18 Feb
2011
18 Feb
'11
11:38 p.m.
On Fri, Feb 18, 2011 at 1:20 PM, Fons Adriaensen <fons(a)linuxaudio.org> wrote:
On Fri, Feb 18, 2011 at 11:30:24AM -0800, Mark Knecht
wrote:
Interesting. Thanks.
I would guess that this sort of solution makes some sort of assumption
of what the actual SPL is. Without that it seems at first blush (to
me) to be not that accurate in an uncalibrated environment, or do you
offer some way to calibrate the meter prior to depending on it? I know
I was pretty surprised how loud 85db SPL was when I got ahold of
calibrated equipment to test with.
I'll look forward to playing with your tools when you release them.
Thanks,
Mark
On Wed, Feb 16, 2011 at 10:45 AM, Fons Adriaensen
<fons(a)linuxaudio.org> wrote:
<SNIP>
In general that is not the case, probably because for most types of
content the mid-frequency range is dominant. And loudness measuring
systems are expected to give good results for typical content, not
to estimate the subjective loudness of single tones or narrow band
signals.
The filter recommended by the ITU and EBU is a combination of a 2nd
order highpass (around 50 Hz) and a shelf filter providing 4 dB gain
at 4 kHz and above. Listening tests have shown this to work quite
well in practice.
Ciao,
--
FA
Some filtering followed by RMS detection will provide a good
approximation to subjective loudness.
<SNIP>
Is this part to implement something like like Fletcher-Munson curves?