Natural Science Forum / Physics / Acoustics / October 2005

No surprise.

Not really. Beranek has made the NCB curves down to 16 Hz.
for NCB-25, we have:

Hz    dB
16    81
31    66
63    51
125    43
250    35
500    30
1000    27
etc.

    The RC curves are bogus, IMHO when it comes to evaluating lowest and
highest frequency noise because they badly stray away from the "single
number" philosophy. Typically at low frequencies, a noise that is rumbly
(as you have), will rate as RC-25(R) while the NC-scale will rate it as
NC-40, for instance.

My rationale these days is the following:

1- Acousticians love NC (or NCB)

2- Mechanical engineers love RC.

3- I know that when they stick to designing to the RC "curve" (really, a
5 dB/octave straight line, though this has recently changed in new ANSI
S12.2), so that the ME's will over design and thereby fend off trouble
at very low and very high frequencies.

4- When I arrive on-site and measure the results, I am likely to get
what we want.

    Angelo Campanella

Herb,

First recognize that a major part of the complaint is not because of the
level  of the sound but because of the beat.  That is, if the sound measured
the same but did not contain the musical beat, it would be much less
objectionable.  You could have a spectrum that is perfectly acceptable based
on the rating curves but still very annoying because of the musical beat
characteristic.

The NC and NCB curves basically assume steady sound levels that are not
pulsating.  One of the reasons that the RC curves are more restrictive at
low frequencies is because the developer had experienced a lot of problems
with HVAC systems that produced pulsating sounds at low frequencies.  He was
trying to compensate for this, but then the RC curves are too restrictive at
low frequencies if the sound is not pulsating.  More recently, Schomer has
proposed the RNC rating system that uses curves more like NCB, but with a
tangency method, and a penalty adjustment if the low frequency sound is
pulsating.  You might look into that.

However, the problem is that if the musical beat is heard, it is very
attention getting and irritating.  The goal is not just to get the music
down to some criteria curve, but to determine a reasonable level of
background sound by such a criteria curve, and then get the music down to
the point that it would be masked by such background level.  Thus, the music
needs to be significantly lower than the acceptable level of background
sound.

This problem isn't directly assessed in the norwegian regulation.
However, the reference standard says that music from bars, restaurant,
etc, in apartments should be evaluate as noise from mechanical
installations.

The noise limits are Leq<25 dBA, Lmax<27dBA and Lmax<47dBC. That's for
classs C apartement, the minimum limits. For apartements with improved
acoustics, class A, thoses limits are reduced by 5dB (20, 22) and it
says that it should not be any disturbing frequence when the 1/1-octave
noise levels are plot against NC- og RC-curve. For NC curves, that
means that the result curve must be 5dB lower than the noise limit in
dBA. hence, NC15 for class A. For Rc-curves, the noise level must not
exceed the curve by more than 5dB in frequences below 500Hz.

Hope this will give you some relevant guidelines.

Alain Bradette