Re: equal loudness calculation

[Julian Rohrhuber <rohrhuber@xxxxxxxxxxxxxx>]

At 12:54 Uhr +0200 18.04.2005, Jose-Javier Lopez wrote:
> >  in this case, I don't want to filter a sound, but simply scale its
> >  amplitude.
> >  Or is there a way to make the formula itself efficient by
> >  implementing it as a filter?
>
> If you have your data in freq domain, just use the equation for weighting.
> In this case you don't need to filter in time domain.

ok, that seems good.


if anyone else needs it, here is the formula mapping freq to amp:

(
var k = 3.5041384e16;
var c1 = pow(20.598997,2);
var c2 = pow(107.65265,2);
var c3 =  pow(737.86223,2);
var c4 = pow(12194.217,2);

{|f|
  var r = pow(f,2);
  var m1 = pow(r,4);
  var n1 = pow(c1 + r, 2);
  var n2 = c2 + r;
  var n3 = c3 + r;
  var n4 = pow(c4 + r, 2);
  var level = k * m1 / (n1 * n2 * n3 * n4);
  sqrt(level)
};

)



> JJ
>
>
>
> ----- Original Message -----
> From: "Julian Rohrhuber" <rohrhuber@xxxxxxxxxxxxxx>
> To: <AUDITORY@xxxxxxxxxxxxxxx>
> Sent: Monday, April 18, 2005 12:18 PM
> Subject: Re: [AUDITORY] equal loudness calculation
>
>
> >  >This is the analitic equation for A-Weighting.
> > > http://www.beis.de/Elektronik/AudioMeasure/WeightingFilters.html#A-Weighting
> >
> >
> >  thank you, this is quite exactly what I was looking for.
> >
> > > Just, pass it to Z domain and use "a" and "b" coef. in a IIR filter.
> >
> >  in this case, I don't want to filter a sound, but simply scale its
> >  amplitude.
> >  Or is there a way to make the formula itself efficient by
> >  implementing it as a filter?
> >
> > > Very efficient !
> > >
> > > JJ
> > >
> > >
> > >
> > > ----- Original Message ----- From: "Julian Rohrhuber"
> > > <rohrhuber@xxxxxxxxxxxxxx>
> > > To: <AUDITORY@xxxxxxxxxxxxxxx>
> > > Sent: Tuesday, April 12, 2005 4:25 PM
> > > Subject: [AUDITORY] equal loudness calculation
> > >
> > >
> > > >   I'm looking for a computationally efficient way to do a frequency
> > > >   dependent amplitude compensation. The emphasis is much less accuracy
> > > >   (especially not for different loudness levels) but more a reasonable
> > > >   approximation - maybe a polynomial.
> > > >
> > > >   Any hints?
> > > >
> > > >   --
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >   .
> >
> >  --
> >
> >
> >
> >
> >
> >
> >
> >
> >  .

--








.