[Ethan Winer]

Folks,

People often ask which affordable microphones I recommend for measuring their room. To answer this question I tested ten popular microphones ranging from an inexpensive DIY model to a Josephson with Microtech Gefell capsule costing $1,800, and wrote up the results for this article: Comparison of Ten Measuring Microphones. Enjoy!

--Ethan

[Russtafarian]

Thanks Ethan.  Good information.

Any recommendations for a low cost phantom power preamp/mixer to interface a Behringer or Nady mic to a computer?

Russ

[TomW16]

Nice job Ethan.  I find the results tremendously helpful as I am building a home theater and the "Nady or Behringer microphones seem to be quite adequate" as you put it.  When it comes time to take measurements of the finished room I'll look into purchasing one of these microphones.

Thanks!  

Tom

[JeffB]

Wow, Thanks Ethan.
Fantastic contribution.

I have been complaining about the rising response in my Visaton B200 for the longest time.
This was based mainly upon seeing the graph put out by Visaton and confirmed with my digital radio shack SPL meter.
Now, I see the radio shack unit has a built in rising response.  So the Visaton's response might actually be better than I thought.
Although my ears still lead me to believe the response is rising.

[JeffB]

I also find it amazing how between 55Hz and 200Hz there is about 30db variation in signal.

[JeffB]

Why does the signal drop off so quickly below 55Hz?
I don't notice this at home with the RatShack Meter.

[Ethan Winer]

I also find it amazing how between 55Hz and 200Hz there is about 30db variation in signal.

This happens in all rooms. Most rooms are even worse! This is my home studio, and though I have a lot of bass traps there are only about 1/4 as many proportionally as in my living room home theater. But even with bass traps, this is what happens.

Why does the signal drop off so quickly below 55Hz? I don't notice this at home with the RatShack Meter.

The speaker we used is flat (-3 dB points) from 37 Hz to 21 KHz, so some of what you see is the fall of the broad peak below 100 Hz, and the rest is the fall-off of the speaker below 40 Hz.

--Ethan

[TerryO]

I also find it amazing how between 55Hz and 200Hz there is about 30db variation in signal.

This happens in all rooms. Most rooms are even worse! This is my home studio, and though I have a lot of bass traps there are only about 1/4 as many proportionally as in my living room home theater. But even with bass traps, this is what happens.

Why does the signal drop off so quickly below 55Hz? I don't notice this at home with the RatShack Meter.

The speaker we used is flat (-3 dB points) from 37 Hz to 21 KHz, so some of what you see is the fall of the broad peak below 100 Hz, and the rest is the fall-off of the speaker below 40 Hz.

--Ethan

Very informative topic, thank you.

I believe, and perhaps it should be emphasised, that your graphs are not the frequency response of the mics themselves, but the in-room response while measuring your speaker under identical conditions and for that reason are for comparative purposes.

I have used my 30 year old Radio Shack meter all along and while it isn't anyone's idea of the "best" I've found it adequate for the relatively minor tasks that I've used it for. Is there any reason why you couldn't create a correction table for the Radio Shack meter, if you could get a friend having a calibrated setup (Praxis?) to do a side by side comparison and create your chart or "table?" This would require to have a friend...with the equipment and the time to methodically do a step by step chart using what ever frequency generation source is chosen.

Best Regards,
TerryO

[Steve]

Very informative topic, thank you.

I believe, and perhaps it should be emphasised, that your graphs are not the frequency response of the mics themselves, but the in-room response while measuring your speaker under identical conditions and for that reason are for comparative purposes.

I have used my 30 year old Radio Shack meter all along and while it isn't anyone's idea of the "best" I've found it adequate for the relatively minor tasks that I've used it for. Is there any reason why you couldn't create a correction table for the Radio Shack meter, if you could get a friend having a calibrated setup (Praxis?) to do a side by side comparison and create your chart or "table?" This would require to have a friend...with the equipment and the time to methodically do a step by step chart using what ever frequency generation source is chosen.

Best Regards,
TerryO

Hi Terry,

Already been done several times. Here are the links.

http://www.digital-recordings.com/audiocd/radio.html

http://www.gearslutz.com/board/post-production-forum/230192-radio-shack-spl-meter-correction-table.html
 
http://www.audioasylum.com/audio/general/messages/49147.html

Notice the maximum variation is approx 11db on the analog graph and 8 db or less on the digital.

Take care.
Steve

[JeffB]

I took a look at these 3 links and the answers they provide are so different as to be useless.
They are also unclear as to whether the corrections are to be added or subtracted.
I consolidated the results here.  They are comma delimited for easy import into Excel.
I believe these numbers represent what should be added to the radio shack reading to arrive at the true reading.
All the values in the first link were negated, the other two links were taken as is.
The second and third link have good agreement below 3000Hz, but separate completely above that.
The first link doesn't correlate with either of the others.
Also none correlate to Ethan's graph.

The first list is
freq., first link, second link, third link.

The second list is
freq, second link, third link.

20,  N/A,7.5,6.2
25,  N/A,5,4.4
31.5,  N/A,3,3
40,  N/A,2.5,2
50,  N/A,1.5,1.3
63,  N/A,1.5,0.8
80,  N/A,1.5,0.5
100,  N/A,2,0.3
125,3.5,0.5,0.2
160,  N/A,-0.5,0.1
200,,-0.5,0
250,0.5,0.5,0
315, N/A,-0.5,0
400, N/A,0,0
500,-4.5,-0.5,0
630, N/A,0,0
750,-2, N/A, N/A
800, N/A,0,0
1000,1,0,0
1250,N/A,0,0
1500,-1.5,N/A,N/A
1600,N/A,-0.5,0.1
2000,1,-1.5,0.2
2500,N/A,-1.5,0.3
3000,-4,N/A,N/A
3150,N/A,-1.5,0.5
4000,-2.5,-2,0.8
5000,-4.5,-2,1.3
6000,-3,N/A,N/A
6300,N/A,-2,2
7000,-5.5,N/A,N/A
8000,-4,-2,3
9000,-3,N/A,N/A
10000,-2.5,-1,4.4
11000,-1.5,N/A,N/A
12500,N/A,0.5,6.2
14000,-2,N/A,N/A
16000,-4,0,8.5
18000,-2.5,N/A,N/A
20000,-1,1,11.2



20,7.5,6.2
25,5,4.4
31.5,3,3
40,2.5,2
50,1.5,1.3
63,1.5,0.8
80,1.5,0.5
100,2,0.3
125,0.5,0.2
160,-0.5,0.1
200,-0.5,0
250,0.5,0
315,-0.5,0
400,0,0
500,-0.5,0
630,0,0
800,0,0
1000,0,0
1250,0,0
1600,-0.5,0.1
2000,-1.5,0.2
2500,-1.5,0.3
3150,-1.5,0.5
4000,-2,0.8
5000,-2,1.3
6300,-2,2
8000,-2,3
10000,-1,4.4
12500,0.5,6.2
16000,0,8.5
20000,1,11.2

[Steve]

I think it is important to understand the graphs were presented to show the maximum deviation measured. A maximum 11db deviation is a far cry from 34 or 40db max variation, especially at mid and higher frequencies.

The first link, Audio-CD can be interpreted. From 125hz to 20khz there are pluses and minuses from true accuracy. For instance, at 250hz, subtract .5db from the reading. At 3khz, add 4db to the reading. At 16khz, add or subtract nothing.

The second link by Peterson consists of the frequency, meter read, the correction in db and the true SPL. (-.5db) means subtract .5db from the meter reading for the true SPL. The other figures not in parenthesis, such as 7.5 at 20hz means add 7.5db for the true SPL. At 2khz subtract 1.5db. At 16khz, add or subtract nothing to the meter reading.

"The second and third link have good agreement below 3000Hz, but separate completely above that."

There are some differences. The second link is for the digital meter while the third link is for the analog meter. At 5khz one says add 4.5db the other says subtract 2.0db. But they are within 6.5db.


"The first link doesn't correlate with either of the others."

The first link is not hugely different from the second link, and the second link is for the digital meter. At 20hz, add 6.2db. At 200hz, no need to add as it is accurate. At 20khz, add 11db to the reading. 11db is the largest deviation measured among the three links.

No they will not correlate to Ethan's graph. The properly working SPL meter, as Ethan contends for the past three weeks,  is not going to have a 34 (analog) to 40db (digital) variation.

TNT Audio States
"Being an inexpensive meter, its accuracy is just fair (± 2 dB @ 114 dB). Also, a wider frequency response would have been useful for us HiFi-nuts."
Spec "Accuracy: ± 2 dB @ 114 dB

The links are for lower spls.

"To verify the accuracy of the newer version, I compared it to an Audio Control 3050 RTA, the same one Tom Nousaine used for years in his test reports for Car Stereo Review until he bought MLSSA. The overall SPL accuracy of all three of my Radio Shack sound level meters -- 2 analog and 1 digital 33- 2055-- were within 1 dB of 75 dBC SPL compared to the Audio Control.... Make sure your meter is set to C weighting. The digital meter and my second analog meter (6 years newer than the test unit in 1996) were the same as the test SLM. Response below 25 Hz done by Eric Busch with sine waves and B&K equipment at Dave Clark's DLC Design in Michigan.

Michael Sims
Prairie State Audio Construction Society

Ethan has had a problem explaining some measurements posted at Stereophile forum three weeks ago and had to fall back on the Radio Shack meter being grossly inaccurate as his explanation.

[Ethan Winer]

Is there any reason why you couldn't create a correction table for the Radio Shack meter

You could do that yourself using my graph, but with a few caveats. First, the RS meter is fine below about 1 KHz. Yes, it falls off a little below 40 Hz, but it's still within a few dB of the good microphones. It's probably not practical to try to improve the response above 10 KHz because it falls off so quickly. More to the point, most people need to measure bass frequencies only, and for that the RS meter is okay as is. And if you do need more accuracy, the $40 Nady or $50 Behringer are very close to the high-end microphones.

--Ethan

[Ethan Winer]

I took a look at these 3 links and the answers they provide are so different as to be useless.

Exactly. I have to think that most of these "correction" graphs are produced by people who have no idea how to measure in-room speaker response. So they buy a Radio Shack meter, measure their speaker, and ASSume what they measure is The Truth and feel the need to share that with the world.

Before I did this test I had only a comparison of my own RS meter to my own AKG calibrated microphone. But now that I have two RS meters to compare - made 20 years apart - it's clear they track very well at low frequencies, proving all those contradictory correction curves are bogus.

--Ethan

[Steve]

I think it is pertinent to examine some other evidence before condemning the other links. There is more to the story than others simply requesting a comparison of mics.

After I presented evidence to the contrary to his "typical" room measurements at Stereophile's entry level forum, string titled "accuracy" (discussion actually started in "upgrade paths" page 4, some 3 weeks ago,

Ethan suddenly recalled that he used a Radio Shack spl meter and that it was the cause of the gross inaccuracies of those measurements, esp at mid to high frequencies, even though he had tested the mic to only 800hz.

So it is to his advantage to present the Radio Shack spl meter as poorly as he can in an attempt to nullify the evidence I presented there concerning his measurements.

it's clear they track very well at low frequencies, proving all those contradictory correction curves are bogus.

We only have your word on it, so no proof the correction curves are bogus.

[BradJudy]

The results don't surprise me.  I had a calibration file created (by a pro) for my old ECM8000 and have had to chance to see ECM8000 calibration files for a few other mics.  A great deal for a quite flat mic - perfect for most hobbyist measurement. 

FWIW: I was using an ART MicroMix pre-amp for a while.  I sold the mic and pre-amp a while back. 

The RadioShack meter is great for what most people use it for - level matching speakers.  That's what I use mine for.  It isn't a great choice for people wanting to do EQ work though.  The correction files can help, but I have often wondered how consistent the RadioShack meters are.  It would be interesting for someone to do something similar to Ethan, but with multiple RadioShack meters bought at very different times and/or locations. 

[Steve]

Hi Brad,

I do not think they are that far off, but certainly not +/- 2db. The links all had their own meters (and there are others) so one can get some sort of idea of the response.

But for any newbies, believe me, the RS meters are not even close to 40db off as Ethan "measures" it to be. By the way, he also tried to change setup conditions of his "typical" room measurements as well, after I presented the evidence on Stereophile.

Anyway, if one notices the slope of the roll off the meter above 5khz, according to Ethan's measurement the drop off is around 12db per octave, which is basically "free falling".

Of course this is nonsense as there is compensation built into the RS meter at both low and high frequencies. So, he could not possibly be measuring what he claims.

Another point is that, on Stereophile forum, Ethan never claimed his use of the RS meter until after I presented evidence of his grossly inaccurate measurements/graphs.

So according to Ethan's arguement of his using the RS meter, he was content to let the public accept his measurements and graphs as accurate, even though he knew the measurements/graphs were grossly inaccurate. Glad I brought it up the evidence on stereophile or the public would still think they were given true/accurate information.

But of course, the point here is that the meter is fairly accurate according to several sources. Nothing to shout about, but not 15 or 40db off either.

[youngho]

Steve, no offense, but you're starting to come off as a bit of a stalker. The whole "I was fast asleep and snoring" nonsense in the Stereophile threads...a little weird...

Incidentally, the TNT link seems to simply quote the RS specs, without any data. No description of methodology was provided. Furthermore, you left off part of the Michael Sims quote here http://www.hometheatershack.com/forums/rew-forum/86-interesting-info-rs-spl-meter-correction-values.html, so your quote is misleading. The key part is probably "I then checked the frequency response, comparing it while set to C weighting and slow, with pink noise, 1/3 octave band by 1/3 octave band, to the Audio Control RTA in the SPL mode. Using the same official PSACS calibrated PSB loudspeaker and a pink noise CD, I made a calibration curve that can be subtracted from the results obtained by the Radio Shack in your living room to obtain accurate, repeatable measurements for about $60, including pink noise CD." The included calibration curve suggests deviation of significantly more than 11 dB between 10-100 Hz.

The second link included an article from Audio Xpress (http://www.audioxpress.com/magsdirx/...a/koya2811.pdf) that I found interesting, particularly because it completely contradicts the TNT link. I wonder if some of the disagreement may stem from different uses of "deviation," since +/- 10 dB means a range of 20 dB. In any case, I wonder if we can all agree on the following:

1. The RS SPL meters are useful for comparing SPLs at a given frequency or for broadband content, i.e. level matching
2. The RS meters are not nearly as useful for measuring frequency response
3. Objective measurements of the RS meters suggest that they may be worse than you say, though perhaps not as bad as Ethan says

Young-Ho

[Steve]

Steve, no offense, but you're starting to come off as a bit of a stalker. The whole "I was fast asleep and snoring" nonsense in the Stereophile threads...a little weird...

I Actually stated

I know Alex. I was fast asleep and snoring when Ethan woke me up with his antics and manipulative/deceptive behaviour again.

A little different than your partial version. Putting yourself in Ethan's camp is not a good idea. Now Let's see the rest of the story and get the truth.


Read Alex's earlier post Young, on page 3.
http://forum.stereophile.com/forum/showflat.php?Cat=0&Board=TheEntryLevel&Number=43995&page=1&fpart=3

Let's put this thread to bed.

I then signed off, quitting. As soon as I posted I am alseep and snoring,

Ethan again posted and started things going again.
 
Then Alex posts

here we go again.

The whole discussion started at Entry level forum, Upgrade Paths, page 4, Ethan broke into the conversation Clifton and I were engaged in. Then Ethan started with the insinuations. We were off subject so string "accuracy" was started, and continued.
 
It is amazing how the Ethan followers always seem to reverse things.
 

Incidentally, the TNT link seems to simply quote the RS specs, without any data.

Check out RS spl meter page and related links. Did I miss the data is given there? I do not see any spec +/- 2db on RS webpage and subsequent links. Here is the main link to the RS meter page.
http://www.radioshack.com/sm-digital-display-sound-level-meter--pi-2103667.html

 The meter is made for more than one company, so I suspect the specs came from them since at least one company lists that spec. And you failed to address the first three links. So what is the big deal Young?

I don't see you questioning a ridiculous 40db total variation from mid to highs measured by Ethan.

No description of methodology was provided. Furthermore, you left off part of the Michael Sims quote here http://www.hometheatershack.com/forums/rew-forum/86-interesting-info-rs-spl-meter-correction-values.html, so your quote is misleading. The key part is probably

"I then checked the frequency response, comparing it while set to C weighting and slow, with pink noise, 1/3 octave band by 1/3 octave band, to the Audio Control RTA in the SPL mode. Using the same official PSACS calibrated PSB loudspeaker and a pink noise CD, I made a calibration curve that can be subtracted from the results obtained by the Radio Shack in your living room to obtain accurate, repeatable measurements for about $60, including pink noise CD."

So? How is that misleading? In the bass, the meter will not have sudden peaks and nulls.  Lower frequency comb filter problems are caused by the room, not the meter/mic. And 7db goes right along, agrees with my other links, while Ethan disagrees with Michael Sims by stating

it's clear they track very well at low frequencies, proving all those contradictory correction curves are bogus.

(which proves nothing)

included calibration curve suggests deviation of significantly more than 11 dB between 10-100 Hz.

What included calibration curve? And who cares about 10hz? Ethan did not even measure to 10hz. Slick trick extending it to 10hz Young, since that is where the most deviation is. At 20hz, the error is much less.

We just saw Ethan claim that the RS meter is accurate in the lows, you did not disagree.
Neither have you addressed the 40db total deviation Ethan claims to have measured from mids to highs.

The second link included an article from Audio Xpress (http://www.audioxpress.com/magsdirx/...a/koya2811.pdf) that I found interesting, particularly because it completely contradicts the TNT link.

You are kidding, right. And what spl level for figure 2???? Check the setup and testing methods. It is laughable. And Ethan claims the lows are accurate. So is ethan wrong on both and highs? I would say so. One thing for sure though, no 40db total deviation from mids to highs.

I wonder if some of the disagreement may stem from different uses of "deviation," since +/- 10 dB means a range of 20 dB. In any case, I wonder if we can all agree on the following:

 +/-  is one of my peeves that manufacturers use. How many times have I stated  "total deviation", which includes both + and -. However, the key is the rate of decline, 12db per octave above 5khz. It is a joke.

Your points are relative at best.

By the way, there is alot more info I presented, like the 12db/octave roll off that you failed to address. The RS meter is not even close to what Ethan claims to have measured.

3. Objective measurements of the RS meters suggest that they may be worse than you say, though perhaps not as bad as Ethan says

Don't think so Young. There is no "perhaps" about it.

Maybe next time we can get a realistic post.

[Ethan Winer]

Steve, no offense, but you're starting to come off as a bit of a stalker.

No kidding, and this is getting pathetic. In truth, I have no idea what Steve is objecting to, or why he feels the need to discredit me. Trying to get a straight answer out of him is a moving target, and I have asked Steve directly exactly what he feels is in error in my posts. The root issue seems to be that Steve doesn't understand that all normal size rooms have 30 dB peak / null spans at bass frequencies. But rather than actually test a few rooms with software and post the results as I suggested, he continues to throw stones at me. First at Stereophile's forum, and now here. I've seen this type of behavior before, but I'll never understand it.

--Ethan

[Steve]

Steve, no offense, but you're starting to come off as a bit of a stalker.

No kidding, and this is getting pathetic. In truth, I have no idea what Steve is objecting to, or why he feels the need to discredit me. Trying to get a straight answer out of him is a moving target, and I have asked Steve directly exactly what he feels is in error in my posts. The root issue seems to be that Steve doesn't understand that all normal size rooms have 30 dB peak / null spans at bass frequencies. But rather than actually test a few rooms with software and post the results as I suggested, he continues to throw stones at me. First at Stereophile's forum, and now here. I've seen this type of behavior before, but I'll never understand it.

--Ethan

Quite simple.
All the viewers have to do is check out

page 4, and Ethan's obsession with going after me.
http://forum.stereophile.com/forum/showflat.php?Cat=0&Board=TheEntryLevel&Number=43089&page=2&fpart=4

Then

http://forum.stereophile.com/forum/showflat.php?Cat=0&Board=TheEntryLevel&Number=43995&page=0&fpart=1

where I post the evidence concerning Ethan's graphs of 35db total variations (Ethan now says 30db) and how one can mimick his measurements. See whether Ethan or I am telling the truth.

Oh, and don't forget the attack on me at (I never even posted in that string, or forum before.)

http://forum.stereophile.com/forum/showflat.php?Cat=0&Number=43592&page=0&fpart=3&vc=1

I quoted his comments and responded at

http://forum.stereophile.com/forum/showflat.php?Cat=0&Number=43592&page=0&fpart=4&vc=1

By the way comparing Ethan's two graphs, one in Stereophile and the other on his webpage shows quite different wideband results. And we are not talking comb filtering. And they both used the same room, same mic, same speakers. Distance/placement is rather irrevelant since we are talking average response.

So you screwed yourself again in front of the public. You have no one to blame but yourself Ethan.