-STD-1474D
5.3.1.2.2 Digital recording. Sampling rate shall be a minimum of 160,000 samples/sec. The analog signal shall be filtered, using a Bessel type, with 40 kHz cutoff frequency. The roll-off rate shall be not less than 36 dB/octave.

It turns out that meters that are affordable (I know affordable means different things to different people) don't meet this requirement. For example the B&K 2270 is only 48,000 samples/sec (this meter is used by Silencer Shop). The Silencerco meter uses a National Instruments product that only does 102,400. Neither of these meters meet MIL STD 1474D requirements. Bummer, cause I wanted the B&K product. Larson Davis puts out a meter they claim is a silencer testing meter, but it doesn't meet the requirement either. Its around 51,200 samples/sec. Its the Larson Davis 831.

B&K says they have another option that will work and do 192,000, so I will look into that and depending on the price I will go that route. Probably over 20K would be my guess. The B&K pulse system is over 40K.

I just found it interesting that the meters that a number of companies and dealers have been pumping out numbers with are not only too slow to catch the peak of a suppressed gunshot, they don't meet the standards of military testing either. 90% of consumers will never know they are getting sold on sound reduction numbers that are bogus. On the low end (22 suppressors and quality 9mm suppressors) this won't usually be a huge issue, but on stuff that is borderline, these digital meters mentioned are producing numbers that could potentially make unsafe impulse noise seem safe. Not a good standard to set for the industry and the consumer.

[doubloon]

A semi-interesting paper that supports your point on sampling rate http://users.phhp.ufl.edu/colleeng/Loba ... m_2014.pdf

It certainly seems possible that numbers published using less accurate equipment would be off. It would be interesting to see side by side tests done with the B&K 2270 and a more accurate impulse meter to see if the B&K 2270 at least occasionally records an accurate number or if they're "mostly accurate".

Do you know if the OSHA standards for measuring peak noise are different from MIL STD?

ETA: I think I found something that says OSHA looks for a 10 microsecond rise time but it's a letter and not the actual standard. Even so a sample rate less than 50k may not be acceptable even to OSHA standards.

[T-Rex]

I'm interested in your discussion, I've been an audiophile for quite some time.

Many companies and individuals tout items claiming Mil-Std specs. When, in fact, Mil-Std is usually the least up-to-date standard.

For sound level metering, I would consult IEC, ANSI and ASTM for the current standards of practice.

These should also show the proper method of instrument setup and calibration to conduct an adequate experiment.

[JasonM]

I'm interested in your discussion, I've been an audiophile for quite some time.

Many companies and individuals tout items claiming Mil-Std specs. When, in fact, Mil-Std is usually the least up-to-date standard.

For sound level metering, I would consult IEC, ANSI and ASTM for the current standards of practice.

These should also show the proper method of instrument setup and calibration to conduct an adequate experiment.

A gunshot is not typical 'sound testing' though. I don't believe there are any ANSI / ASTM / etc. standards for sound testing silencers.

Mil-std is what we have.

The addition of more sound collection points (shooter's ear, etc) are a big step in the right direction. But until there are legit standards in terms of method, equipment, and process that EVERYONE follows, comparing sound numbers is essentially useless.

[JasonM]

90% of consumers will never know they are getting sold on sound reduction numbers that are bogus. On the low end (22 suppressors and quality 9mm suppressors) this won't usually be a huge issue, but on stuff that is borderline, these digital meters mentioned are producing numbers that could potentially make unsafe impulse noise seem safe. Not a good standard to set for the industry and the consumer.

First, hi John, nice to see you on here.

second, I don't think most people care or understand enough about sound to need to know the truth. It's unfortunate.

If someone is enamored with a manufacturer, they believe all the hype.

Fortunately, this doesn't apply to those that make overly-outrageous claims. Like Spikes and Jesse James.

But a few dB here and there are never called out. 99.999999999% of consumers have never seen a real sound meter, never mind own one and so if their can is giving them 135dB and the manuf claimed 132dB, they'll never know.

And I agree about safety, but it goes beyond that because people will buy a "137dB" 556 can and think they're safe...

but then it all comes down to the fact that most people don't actually shoot. and even a 150dB can looks good in arfcom posts.

[T-Rex]

A gunshot is not typical 'sound testing' though. I don't believe there are any ANSI / ASTM / etc. standards for sound testing silencers.
Mil-std is what we have.
The addition of more sound collection points (shooter's ear, etc) are a big step in the right direction. But until there are legit standards in terms of method, equipment, and process that EVERYONE follows, comparing sound numbers is essentially useless.

This is not completely true. Even the EPA has standards and testing for gunfire (IMPULSE rating). They are the ones that set the standards for hearing protection, specifically aimed towards firearm hearing protection.
They have a set of standards for testing which deals, specifically, with outdoor shooting ranges.
Their measurement formula is:



There is an entire paper on the procedure as well as the standards.


ANSI/ASA S12.42-2010 deals directly with hearing protection devices that are purposed for Impulsive Noise, ie gunfire. This is the standard used by NIOSH. The levels used in the testing are non-linear and fall right in the range of an unsuppressed, centerfire rifle.

Between ANSI and IEC, there are at least a dozen standards for, not only, the testing equipment, but the method of calibration and use in testing, for different experiments/scenarios.

MIL-STD-1474D is a generic procedure, not specified for one application, other than human hearing damage.
I think the methods set by internationally recognized organizations would be better suited, at least for an "official" test.

These might shed light on the subject, better than I can explain here.
A-Weighting Filter For Audio Measurements
The Difficulties in Evaluating A-Weighted Sound Level Measurements
AHAAH Report

[JasonM]

A gunshot is not typical 'sound testing' though. I don't believe there are any ANSI / ASTM / etc. standards for sound testing silencers.
Mil-std is what we have.
The addition of more sound collection points (shooter's ear, etc) are a big step in the right direction. But until there are legit standards in terms of method, equipment, and process that EVERYONE follows, comparing sound numbers is essentially useless.

This is not completely true. Even the EPA has standards and testing for gunfire (IMPULSE rating). They are the ones that set the standards for hearing protection, specifically aimed towards firearm hearing protection.
They have a set of standards for testing which deals, specifically, with outdoor shooting ranges.
Their measurement formula is:



There is an entire paper on the procedure as well as the standards.


ANSI/ASA S12.42-2010 deals directly with hearing protection devices that are purposed for Impulsive Noise, ie gunfire. This is the standard used by NIOSH. The levels used in the testing are non-linear and fall right in the range of an unsuppressed, centerfire rifle.

Between ANSI and IEC, there are at least a dozen standards for, not only, the testing equipment, but the method of calibration and use in testing, for different experiments/scenarios.

MIL-STD-1474D is a generic procedure, not specified for one application, other than human hearing damage.
I think the methods set by internationally recognized organizations would be better suited, at least for an "official" test.

These might shed light on the subject, better than I can explain here.
A-Weighting Filter For Audio Measurements
The Difficulties in Evaluating A-Weighted Sound Level Measurements
AHAAH Report

I understand the impulse sound measurement, but it needs to be expanded/specialized for gunshots and silencer performance testing. Sounds like a good job for the American Suppressor Association.

[T-Rex]

I understand the impulse sound measurement, but it needs to be expanded/specialized for gunshots and silencer performance testing. Sounds like a good job for the American Suppressor Association.

I, wholeheartedly, agree.

Obviously the shooter's ear is priority.
Maybe 5ft behind shooter and several, set, intervals perpendicular to the firearm, should suffice. In a safe environment, no one should be forward of the shooter.
I feel more than, maybe, a 100yd line-of-sight test is unnecessary. But, this info could be used to strengthen the "why should silencers be legal" standpoint.

[Historian]

I understand the impulse sound measurement, but it needs to be expanded/specialized for gunshots and silencer performance testing. Sounds like a good job for the American Suppressor Association.

I, wholeheartedly, agree.

Obviously the shooter's ear is priority.
Maybe 5ft behind shooter and several, set, intervals perpendicular to the firearm, should suffice. In a safe environment, no one should be forward of the shooter.
I feel more than, maybe, a 100yd line-of-sight test is unnecessary. But, this info could be used to strengthen the "why should silencers be legal*" standpoint.

+1

Two articles might also be of interest:

1. Wall Street Journal Tuesday, March 10, 2015, D1:
"Hearing Risk From How Loud You Listen,and How Long".

For example 'danger zone' "105 decibels/4 min. - chainsaw" and
" 110 decibels/1min. 30sec. - Shouting into the ear (#$%*&@$ DI's )"

2. "The Brain"s Way of Healing" - Norman Doidge.
Example, opera singers develop deafness singing at 135 decibels.

The book also should be read by folks with children having autism and
older people after a stroke: brain can be retrained and new connections
induced externally to heal leading to recovery. What Doidge is saying is supported by prescient
basic research initiated by associates in 1960's. Key is via ears/electronic encoders.


* And mandated use, not impeded by misguided totemism.

[Skorch]

I am about to perform a series of sound tests for my own personal research on some form 1s. This forum peaked my interest as I was going to use a Larsen Davis 831, and I thought this would be adequate. It was mentioned in the article referenced earlier in this forum.

http://www.cdc.gov/niosh/hhe/reports/pd ... 1-2898.pdf is another article about studying a range.

I have contacted Silencershop if they would mind sharing some details as to their sound testing, but have gotten no response.

As you mentioned I would be considered general public, and do not have an expertise in sound measurement. I am doing this merely for comparison methods, but I would like to know if I will get similar results to that of places like Silencershop.

Could you guys recommend what weighting to use A, C, or Z? I was thinking C as it is above 100db, but the article reference used all A.

http://dtic.mil/ndia/2009infantrysmalla ... xi8453.pdf references inverse square law II to use since my pre-amp is maxed at 140db for measuring firearms unsuppressed.

What do you guys think? I apologize if this creates a side subject. This looked to be the most relevant topic at this moment.

[John Titsworth]

90% of consumers will never know they are getting sold on sound reduction numbers that are bogus. On the low end (22 suppressors and quality 9mm suppressors) this won't usually be a huge issue, but on stuff that is borderline, these digital meters mentioned are producing numbers that could potentially make unsafe impulse noise seem safe. Not a good standard to set for the industry and the consumer.

First, hi John, nice to see you on here.

second, I don't think most people care or understand enough about sound to need to know the truth. It's unfortunate.

If someone is enamored with a manufacturer, they believe all the hype.

Fortunately, this doesn't apply to those that make overly-outrageous claims. Like Spikes and Jesse James.

But a few dB here and there are never called out. 99.999999999% of consumers have never seen a real sound meter, never mind own one and so if their can is giving them 135dB and the manuf claimed 132dB, they'll never know.

And I agree about safety, but it goes beyond that because people will buy a "137dB" 556 can and think they're safe...

but then it all comes down to the fact that most people don't actually shoot. and even a 150dB can looks good in arfcom posts.

Thanks Jason. I don't get on this or any other forum very often anymore. I totally know what you are saying about posts on forums like arfcom. Its usually just a bunch of "buy this cause its awesome and I have one!" Gets old to those of us that have been in the game for a while. Oh well...

There are no real tests being done anymore, except "digital" tests (AKA MARKETING TESTS) and those obviously are not adequate and are not believable. There was a recent post by silencer shop on arfcom showing the SiCo Omega at 141dB on a 308 and they redid the test to try to conform to the new standards that the ASA is pushing (at the request of SiCo). They re-did the tests and got closer, but they are all using digital meters that don't meet Mil Std 1474D. People objected, but its just business as usual.

There was an industry meeting at Shot this year. I did not attend, but spoke to manufacturers who did. The push was to eliminate FRP from the entire equation and move to 4 shot averages AND measure at the muzzle of the host weapon regardless of the suppressor length. They also want to change the weighting network. I am very on board with that. We should at a minimum change to C weighting or none at all. The numbers are better for hearing loss with suppressors. If the standard should be changed it should include only sound meters that meet current Mil-Spec 1474D, use linear or C weighting and have transducers at the reference location and at both ears. The standard should not be changed to encourage buyers to ignore FRP (any hunter knows this may be the only shot you get and is the most important shot).

[Tuukka]

90% of consumers will never know they are getting sold on sound reduction numbers that are bogus. On the low end (22 suppressors and quality 9mm suppressors) this won't usually be a huge issue, but on stuff that is borderline, these digital meters mentioned are producing numbers that could potentially make unsafe impulse noise seem safe. Not a good standard to set for the industry and the consumer.

First, hi John, nice to see you on here.

second, I don't think most people care or understand enough about sound to need to know the truth. It's unfortunate.

If someone is enamored with a manufacturer, they believe all the hype.

Fortunately, this doesn't apply to those that make overly-outrageous claims. Like Spikes and Jesse James.

But a few dB here and there are never called out. 99.999999999% of consumers have never seen a real sound meter, never mind own one and so if their can is giving them 135dB and the manuf claimed 132dB, they'll never know.

And I agree about safety, but it goes beyond that because people will buy a "137dB" 556 can and think they're safe...

but then it all comes down to the fact that most people don't actually shoot. and even a 150dB can looks good in arfcom posts.

Thanks Jason. I don't get on this or any other forum very often anymore. I totally know what you are saying about posts on forums like arfcom. Its usually just a bunch of "buy this cause its awesome and I have one!" Gets old to those of us that have been in the game for a while. Oh well...

There are no real tests being done anymore, except "digital" tests (AKA MARKETING TESTS) and those obviously are not adequate and are not believable. There was a recent post by silencer shop on arfcom showing the SiCo Omega at 141dB on a 308 and they redid the test to try to conform to the new standards that the ASA is pushing (at the request of SiCo). They re-did the tests and got closer, but they are all using digital meters that don't meet Mil Std 1474D. People objected, but its just business as usual.

There was an industry meeting at Shot this year. I did not attend, but spoke to manufacturers who did. The push was to eliminate FRP from the entire equation and move to 4 shot averages AND measure at the muzzle of the host weapon regardless of the suppressor length. They also want to change the weighting network. I am very on board with that. We should at a minimum change to C weighting or none at all. The numbers are better for hearing loss with suppressors. If the standard should be changed it should include only sound meters that meet current Mil-Spec 1474D, use linear or C weighting and have transducers at the reference location and at both ears. The standard should not be changed to encourage buyers to ignore FRP (any hunter knows this may be the only shot you get and is the most important shot).

C-weighting has been typically used on this side of the pond in work health & safety side and in many of the suppressor comparison tests as well.

With regards to the MIL-STD 1474D and the reference locations.

Many manufacturers seem to use the 1m to the side of the weapon muzzle and suppressor muzzle as the #1 reference point, which it is not.

The 15 cm from the shooters ear on a line between the ear and noise source. The standard mentioning either the muzzle or breech.

This is why we we also test at both ears and I would hope all reasonable / competent manufacturers do also.

The 1 m to the side is not defined as such in the standard, but seems to have evolved as a combination of the two secondary reference locations. Do not know who started it originally, I would wager Gemtech?

With regards to testing at the side of the muzzle, looks like there are folks who are pushing for testing at the rifle muzzle, instead of the suppressor muzzle, when conducting extensive suppressor comparisons, on the argument that the meter is not then moved. However reliable measures can be used to align the meter at each suppressors muzzle.

We feel that to be fair to every manufacturer and suppressor type/length, that measurement should be done still at the suppressor muzzle.

The above does not change anything shooters ear wise, which is the key thing, health wise.

Best Regards!

Tuukka Jokinen
Ase Utra sound suppressors

[k31user]

You mean manufacturers stretch the truth to sell product? Say it isn't so.

[John Titsworth]

You mean manufacturers stretch the truth to sell product? Say it isn't so.

HAHA. Some do and some don't. Its all about that marketing...

[John Titsworth]

C-weighting has been typically used on this side of the pond in work health & safety side and in many of the suppressor comparison tests as well.

With regards to the MIL-STD 1474D and the reference locations.

Many manufacturers seem to use the 1m to the side of the weapon muzzle and suppressor muzzle as the #1 reference point, which it is not.

The 15 cm from the shooters ear on a line between the ear and noise source. The standard mentioning either the muzzle or breech.

This is why we we also test at both ears and I would hope all reasonable / competent manufacturers do also.

The 1 m to the side is not defined as such in the standard, but seems to have evolved as a combination of the two secondary reference locations. Do not know who started it originally, I would wager Gemtech?

With regards to testing at the side of the muzzle, looks like there are folks who are pushing for testing at the rifle muzzle, instead of the suppressor muzzle, when conducting extensive suppressor comparisons, on the argument that the meter is not then moved. However reliable measures can be used to align the meter at each suppressors muzzle.

We feel that to be fair to every manufacturer and suppressor type/length, that measurement should be done still at the suppressor muzzle.

The above does not change anything shooters ear wise, which is the key thing, health wise.

Best Regards!

Tuukka Jokinen
Ase Utra sound suppressors

5.4.5 Transducer locations. For shoulder-fired and hand-held weapons, transducers shall be located at the center of each operator or crewmember's probable head location. For other weapons the transducer shall be positioned 1.60 m above the ground surface; for sitting locations it shall be 80 cm above the seat. When the operator must be present, the measurement shall be made 15 cm from the ear closest to the noise source (i.e., muzzle or breech, as the case may be) on a line between the operator's ear and the noise source.


5.4.5.1 Reference transducer.
a. If required, a transducer shall be placed 200 cm to the side of the major noise source of the weapon (e.g., perpendicular to the muzzle for closed breech systems and perpendicular to the rear for rocket launchers), with the weapon and the sensor 160 cm above the ground.

b. If required, a reference transducer shall be located on the 135-degree or 225-degree radial (taking the line-of-fire as 0 degrees). For weapons of bore diameter greater than 20 mm, this transducer shall be located at a distance of 50 bore diameters measured from a point directly under the muzzle with a tube (barrel) elevation of 0 degrees at a height of 160 cm. For weapons of smaller bore diameter, the transducer shall be located at the same elevation as the muzzle, at a distance of 100 cm.

The reference location is not any sort of guess or myth, but its not super clearly defined either. A reference location is certainly necessary and I think its a good idea. I removes many of the variables associated with testing. The location of the transducer is very important, as you know from years of testing.

We can see that a reference transducer (microphone) is called for at 160cm (1.6 meters) perpendicular to the muzzle (in our case the attached suppressor is now the muzzle of the suppressor not the muzzle of the host weapons barrel threads) at a distance of 100cm (1.0 meters). Gemtech may have started it, but a reference location is certainly a good idea not only because it removes suppressor length and host weapon size/length from the equation, it also would let us refer to previous testing.

My new system will include reference location testing and simultaneous right and left ear location testing. The system exceeds the rise time requirement and exceeds the 160,000 samples per second requirement of 1474D by over 20%. The rise time will be 5 MICRO seconds. Since its digital and quicker, I can perform very fast super accurate testing.

I think an industry wide shootout test is in our future.

[ncorry]

Wow, I just spent about 30 mins of otherwise billable time reading this and some of the links. #1- I had no freaking clue that sound meters cost that damn much. #2- to save other people some time, the 147D mil standard that is referenced is available at http://www.silencertalk.com/docs/mil-std-1474d.pdf and Section 5.3.1.2.2 that John references is on page 45 in the Adobe do (page 40 using the paper's page numbers). Oh, and a transducer appears to be a real fancy way of saying microphone.

OK- back to the people that know what they're talking about. And Mr. Titsworth, good seeing you on-line, although we've been e-mailing back and forth recently using 100% Arkansas internets.

[John Titsworth]

Wow, I just spent about 30 mins of otherwise billable time reading this and some of the links. #1- I had no freaking clue that sound meters cost that damn much. #2- to save other people some time, the 147D mil standard that is referenced is available at http://www.silencertalk.com/docs/mil-std-1474d.pdf and Section 5.3.1.2.2 that John references is on page 45 in the Adobe do (page 40 using the paper's page numbers). Oh, and a transducer appears to be a real fancy way of saying microphone.

OK- back to the people that know what they're talking about. And Mr. Titsworth, good seeing you on-line, although we've been e-mailing back and forth recently using 100% Arkansas internets.

Good to talk via the internets in AR, glad they work!

My new system will probably be close to 20K. Not cheap, but not as expensive as the Pulse system at 46K+. When I get it you can some do some shooting on it. Mid April it will be here. Its about time for another Silencer Research style shootout with real sound testing equipment.

[Grounded]

I have been looking into getting some new sound metering equipment lately. After reviewing MIL STD 1474D and reading the requirements for digital recording.

MIL-STD-1474D
5.3.1.2.2 Digital recording. Sampling rate shall be a minimum of 160,000 samples/sec. The analog signal shall be filtered, using a Bessel type, with 40 kHz cutoff frequency. The roll-off rate shall be not less than 36 dB/octave.

It turns out that meters that are affordable (I know affordable means different things to different people) don't meet this requirement. For example the B&K 2270 is only 48,000 samples/sec (this meter is used by Silencer Shop). The Silencerco meter uses a National Instruments product that only does 102,400. Neither of these meters meet MIL STD 1474D requirements. Bummer, cause I wanted the B&K product. Larson Davis puts out a meter they claim is a silencer testing meter, but it doesn't meet the requirement either. Its around 51,200 samples/sec. Its the Larson Davis 831.

B&K says they have another option that will work and do 192,000, so I will look into that and depending on the price I will go that route. Probably over 20K would be my guess. The B&K pulse system is over 40K.

I just found it interesting that the meters that a number of companies and dealers have been pumping out numbers with are not only too slow to catch the peak of a suppressed gunshot, they don't meet the standards of military testing either. 90% of consumers will never know they are getting sold on sound reduction numbers that are bogus. On the low end (22 suppressors and quality 9mm suppressors) this won't usually be a huge issue, but on stuff that is borderline, these digital meters mentioned are producing numbers that could potentially make unsafe impulse noise seem safe. Not a good standard to set for the industry and the consumer.

\

I'm currently running the following...

B&K 4944-A Mic
4231 calibration
3052-13-3/0 LAN-XI Input module @ 102.4kHz


I've been happy with the results as they have been consistent and I can not only monitor dB but B Duration as well. It samples fast enough to catch the initial shot as well as the reflection off of the ground and nearby objects.

[Mongo]

I always wondered about designing suppressor muzzles to spoof the mil std test into yielding lower numbers by directionally controlling the sound impulse. You could make a suppressor sound quieter in testing yet actually be much louder to down range observers. I guess it comes down determining the goal of the sound testing, for down range noise or user observed sound levels.

[Grounded]

I always wondered about designing suppressor muzzles to spoof the mil std test into yielding lower numbers by directionally controlling the sound impulse. You could make a suppressor sound quieter in testing yet actually be much louder to down range observers. I guess it comes down determining the goal of the sound testing, for down range noise or user observed sound levels.

I think the end goal would be quieter report in a 180° sweep from the muzzle. Anything directed downrange and away from the shooter would be acceptable IMO.

[SpeedyR]

very interesting discussions about the standards. I have worked in the motorcycle industry (dynomometers for HP) and outdoor industry (FL-1 standard for lights, NFPA/ANSI testing for rescue equipment) and spent a lot of time looking at different issues.

It would seem that the ASA would be interested in creating an industry standard- similar to what the light industry did with the ANSI/NEMA FL-1 standard for light output. While not an ideal standard, it does at least get most on the same playing field. Like a dyno for horsepower though, being able to test things at the same time with the same equipment and compare results can usually yield some decent results. it's not always about the final numbers, but more about how A compares to B in the same situation. Even then it's hard to take into account things like tone when looking at results.

It would seem to have a few numbers from the testing would be ideal. First round number, then maybe a 5 string or 10 string averaged, etc. For hunters and other "one shot" users, the only number that would matter is the first round. For many the results after 5 or 10 rounds might be more important. or 30?

as for the sampling, the third link in one of the first posts (AHAAH report) showed the SPL for 5.56, 7.62 and a few other rounds in miliseconds.. didn't have time to read but might be interesting to convert/compare this info to the sampling for the standards and see if the measurements of the hardware is fine/detailed enough to capture the peaks...

http://www.arl.army.mil/www/default.cfm?page=351

but good to bring up some of these issues and consider the ramifications.

[rimshaker]

AAC actually listed their sound equipment in their old 2012 catalog:

B&K 3052-A-030 Pulse System set on "A" weighted and peak hold
B&K 2670 1/4" microphone

[T-Rex]

AAC actually listed their sound equipment in their old 2012 catalog:

B&K 3052-A-030 Pulse System set on "A" weighted and peak hold
B&K 2670 1/4" microphone

I believe SilencerShop uses the B&K 2270

[KiA]

pardon my ignorance, i can't get over the following thought:
if the sound impulse is too quick to require a sampling rate of >44khz, wouldn't that make it high frequency inaudible by human ear?

[doubloon]

Sound is pressure.

http://news.sciencemag.org/biology/2014 ... -your-ears

http://dujs.dartmouth.edu/biological_sc ... annot-hear