naturerecordists
[Top] [All Lists]

Re: testing mic self-noise

Subject: Re: testing mic self-noise
From: Klas Strandberg <>
Date: Wed, 28 May 2003 12:56:01 +0200
I wrote, about two years ago to this list:

"This is how noise is measured: 

You need a special noise meter instrument, containing a filter normalized to
db(A) or (CCIR) 468
 
1/ In a controlled way, the microphone is exposed to a sound pressure of 1
pascal = 94 db SPL. 
2/ The meter instrument is set so that the db meter shows 94 db.
3/ You switch off the 94 db sound pressure and measure the remaining noise,
through the wanted filter, db(A) or (CCIR) 468.

This remaining noise is the inherent noise of the microphone, measured db(A)
or CCIR.

Measuring a standard 10 mm electret, - without filters - the meter will
probably stop around some 35 db. 
But this "un-weighted" measurement is not used very much. Instead, the noise
is filtered, either db(A) or CCIR.
After this filtering, the meter will probably stop at some 23 db(A) or  30
db (CCIR) 486. How the two figures differ, depends on the high frequency
content of the noise. 
The differences between the three measurements, 1/ un-weighted 2/ weighted
db(A) and 3/ weighted CCIR 486 depends on the frequency spectrum of the
inherent noise which the microphone performs. 

If the inherent noise mainly contain low frequencies, the differences
between db(A) and CCIR tend to be small. The "hissier" the noise is, (the
more audible it is for the human ear) the bigger differences you get between
the 2 measurements

Then a problem: The Japanese etc insist on giving noise figures in terms of
"signal-to-noise ratio". This is very tricky, as those figures many times
are measured and filtered differently. When comparing "signal-to-noise
ratio" figures between different microphones, you also have to make sure
that the measuring methods and filtering have been the same. Personally, I
never trust such figures.

Klas.

------------

At 22:34 2003-05-27 -0500, you wrote:
>>cmursic wrote:
>>>  Thank you, Walter, for a very erudite explanation.  As I suspected,
>>>  such measurements are beyond my means.  But I'll try the 
>>>  quick "pillow" method you describe with my Rode NT4 and see what I
>>>  get.
>>
>>It would be a waste of money to set all that up unless you were a
>>manufacturer or such like. In the case of the NT4, the specs include
>>what they got.
>>
>>The pillow method will get you most all you really need. Once you have a
>>mic the published specs get to be a lot less important than what you
>>find just using it.
>>
>>I'll be interested to know what sort of sound the NT4's self noise is.
>>That's something it's very hard to get info on.
>>
>>Walt
>>
>>
>
>>
>
>
>Hi cmursic--
>
>Great explanation Walt! It's going to the FieldMicList archive!
>
>To accelerate what you can learn about self noise, get at least one 
>other mic with a known noise rating, two more are even better.  Go 
>outside when the ambience sound level is low like in the middle of a 
>windless night (~30-40 dB here in Milwaukee) and record just the air 
>for about 15 minutes each. Set mic record gain at max because the 
>noise introduced by the preamp is constant and hi to full gain is 
>when the noise problems really show. Transfer them all to digital 
>files, listen to them back to back and perhaps use "find peak," to 
>measure the relative saturation levels. Some mics are a lot hotter 
>and this is good to note.  Then make copies of the files, pick a 
>section from each when the sound level is the lowest and crop and 
>normalize this section in all three files to -20dB and play them back 
>on your normal, reference monitor speakers at comfortable room level. 
>You'll have no trouble hearing differences. The broadband width 
>"air"also reveals the transparency-- how much you can hear into the 
>space.
>
>from earlier in the string, Walt wrote:
>>A desirable sound for self noise
>>would be a smooth hiss, but some mics sputter or crackle. With a smooth
>>hiss, the self noise can actually be audible and not upset the listener.
>
>Very true. Narrow bands right around 8K are pretty common. Use a 
>narrow peak or notch parametric equalizer to find these spots in all 
>three files.  These can be removed effectively with successive very 
>narrow notches-- avoid a fat single notch if the actual noise is in a 
>narrow band (~200-500Hz wide). The sputtering/uneven stuff from some 
>mics falls down to 3K and  is quite hard to remove without really 
>making the recording sound dull.  The Rode NT3's I've used have some 
>sputtering noises in this range and they're rated at 16dB(A)*.  I've 
>seen some mics with curious narrow noise bands as low as 4700 Hz.
>
>I'm quite sure self noise affects transparency in the low range too 
>when field saturation levels in the recordings are low (~<6%). Noise 
>in the low range makes its hard to balance the acoustic space 
>realistically, especially in relation to individual sounds in the 
>lower mid range like the human voice. Lots of folks roll-off the 
>bottom end but the space pretty much disappears and its an important 
>part of the field experience.  I recently found that I could 
>significantly boost the 125-600 Hz range in a recording made in a 
>quiet location to make a  tinny frog chorus sound more natural and 
>less harsh on the ears.  The boost added warm "air" to recordings 
>when made with mics in the 16dB(A) self noise range, and worked 
>really well with some mkh-20 recordings too  (10dBA). But when I 
>tried the boost with recordings made with my Crown PZM's (23dBA), it 
>made instant grunge.
>
>The hiss does jump out first, but if you compare clarity at other 
>frequencies, you may notice this too.
>
>Rob D.
>
>* The sensitivty of the Rode NT3  is 12dB(A), about 4dB less than 
>some other mics with the same self noise rating I regularly use. 
>Whenever the record gain is high, its easy to hear this difference in 
>sensitivity because its translates into 4dB? of additional noise when 
>the sound is played back at final mix level.
>
>
>To unsubscribe from this group, send an email to:
>
>
> 
>
>Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 
>
>
>
Telinga Microphones, Botarbo,
S-748 96 Tobo, Sweden.
Phone & fax int + 295 310 01
email: 
       



________________________________________________________________________
________________________________________________________________________

<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU