# dB and dB(A)

Started by October 3, 2011
```Hello forum,

I understand what dB and dBm are. I am still a little confused about dB(A).

I know that the sensitivity of the human ear to sound depends on
frequency: two sounds of the same sound pressure but different frequencies
may appear one louder than the other. Humans hear high frequency noise much
better than low frequency noise.

An A-weighting filter deemphasizes low frequencies. Decibels measured using
this filter are A-weighted and are called dB(A).

But how can I get a practical sense of what, say 30 dB(A), 20 dB(A), 60
dB(A) are? Is there an easy way to determine how much bigger is a certain
dB(A) with respect to another?

For instance, 3 dB is easy: it means twice as much power. 10 dB means 10X.
20 dB means 200X....etc...
But 10 dB(A) means twice as loud, I think, in terms  human loudness
perception. IF the noise limit is 80 dB(A), what if a noise meter measures
100 dB(A), i.e. 20 dB(A) above the limit?  In terms of loudness, how can I
quantify it? It 2X, 3X, ...times noisier than allowed?

Is there an easy way to convert from dB(A) to dB?

thanks
fisico32
```
```
"fisico32"  wrote in message
news:7bKdnQ2DOdin3RfTnZ2dnUVZ_g2dnZ2d@giganews.com...

Hello forum,

I understand what dB and dBm are. I am still a little confused about dB(A).

I know that the sensitivity of the human ear to sound depends on
frequency: two sounds of the same sound pressure but different frequencies
may appear one louder than the other. Humans hear high frequency noise much
better than low frequency noise.

An A-weighting filter deemphasizes low frequencies. Decibels measured using
this filter are A-weighted and are called dB(A).

From Wikipedia, http://en.wikipedia.org/wiki/A-weighting

"A-weighted decibels are abbreviated dB(A) or dBA. When acoustic (calibrated
microphone) measurements
are being referred to, then the units used will be dB SPL referenced to 20
micropascals = 0 dB SPL"

Also see http://en.wikipedia.org/wiki/Decibel

But how can I get a practical sense of what, say 30 dB(A), 20 dB(A), 60
dB(A) are? Is there an easy way to determine how much bigger is a certain
dB(A) with respect to another?

http://en.wikipedia.org/wiki/Sound_pressure#Examples_of_sound_pressure_and_sound_pressure_levels

For instance, 3 dB is easy: it means twice as much power. 10 dB means 10X.
20 dB means 200X....etc...

20 dB would be 100 times, not 200...

But 10 dB(A) means twice as loud, I think, in terms  human loudness
perception. IF the noise limit is 80 dB(A), what if a noise meter measures
100 dB(A), i.e. 20 dB(A) above the limit?  In terms of loudness, how can I
quantify it? It 2X, 3X, ...times noisier than allowed?

Why do you say this?  3 dBA is twice as loud

Is there an easy way to convert from dB(A) to dB?

thanks
fisico32

```
```On 10/3/2011 4:07 PM, fisico32 wrote:
> Hello forum,
>
> I understand what dB and dBm are. I am still a little confused about dB(A).
>
>   I know that the sensitivity of the human ear to sound depends on
> frequency: two sounds of the same sound pressure but different frequencies
> may appear one louder than the other. Humans hear high frequency noise much
> better than low frequency noise.
>
> An A-weighting filter deemphasizes low frequencies. Decibels measured using
> this filter are A-weighted and are called dB(A).
>
> But how can I get a practical sense of what, say 30 dB(A), 20 dB(A), 60
> dB(A) are? Is there an easy way to determine how much bigger is a certain
> dB(A) with respect to another?
>
> For instance, 3 dB is easy: it means twice as much power. 10 dB means 10X.
> 20 dB means 200X....etc...
> But 10 dB(A) means twice as loud, I think, in terms  human loudness
> perception. IF the noise limit is 80 dB(A), what if a noise meter measures
> 100 dB(A), i.e. 20 dB(A) above the limit?  In terms of loudness, how can I
> quantify it? It 2X, 3X, ...times noisier than allowed?
>
> Is there an easy way to convert from dB(A) to dB?
>
> thanks
> fisico32

What I would say is:

dB(A) ARE dB!!  It's just that there is a particular reference weighting
filter involved.

I think you should start by understanding that this refers to a
broadband measurement - just to get away from any thoughts of spectral
density, etc.  And that the measurement, if based on amplitude, most
likely uses RMS values.

Because it's relative to a weighting filter then you still have to
define what is zero dB(A).  The common reference level is 20
micropascals in air.  But, if you don't have such an amplitude to deal
with then you may have to define your own.  Since dB is a ratio anyway,
it's not too inconvenient to do so - you just have to define what you've
done.  Then, converting to a different reference is only a matter of
adding or subtracting dB .. keeping the same weighting of course.

Fred

```
```On 10/4/2011 12:28 PM, Fred Marshall wrote:
> On 10/3/2011 4:07 PM, fisico32 wrote:
>> Hello forum,
>>
>> I understand what dB and dBm are. I am still a little confused about
>> dB(A).
>>
>> I know that the sensitivity of the human ear to sound depends on
>> frequency: two sounds of the same sound pressure but different
>> frequencies
>> may appear one louder than the other. Humans hear high frequency noise
>> much
>> better than low frequency noise.
>>
>> An A-weighting filter deemphasizes low frequencies. Decibels measured
>> using
>> this filter are A-weighted and are called dB(A).
>>
>> But how can I get a practical sense of what, say 30 dB(A), 20 dB(A), 60
>> dB(A) are? Is there an easy way to determine how much bigger is a certain
>> dB(A) with respect to another?
>>
>> For instance, 3 dB is easy: it means twice as much power. 10 dB means
>> 10X.
>> 20 dB means 200X....etc...
>> But 10 dB(A) means twice as loud, I think, in terms human loudness
>> perception. IF the noise limit is 80 dB(A), what if a noise meter
>> measures
>> 100 dB(A), i.e. 20 dB(A) above the limit? In terms of loudness, how can I
>> quantify it? It 2X, 3X, ...times noisier than allowed?
>>
>> Is there an easy way to convert from dB(A) to dB?
>>
>> thanks
>> fisico32
>
> What I would say is:
>
> dB(A) ARE dB!! It's just that there is a particular reference weighting
> filter involved.
>
> I think you should start by understanding that this refers to a
> broadband measurement - just to get away from any thoughts of spectral
> density, etc. And that the measurement, if based on amplitude, most
> likely uses RMS values.
>
> Because it's relative to a weighting filter then you still have to
> define what is zero dB(A). The common reference level is 20 micropascals
> in air. But, if you don't have such an amplitude to deal with then you
> may have to define your own. Since dB is a ratio anyway, it's not too
> inconvenient to do so - you just have to define what you've done. Then,
> converting to a different reference is only a matter of adding or
> subtracting dB .. keeping the same weighting of course.

Since dB(A) measurements are made after a (frequency-selective)
A-weighting filter, withought specifying a frequency, how can it be
related to a particular sound pressure?

Jerry
--
Engineering is the art of making what you want from things you can get.
```
```On 10/4/2011 6:34 PM, Jerry Avins wrote:
> Since dB(A) measurements are made after a (frequency-selective)
> A-weighting filter, withought specifying a frequency, how can it be
> related to a particular sound pressure?
>
> Jerry

RMS

Fred
```
```On 10/4/2011 10:10 PM, Fred Marshall wrote:
> On 10/4/2011 6:34 PM, Jerry Avins wrote:
>> Since dB(A) measurements are made after a (frequency-selective)
>> A-weighting filter, without specifying a frequency, how can it be
>> related to a particular sound pressure?
>>
>> Jerry
>
> RMS
>
> Fred

That's too terse for me to get what you mean. I was probably too terse
too. According to the curve at
http://en.wikipedia.org/wiki/File:Acoustic_weighting_curves.svg, the
same sound pressure that yields 0 dB(A) at 1000 Hz yields -30 dB(A) at
50 Hz. What have I missed?

Jerry
--
Engineering is the art of making what you want from things you can get.
```
```On Oct 5, 10:11&#2013266080;am, Jerry Avins <j...@ieee.org> wrote:
> On 10/4/2011 10:10 PM, Fred Marshall wrote:
>
> > On 10/4/2011 6:34 PM, Jerry Avins wrote:
> >> Since dB(A) measurements are made after a (frequency-selective)
> >> A-weighting filter, without specifying a frequency, how can it be
> >> related to a particular sound pressure?
>
> >> Jerry
>
> > RMS
>
> > Fred
>
> That's too terse for me to get what you mean. I was probably too terse
> too. According to the curve
athttp://en.wikipedia.org/wiki/File:Acoustic_weighting_curves.svg, the
> same sound pressure that yields 0 dB(A) at 1000 Hz yields -30 dB(A) at
> 50 Hz. What have I missed?
>
> Jerry
> --
> Engineering is the art of making what you want from things you can get.

That's what the weighing does, it's trying to match the human
frequency response, so each frequency gets weighted by a different
amount.

HTH
-Michael Shonle
```
```On 10/5/2011 10:30 AM, Michael Shonle wrote:
> On Oct 5, 10:11 am, Jerry Avins<j...@ieee.org>  wrote:
>> On 10/4/2011 10:10 PM, Fred Marshall wrote:
>>
>>> On 10/4/2011 6:34 PM, Jerry Avins wrote:
>>>> Since dB(A) measurements are made after a (frequency-selective)
>>>> A-weighting filter, without specifying a frequency, how can it be
>>>> related to a particular sound pressure?
>>
>>>> Jerry
>>
>>> RMS
>>
>>> Fred
>>
>> That's too terse for me to get what you mean. I was probably too terse
>> too. According to the curve
athttp://en.wikipedia.org/wiki/File:Acoustic_weighting_curves.svg, the
>> same sound pressure that yields 0 dB(A) at 1000 Hz yields -30 dB(A) at
>> 50 Hz. What have I missed?
>>
>> Jerry
>> --
>> Engineering is the art of making what you want from things you can get.
>
> That's what the weighing does, it's trying to match the human
> frequency response, so each frequency gets weighted by a different
> amount.
>
> HTH

Unfortunately, no. It seems to me that the dB(A) measurement depends
both on sound pressure and frequency. Freds statement, "Because it's
relative to a weighting filter then you still have to define what is
zero dB(A).  The common reference level is 20 micropascals in air"
prompts me to ask 20 micropascals at what frequency?

Jerry
--
Engineering is the art of making what you want from things you can get.
```
```On 10/5/2011 3:06 PM, Jerry Avins wrote:
Jerry,

I wasn't just terse but too glib as well.

Let us say that we are living in an A-weighted context.
Let us say that we are dealing with a single signal (Note I didn't say
"sinusoid") - or otherwise taking ratios might not be very meaningful
and particularly when there is weighting involved.

I should remember this.  Here's something somewhat similar:

Transmit a sinusoid at frequency f0 through a lossy medium where loss is
a function of frequency.  Such as acoustic waves in water.  Let us say
that the source level is 0 db (1 microPascal).
Receive the same signal with a calibrated hydrophone at 2 miles and
determine the attenuation in dB.

Now, transmit a sinusoid at another frequency, say at 10*f0 and perform
the same attenuation measurement accounting for possible variations in

So, we already know the answers should be different because of the medium.

Can we say that the signal at 10*f0 is X dB below the signal at f0?
Probably yes.

OK. Now let's transmit bandpass white noise that ranges from f0 to 10*f0
that has been subsequently A-weighted.  And let's say that the source
level is 20dB re 1uPa per sqrtHz at f0.  Having said this, we know the
entire spectral character at the source and should be able to calculate
or measure, either one, the rms value of the source signal.

Now we do the attenuation measurement again.  About all we can measure
is the rms level of the received signal.  It's both A-weighted and
weighted by the medium.

Might we compare the rms source level to the rms received level?
I think so.  How else might we compare them in a gross sense?

Anyway, that's what I was thinking.....

Fred

```