[klaatumail] Re: My final review of 3:47 EST on Klaatunes Records
- From: "Bradley, David" <David_Bradley@xxxxxxxxxxxxxxx>
- To: "'klaatumail@xxxxxxxxxxxxx'" <klaatumail@xxxxxxxxxxxxx>
- Date: Wed, 5 Oct 2011 17:48:33 +0000
> Dave - can you explain how the dynamic range numbers you are quoting are
> measured?
I don't know the mathematics used in the calculation, but the audio editing
software I use for my audio restoration and preservation work has a
"statistics" function which will use either a square wave, or a sine wave, and
calculate RMS for several different types of values.
Minimum RMS Power
Maximum RMS Power
Average RMS Power
The statistics I quote are usually Maximum RMS Power minus Average RMS Power.
The root-mean-square amplitude (RMS amplitude) of a signal relates to the
average amplitude of a sound wave over a given period. This period can either
be the entire wave or a portion that you have selected.
You may see something like a minimum of -93 dB, a maximum of -2 dB, and an
average of -11 dB. I take max of -2 and subtract -11 to show a 9 decibel range
for the dynamic range. The minimum is usually not useful as the songs fade out
to very low values.
So when I say there is 13-15 dB of dynamic range, what I'm indicating is that
the difference between the maximum RMS power and the average RMS power is
someone between 13 and 15 dB. Keep in mind that the two channels of audio can
have different values, so I have a range which incorporates both of the
channels. If the signal is mono, then there'd be a definite value. Since it's
stereo, I state the lower and the higher of the two numbers I come up with as a
range. So if left was 13 and right was 15 I'd say 13-15. If left was 15 and
right was 13 I'd still say 13-15.
Keep in mind that RMS, over the length of the song, will show averages, but
things like Maximum are based on the loudest point in the song. That loudest
point MAY be 0 on the scale, but the RMS power that is shown takes into account
more than just the one sample that is a 0. Usually, in the case of the default
settings with the software I use, the period of time is 50 ms. So while you
may have one sample that hits 0, over a 50 ms time period (which would be 2,205
samples per channel) you may average -2 dB RMS.
1 second at CD resolution is 44,100 samples per channel. 1 ms would be 44.1
samples. 50 ms would be 2,205 (44.1*50).
Dave
Other related posts:
