RMS -> SPL

Hi to everyone,
I've to calculate the SPL value of the waveform from the digital values of
audio samples incoming from the audiocard (Echo Layla24). I get the audio
sample buffer using Asio protocol. I can calculate the RMS value of buffer,
but how can i calculate the SPL value?? Is there a way to obtain the SPL
from RMS??

Thanks in advance!

"Emanuele" <eman1975@libero.it> wrote in
news:ivFjb.42076$vO5.1531943@twister1.libero.it: 

> Hi to everyone,
> I've to calculate the SPL value of the waveform from the digital
> values of audio samples incoming from the audiocard (Echo Layla24). I
> get the audio sample buffer using Asio protocol. I can calculate the
> RMS value of buffer, but how can i calculate the SPL value?? Is there
> a way to obtain the SPL from RMS??
> 
> Thanks in advance!
> 
> 
> 

To calculate the SPL (sound pressure level), you need to know the 
sensitivity of the transducer (mic?) and the gain structure of the 
soundcard. With a typical sound card this is sometimes more difficult 
than you think. The easiest way is to measure the output of a acoustical 
calibrator and compare.

Often a weighting network is applied (A weighting is most common). The 
value is usually expressed in dB re 20uPa. 



-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

"Al Clark" <dsp@danvillesignal.com> wrote in message
news:Xns9416BBF1D4432aclarkdanvillesignal@66.133.130.30...
> "Emanuele" <eman1975@libero.it> wrote in
> news:ivFjb.42076$vO5.1531943@twister1.libero.it:
>
> > Hi to everyone,
> > I've to calculate the SPL value of the waveform from the digital
> > values of audio samples incoming from the audiocard (Echo Layla24). I
> > get the audio sample buffer using Asio protocol. I can calculate the
> > RMS value of buffer, but how can i calculate the SPL value?? Is there
> > a way to obtain the SPL from RMS??
> >
> > Thanks in advance!
> >
> >
> >
>
> To calculate the SPL (sound pressure level), you need to know the
> sensitivity of the transducer (mic?) and the gain structure of the
> soundcard. With a typical sound card this is sometimes more difficult
> than you think. The easiest way is to measure the output of a acoustical
> calibrator and compare.
>
> Often a weighting network is applied (A weighting is most common). The
> value is usually expressed in dB re 20uPa.

The exact formula is

SPL = 20*log10( RMS/(20e-6) )
where,
RMS = SQRT(SUM(  S1^2 + S2^2 + S3^2 + ... + Sn^2 )/N), where Si is the i-th
sample of N samples.
where
Si are the values scaled according the the calibration factor you determine
from above.

This basically means you will need to decide on a time period (number of
samples) to applied the RMS calculation to.

"J" <j@n.c> wrote in message news:QbOdne-cGNZwyxKiRVn-sA@comcast.com...
>
> "Al Clark" <dsp@danvillesignal.com> wrote in message
> news:Xns9416BBF1D4432aclarkdanvillesignal@66.133.130.30...
> >
> > Often a weighting network is applied (A weighting is most common). The
> > value is usually expressed in dB re 20uPa.
>

I should add that it is preferred to A-weight a signal to make it relate
better to perceived loudness (although there are more accurate weighting
scheme A-weighting is a common standard in North America).

"Al Clark" <dsp@danvillesignal.com> a &#4294967295;crit >
| To calculate the SPL (sound pressure level), you need to know the
| sensitivity of the transducer (mic?) and the gain structure of the
| soundcard. With a typical sound card this is sometimes more
difficult
| than you think. The easiest way is to measure the output of a
acoustical
| calibrator and compare.
| Often a weighting network is applied (A weighting is most common).
The
| value is usually expressed in dB re 20uPa.

I agree.

Computing A-weighting curve coefficients (for fixed point DSP) was not
an easy task ! (but i succed with 4 biquads).
I first compute RMS on a small time period then compute LEQ
(mean(RMS)) on 1 second.
I compute log(LEQ) value and add a calibration constant (depending on
the pre-amplifiers)

If you want to calculate SPL you should use a time-constant too when
computing RMS (sound level meter usually displays fast or slow SPL)
Am I wrong ?

Bonne Chance ! :o)
Curl

"Curl" <Mr.Bilou@microsoft.fr> wrote in news:3f8fa6a2$0$27605
$626a54ce@news.free.fr:

> 
> "Al Clark" <dsp@danvillesignal.com> a &#4294967295;crit >
>| To calculate the SPL (sound pressure level), you need to know the
>| sensitivity of the transducer (mic?) and the gain structure of the
>| soundcard. With a typical sound card this is sometimes more
> difficult
>| than you think. The easiest way is to measure the output of a
> acoustical
>| calibrator and compare.
>| Often a weighting network is applied (A weighting is most common).
> The
>| value is usually expressed in dB re 20uPa.
> 
> I agree.
> 
> Computing A-weighting curve coefficients (for fixed point DSP) was not
> an easy task ! (but i succed with 4 biquads).
> I first compute RMS on a small time period then compute LEQ
> (mean(RMS)) on 1 second.
> I compute log(LEQ) value and add a calibration constant (depending on
> the pre-amplifiers)
> 
> If you want to calculate SPL you should use a time-constant too when
> computing RMS (sound level meter usually displays fast or slow SPL)
> Am I wrong ?
> 
> Bonne Chance ! :o)
> Curl
> 
> 
> 

Leq is a linear average and Fast (125ms) and Slow (1s) time constants 
refer to exponential averaging. Both are routinely used in acoustics 
measurements.

I would be interested in your A Weight filter solution and approach.




-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com