Recent discussions have got me thinking - one of the biggest problems on these forums is defining "quiet". A HDD that one person considers quiet may be relatively loud to another. One person may get a very quiet sample, while one may get a louder one.

The obvious solution is to use a sound level meter, but good ones are expensive. Is there such a thing as an accurate and cheap meter for low-level monitoring, as required for SPCR? If an affordable and readily available model could be found, it would make life for dedicated silencers a lot easier.

Hello

I have been thinking about this too.
I havent really been able to find much.

Conrad.de has these

None of the can measure below 30dB though.

The problem is you want to measure down to 15 or 20 dba but no meter I've seen for sale to mear mortals is reliable for even measuring 30 dba. It looks like the Nady DSM-1 is the best value for the home user but even then you won't be able to measure anything under 35 dBA with any great certainty.

There is no such thing as a cheap SPL that measures SPCR levels reliably. I spent the last little while doing tons of Google searches and this is what I came up with.



Bruel and Kjaer makes sound level meters that I can't even find prices for easily. I'm sure there are plenty of Brands and models not present on this list.

I too have been looking online for a solution, particularly a DIY one. Unfortunately, there do not seem to be any useful hobby projects on which to base a design.

Considering how cheap and accurate other kinds of meters are, it's surprising that SPL meters are so difficult and expensive to make.

I'm not sure how you think this will help? so if one person on the forums says something is loud and another says it is quiet, the SLM will settle the matter? given that the measurement circumstances would have to be exactly identical to be comparable (background noise, distance from source, same size room and treated/untreated walls) the lack of cheap, low-dB meters is not the only obstacle.

i looked it up once in my area... if you look hard enough, you can probably rent one...

Lets say I have 10 PSUs and 30 fans laying around my apt. I don't trust my memory to remember which one was the quietest after comparing more than two or three so a SPL would give me a way to quantify the differences and put them on paper. That alone is worth it for me.

Of course the subjective result could override the SPL measurement but I can at least use the SPL to narrow down the list in a objective fashion before I worry about subjective comparisons.

It might also help for those "should I RMA this" threads. If you used a SPL meter in conjunction with a wattage meter and knowledge of the case and fans in use along with some temp data you could make a more informed decision about a system you can't hear.

Sure you might get someone that won't post enough detail but you might get these two scenarios

1: My Corsair makes a buzzing noise and my SPL says ambient is 15 dBA

2: My Corsair makes a buzzing noise and my SPL says ambient is 45 dBA

I'd tell person 1 they may just be too picky. I'd tell person 2 they clearly need to RMA the PSU.

The problem is with a cheap Radio Shack SPL is both people would say they'll say their ambient is 40 to 45 dBA even if it isn't. You don't get accurate, or even useful information from a SPL that doesn't go below 30 dBA when you are discussing SPCR level components.

And without an SPL you get people comparing items with made up dBA numbers that have no objective basis. They mean well but you just can't trust every joe that says x is about 17 dBA.

I've found some more meters to add to the price lists and a primer article on understanding their use, ratings, and such

http://www.prosoundweb.com/studyhall/la ... /slm.shtml



The standouts are



There may be better choices but if so they don't show up well on the google searches I chose.

For a casual user the Scosche SPL1000 or Nady DSM-1 might be enough for to settle an argument or satisfy curiosity.

For a more serious use the Extech 407738 claims to do under 30 dba for under $300.

And if I were trying to keep up with SPCR I'd be finding prices on all those stupid SPL manufacturers that don't sell to the retail market. God only knows who makes what at what price. But you can definitely get something that will read under 20 dba if you try hard enough and have plenty of cash.

It's really down to the microphone used; Cheap sound level meters tend to use simple electret microphone capsules which, while adequate for normal levels, are too intrinsically noisy for low level measurements.

The problem most of us would have, is finding somewhere quiet enough to make meaningful measurements (Mike's lucky to have ambient levels below 25dB). For instance, here in my workshop on a quiet Sunday evening, no machines running, breathing gently, my B&K 2206 is reading ~30dBA (30dB is the lowest setting - the scale actually goes 10dB above and below this). If I inhale/exhale too fast or try to type, the meter hits the end stops (>40dB).

20dBA(spl) is extremely quiet.

Finding a quiet room can be done for most if they really want to test. If you had to you could use a large closet or a large bathroom that has no windows.

To put the levels you want to measure in perspective:



If your ambient is 30 dBA you shouldn't be able to hear your hard drive right?

wrong, with the pc running (with two 10,000 rpm raptors - on the floor beside me) the ambient level increases by about 2dB to ~32dBA, the whining of the Hdd/fans are plainly audble .

As I've said before on this forum, ambient noise doesn't mask sound.

It's the lack of a SPL in my hand that makes that concept foreign to me.

I easily forget that sound is additive.

I need to have a SPL around for a year or two so I can test all the noises I think about on a regular basis and get rid of my misconceptions...

Also, noise at different frequencies does not mask as well as noise at the same frequency. Thus, two fans will only be minimally louder, but a fan and a HDD at the same dB level but different frequencies will be much more noticeable.

"Coctail party effect": you can track, and understand, a conversation that's much lower than the ambient sound level at a large party. There's an evolutionary reason for this: folks with better discrimination of sound sources survived better when predators (2 or 4 legs) snuck up on them.

You are a descendant of a long line of folks with above-average sound discrimination (in their generation). Every generation, the average discrimination got just a tad better - which is why you can hear a 20dBA disk drive in a 30dBA environment quite easily.

To add to the list of SLMs with low A-wtd noise levels:

The Larson Davis models 812 and 820 have a noise base of 17.5dBA.

The Casella CEL 430.A1 has a noise level of 16.5dBA.

All are Type 1 SLMs and not cheap. All use 1/2" free-field microphones with 200V polarization. At the moment the dollar value is moving around a lot, so let's just say these units sell for very roughly $3300 for a basic model with no calibrator etc.

Free-field mikes are what you want; pressure microphones are what you don't want. Trust me!

interesting as that's what anandtech use and they claim a 15dB(A) noise floor:

http://www.anandtech.com/casecoolingpsu ... i=3021&p=7

Googling "microtech gefell" reveals they make studio microphones and amplifiers, not sound level meters. Anandtech built its own-design ~1 meter cubed "quiet" enclosure. This stuff will not produce data you can present in court.

"Type 1" and "Type 2" are legal standards. Data taken with these SLMs can be presented in court.

What Anandtech has is not a sound level meter, as SLMs are generally defined. Not to say it isn't good stuff. Just doesn't meet any legal standard.

Good find (the linked article about SLMs) and good summary in that post.

I doubt there's been much evolutionary gain in this regard the last hundred years or more, certainly none in the last 50. Human spaces are mostly devoid of non-human predators... and with the latter, it's usually less about physical sneakiness. Not if any of the extreme crime Am. TV shows are to be believed.

Mike, I'm talking about evolution over the past 100+ million years. Homo sap descended from a predecessor species, which in turn was descended from the previous predecessor species etc. The ability to detect approaching predators has been important, evolutionary speaking, for over a hundred million years. Honest.

That's likely 7+ million generations. A 0.001% survival advantage in each generation, repeated 7 million times, works wonders for hearing acuity.

The last hundred years is nothing in evolutionary terms!

edit: My reason for choosing 100+ million years:

Our hearing acuity began evolving when three conditions were met: when our ancestor had legs, ears, and a need to avoid predators. We are mammals, evolved from the mammals that we know from the fossil record were contemporaneous with the dinosaurs.

I'd just like to add that another cheap trick to get very high sensitivity is to use a large(8-12 inch) 8 or 16 ohm speaker as a microphone. You can pick up sounds from the entire house with a setup like this.

(measuring it accurately will require an oscilloscope or similar, though)

I've updated the price list.

The SPL primer seems to be gone from the original site so here is an archive copy

http://web.archive.org/web/200603180612 ... /slm.shtml



The standouts are



There may be better choices but if so they don't show up well on the Google searches I chose and/or they are new as I haven't done the brute force searches since my first post on this subject.

For a casual user the Scosche SPL1000 or Nady DSM-1 might be enough for to settle an argument or satisfy curiosity.

For a more serious use the Extech 407738 claims to do under 30 dba for under $300.

And if I were trying to keep up with SPCR I'd be finding prices on all those stupid SPL manufacturers that don't sell to the retail market. God only knows who makes what at what price. But you can definitely get something that will read under 20 dba if you try hard enough and have plenty of cash.

Thanks to those in the thread that pointed me to Norsonic, Larson Davis, and Casella which I had no prior knowledge of.

I just rate things on a personal 1-10 scale, sort of how MikeC does on the recommended page.

If you have 30 fans. just rate them. this sounds like an 8-, this sounds slightly better, a solid 8, this next one is basically inaudible, so a 9+. So if you get to like fan 7 or so, and cant remember if its better or worse than something, you can go back and compare to something you gave a ranking to, to get a rank for it. So say you listen to fan 10, but cant remember if its louder or quieter than fan 1, which you gave an 8-, compare it to fan 1. If its louder, give a lower score, or compare it to the next lowest scored fan. Etc etc.

I mean your ultimately going to have to do this process ANYHOW even with meassured SPL's since SPL doesnt show everything. SPL is fairly meaningless except for reference only to other items that were measured by the same person and gear under the same conditions. You cant compare them to anyone else's SPL readings since you'll never get all the ambient conditions the same.

The SPL primer seems to be gone from the original site so here is an archive copy

http://web.archive.org/web/200603180612 ... /slm.shtml

Again, good work to keep this uptodate, dhanson865!

The Extech 407738 you mentioned seems decent, and I've seen it for as low as $250. For tech geeks, it's not a bad tool to have handy, you'll find all kinds of uses for it once you have one on hand. (Not to mention the people who'll want to borrow it for this that or the other...)

Norsonic and Wartsila makes quality meters also.

Is an accuracy of +-1.4dB considered "decent"? It's poor enough to provide wide margins of error when computing the difference between two sounds <5dB apart. (For example calculating the contribution from source X on top of the background noise.)

BTW, what's the accuracy of the equipment you use in the anechoic chamber?

Cheers
Olle

The accuracy of the Nady DS-1 is worse: + 1.5dB (under reference conditions)

Same goes for the Galaxy CM-140 (1.5dB). Check out the graph at soundandvisionmag

Protek SL1700 says ±1.5dB as well.

I don't eve see an accuracy claim in the specs for the Quest Technologies units

I doubt you'll find better accuracy in any of the sub $1000 units

That graph on the review of the CM-140 makes me wonder if the ±1.5dB is not per overall measurement but on a frequency by frequency basis. As in if you measure the same 700hz noise 3 days in a row the reading wouldn't change but if you measured a 700hz noise and a 6000hz noise separately the accuracy at those frequencys would differ.

If so then I'd be happy to have the Extech 407738 even at ±1.4dB accuracy. Repeatable non varying inaccuracy can be mapped and corrected for manually.

I think it is still good enough for non-critical noise measurements. MEasurements from multiple positions/distances can help reduce error. The lab gear -- the main source error would be users (ie, not setting the correct distance, ignoring or not hearing noises that might impact the readings, too many reflective things in the chamber, etc) -- and maybe the mic calibrator, which I've actually checked against several references as being better than 0.5 dB accurate. (See http://www.silentpcreview.com/New_Audio_Test_Gear_2008 ) The rest of the gear is probably at least 0.2 dB accurate.

I've updated this list to include newer PSUs measured in the anechoic chamber



I think this shows why SPCR users wouldn't benefit much from any meter that can't measure below 30 dBA. If your sound meter can't tell the difference between a 1600 RPM fan and a 800 RPM fan at 1 meter what good does it do you?