[LAD] Calculate R M S (Alfs Kurmis)

[Dominique Michel]

Le Fri, 18 Feb 2011 21:05:36 +0000,
Fons Adriaensen <fons at linuxaudio.org> a écrit :

> On Sat, Feb 19, 2011 at 09:16:32AM +1300, Jeff McClintock wrote:
> 
> > > From: Fons Adriaensen <fons at linuxaudio.org>
> > > On Fri, Feb 18, 2011 at 07:36:44AM +1300, Jeff McClintock wrote:
> > > 
> > > > With a RMS VU Meter you measure a 1KHz tone as a reference.
> > > 
> > > A contradiction... A VU does not measure RMS, whatever does
> > > measure RMS is not a VU.
> > 
> > Isn't a VU Meter a standard root-mean-square function followed by a
> > 300ms integration to give it some 'weight'? ...calibrated against a
> > 1kHz tone?
> 
> No. VU meters were used in the times when audio equipment was using
> tubes (valves) so they could not use complex electronic processing,
> at most an amplifier stage to drive a bridge rectifier and a passive
> moving coil meter. So, ignoring the rectifier threshold, the current
> driving the meter would be the absolute value of the signal, not the
> square of it.

Such indicators shows the average value of the signal, not the
instantaneous value. In case of DC, they are the same, but they are
quite differents in case of AC signals like audio. That is why the
needle should never exceed -6db when recording with old analog
equipment.
The exact value you should not exceed depend of the kind of music you
are recording.

> The meter itself is equivalent to a second order lowpass
> filter, and its response was quite strictly specified. For a steady
> signal, it should rise to 99% of the final value in 300ms, and
> overshoot it by 1 to 1.5% before falling back to the real value. The
> overshoot isn't a detail - it has quite a marked effect on the
> response. Many software 'VU' meters use a first order lowpass - this
> doesn't even come close to the response of a the real thing.
> 
> Ciao,
> 


-- 
"We have the heroes we deserve."